ATTACKllG CH(SS

KllG'S llDIAI VOlUlf 1

DAVID VIGORITO
EVERYMAN CHESS
www.everymanchess.com
First published in 2010 by Gloucester Publishers plc (formerly Everyman Publishers
pk), North burgh House, 10 Northburgh Street, London EC1 V OAT

Copyright© 2010 David Vigorito

The right of David Vigorito to be identified as the author of this work has been as­
serted in accordance with the Copyrights, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system or transmitted in any form or by any means, electronic,
electrostatic, magnetic tape, photocopying, recording or otherwise, without prior
permission of the publisher.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.

ISBN: 978 1 85744 645 6

Distributed in North America by The Globe Pequot Press, P.O Box 480,
246 Goose Lane, Guilford, CT 06437-0480.

All other sales enquiries should be directed to Everyman Chess, North burgh House,
10 Northburgh Street, London EC1 V OAT
tel: 020 7253 7887 fax: 020 7490 3708
email: info@everymanchess.com; website: www.everymanchess.com

Everyman is the registered trade mark of Random House Inc. and is used in this
work under licence from Random House Inc.

This book is dedicated to two people: for my darling wife Heather, for all her love
and support; and for my dear friend Joe Fang.for getting me well on my way in the
King's Indian Defence with many long nights of speed chess back in the olden days...

Everyman Chess Series

Chief advisor: Byron Jacobs
Commissioning editor: John Emms
Assistant editor: Richard Palliser

Typeset and edited by First Rank Publishing, Brighton.

Cover design by Horatio Monteverde.
Printed and bound in Great Britain by Clays, Bungay, Suffolk.
Contents

Bibliography 5
Introduction 7

Part I: The Classical Variation

1 d4 lt:Jf6 l c4 g6 3 lt:Jc3 il.g7 4 e4 d6 S lt:Jf3 o-o
6 ii.el es 7 0-0 lt:Jc6 10

1 Mar del Plata Variation: 9 tt:Je1 lt:Jd7 10 tt:Jd3 12

2 Mar del Plata Variation: 9 tt:Je1 lt:Jd7 10 f3 38
3 Mar del Plata Variation: 9 tt:Je1 lt:Jd7 10 .ie3 47
4 Mar del Plata Variation: 9 tt:Jd2 76
s Mar del Plata Variation: 9 b4 tt:Jhs without 10 .:tel 92
6 Mar del Plata Variation: 9 b4 tt:Jhs 10 .!:tel 105
7 Mar del Plata Variation: White's Other 9th Moves 144
8 White's Eighth Move Deviations 156

Part II: The Classical Variation

1 d4 lt:Jf6 l c4 g6 3 lt:Jc3 il.g7 4 e4 d6 S lt:Jf3 o-o
6 ii.el es without 7 o-o 167

9 Gligoric Variation: 7 .ie3 lt:Jg4 8 ii.gs f6 9 ii.cl tt:Jc6, 9 ilh4 tt:Jc6 168
10 Gligoric Variation: 7 .ie3 lt:Jg4 8 ii.gs f6 9 il.h4 gs 184
11 Petrosian Variation: 7 dS as without 8 ii.gs 202
12 Petrosian Variation: 7 dS as 8 ii.gs 222
13 Exchange Variation : 7 dxes dxes 8 '1i'xd8 .l:txd8 238

Part Ill: The Samisch Variation

1 d4 lt:Jf6 2 c4 g6 3 lt:Jc3 .ig7 4 e4 d6 S f3 o-o 6 .ie3 tt:Jc6
7 tt:Jge2 a6 8 '1i'd2 :tbs 264

14 Panno Variation: 9 h4 without 9 ... h s 266

1S Panno Variation: 9 h4 h S 10 o-o-o 2 73
16 Panno Variation: 9 h4 hS 10 tt:J c1 293
17 Panno Variation: 9 tt:J c1 302
18 Panno Variation: 9 0-0-0, 9 .ih6 and 9 g4 320
19 Panno Variation: 9 a3, 9 !:!.bl and 9 .:tel 331
20 Panno Variation: Other lines 344

Index of Variations 3 64
 B.ibliography

Books
6 ... liJc6 in the Sii.[nisch Variation, King's Indian Defense, John Watson (Chess Enter­
prises 1982)
Attack ing Manual I, Jacob Aagaard (Quality Chess 2008)
Beat the KID, Jan Markos (Quality Chess 2009)
Beating the Anti-King's Indians, Joe Gallagher (Batsford 1996)
Beating the Fianchetto Defences, Efstratios Grivas (Gambit 2006)
Dangerous Weapons: The King's Indian, Richard Palliser, Glenn Flear & Yelena
Dembo (Everyman Chess 2009)
King's Indian Defence: 4 e4, Efim Geller (Batsford 1980)
King's Indian Defence: Ma r del Plata Variation, Svetozar Gligoric (Batsford 2003)
My Great Predecessors Part Ill, Garry Kasparov (Everyman Chess 2004)
My Great Predecessors Pa rt IV, Garry Kasparov (Everyman Chess 2005)
Play 1 d4!, Richard Palliser (Batsford 2003)
Play the King's Indian, Joe Gallagher (Everyman Chess 2004)
Starting Out: 1 d4!, John Cox (Everyman Chess 2006)
Starting Out: The King's Indian, Joe Gallagher (Everyman Chess 2002)
The Art of the King's Indian, Eduard Gufeld (Batsford 2002)
The Classical King's Indian Uncovered, Krzysztof Panczyk and Jacek llczuk (Every­
man Chess 2009)
The Controversial Sii.misch King's Indian, Chris Ward (Batsford 2004)
The King's Indian: A Complete Black Repertoire, Victor Bologan (Chess Stars 2009)
The King's Indianfor the Attacking Player, Graham Burgess (Batsford 1993)
The Main Line King's Indian, John Nunn & Graham Burgess (Batsford 1996)
The New Classical King's Indian, John Nunn & Graham Burgess (Batsford 1998)
The Sii.misch King's Indian, Joe Gallagher (Batsford 1995)
The Sii.misch King's Indian Uncovered, Alexander Cherniaev and Eduard Prokuronov
(Everyman Chess 2008)
Understanding The King's Indian, Mikhail Golubev (Gambit 2005)

5
A t ta c king C h e s s : The King 's In dian, V o l u m e 1

Winning with the King's Indian, Eduard Gufeld (Macmillan 1991)

Winning with the King's Indian, Andrew Martin (Caissa 1989)

Periodicals
Chess Informant through Volume 105
New In Chess Magazine through issue 2010/4
New In Chess Yearbook through Volume 94

Electronic Resources
Mega Database 2009 (ChessBase)
The King's Indian, Viktor Bologan (ChessBase 2009)
Chessbase.com
Chesscafe.com
Chess Lecture.com
Chess Publishing .com
The Week in Chess through issue 815

6
 Introduction

The King's Indian Defence is one of the richest openings in all of chess theory.
Black does not play to equalize as he does in the classical defences. Rather he seeks
to unbalance the game from the outset. The last decade has seen a revitalization
of the King's Indian, as even top players are often trying to win with the black
pieces. Compared to the classical openings, the price of each move is quite high
and a mistake by either side can easily lead to disaster.
The King's Indian has always been considered a somewhat risky opening, but
despite that common sentiment, the King's Indian has an impressive pedigree.
While this dynamic system was pioneered in the 1950s by Russian and Yugoslav
players such as David Bronstein, Efim Geller and Svetozar Gligoric, the two big
names that are often attached to the King's Indian are those of its World Cham­
pion practitioners, Robert Fischer and Garry Kasparov. Whereas Fischer's retire­
ment signalled the end of his King's Indian era, Kasparov gave up our favourite
opening while he was still an active player, which 'indicated' its unsoundness. At
least that was the general feeling after he lost a well-known game in 1997 to
Kramnik in the then dreaded 'Bayonet' system.
In fact Kasparov stated something to the effect that the Sicilian and King's In­
dian were too much to keep up with at the level he was playing at, and so he stuck
with the Sicilian while heading for more solid systems in the closed openings.
Nowadays young players are not so worried about this; with advances in technol­
ogy many modem talents play both the Sicilian and the King's Indian, as well as
other sharp defences.
Opening fashions come and go. The beginning of the new millennium brought
forward a great new champion of the King's Indian Defence in Teimour Radjabov.
Like Kasparov, Radjabov hails from the city of Baku in Azerbaijan. Radjabov really
took over where Kasparov left off, even scoring well in the aforementioned Bayo­
net (see Chapters 5 and 6). Radjabov's success influenced the younger generation
as well as the old guard, and nowadays most of the top players have been found at
one time or another on the black side of the King's Indian.

7
A ttac king Ch ess: Th e King 's In dian, Vol u m e 1

The King's Indian Defence has always been an opening I've felt greatly attached
to. Despite the fact that I have written extensively on the Slav Defences, the King's
Indian was my first real defence to 1 d4. While the King's Indian is considered to
be a 'tactical' opening, I have always considered it to be very strategic in nature. It
is an opening where a feeling for piece placement and pawn structure is very im­
portant. There are many thematic ideas and although the opening lends itself to
frequent complications, the tactics have always seemed 'logical' to me. So, while it
is true that when I 'grew up' I began to rely more on the solid Slav systems, it is
always useful to have a sharp weapon available, especially when one really wants
to try to win with Black.
Even though the King's Indian is a complicated opening, I do not think it is so
difficult to learn. For one thing, it is relatively 'move order proof'. That is, the King's
Indian set-up can be employed against 1 d4, 1 c4, or 1 tt:Jf3. Also, the King's Indian
lends itself to just a handful of pawn structures, so the ideas are easier to assimi­
late.
In this book, as well as the second volume, I will generally focus on the main
lines. The reason for this is that I think the best way to learn an opening is to study
the main lines. It is easy to add other secondary systems later. The biggest exam­
ple of this is in the Classical Variation, 1 d4 tt:Jf6 2 c4 g6 3 tt:J c3 .ig7 4 e4 d6 5 lt:Jf3
o-o 6 ile2 es 7 o-o. Here I have gone for what is no doubt the main line, 7 ...tt:Jc6.
This is the most uncompromising approach and it is also the most difficult. The
main alternative is 7... tt:Ja6. This line is safer, easier to learn and may well be just
as good as 7... tt:Jc6. In fact, I have pl ayed 7 ... tt:Ja6 myself on several occasions. Nev­
ertheless, I think it is best for the aspiring King's Indian player to learn the main
lines. One great thing about the King's Indian is its flexibility - if you learn the
main systems, it is easy to expand your repertoire by adding additional lines with­
out having to learn a whole new opening.
In the Sa mi sch Variation, 1 d4 tt:Jf6 2 c4 g6 3 tt:Jc3 .ig7 4 e4 d6 5 f3 o-o 6 .ie3,
the main line nowadays is 6... cs. This was not always the case, and 6 ...tt:J c6 and
6 ... es used to be considered Black's two main systems. I have elected not to go
with 6 ... cs even though it may well be the best move. While White often steers the
game into a Benoni structure, it is also possible for White to simply grab a pawn
while exchanging queens as well. Modem practice has clearly shown that Black
gets sufficient compen sation for the pawn, but some White players are rather well
prepared in these endings. If Black is well prepared too and has a good under­
standing of these positions, he should certainly be able to draw, but I believe that
heading into a pawn-down endgame straight out of the opening is probably not
to everyone's taste.

8
 In tro d u c tion

Moreover, 6...cs against the Samisch has been very well covered in modem
King's Indian literature and I did not have a lot to add to what is already out there,
especially as the lack of popularity of the Samisch at high level has not produced
much fresh material to examine. The classical 6 ... es must still be playable, but I
always thought it was easier to play White in these lines and so decided to go for
the Panno Variation with 6 ... tt:Jc6. This is an uncompromising system that still has
a lot of fresh territory to explore.
Note that the second volume will cover the Fianchetto Variation, the Four
Pawns Attack, the Averbakh and all of White's other tries.
I would like to thank a few people for their help with this book: John Emms, for
suggesting the topic, allowing me to split the book into two parts and for his pa­
tience; Richard Palliser, for listening to me rant and rave about various lines that
may or may not have found their way into this book; Joe Fang, for the use of his
impressive library; and Vasik Rajlich, for keeping me well supplied with Rybka 3
and 4. Thanks to you all!

David Vigorito
Somerville, Massachusetts,
October 2010

9
 Part 1
The Classical Variation
1 d4 lbf6 2 c4 g6 3 ttJc3 i..g7 4 e4 d6
s tt:Jf3 o-o 6 il.e2 es 7 o-o lbc6

6...tt:Ja6 and 6 ...tt:Jbd7, which aim to

avoid certain lines in Part II (most nota­
bly the Exchange Variation), but in my
opinion we should dive headfirst into
the most critical lines. After 7 0-0 (other
moves are considered in Part I I ) we will
play 7 ... tt:Jc6!. Here White usually plays 8
ds (other moves are discussed in Chap­
ter Eight), and after 8 .. lt:Je7 we have the
.

Mar del Plata Variation.

The Classical Variation 1 d4 tt:Jf6 2 c4

g6 3 ltJC3 ilg7 4 e4 d6 5 tt:Jf3 0-0 6 .ie2
is one of White's most principled sys­
tems against the King's Indian Defence.
White develops in a logical manner by
simply developing his kingside and
preparing to castle. Despite this appar­
ently modest approach, the Classical is
a very complex set of systems and sub­
systems which can lead to many differ-
ent types of position. Most of the world's top players have
We will focus on the main move, been found on one side or another of
6 ... es!. There are other moves, such as this position (some players like both

10

sides), because the play is very rich
both strategically and tactically. In
general White will play on the queen­
side and Black on the kingside, but
there are many instances where one
side can take the initiative on their
'weak' side.
One feature that is specific to the
Mar del Plata Variation is the position
of Black's e7-knight. In many King's
Indian lines with a blocked centre,
Black's queen knight would head for
the natural cs-square by ...tt:Jb8-d7-c5
or ... ltJb8-a6-cS (see, for example, the
Petrosian Variation of Chapters 11 and
12). In the Mar del Plata, Black has
played 7 ...tt:J c6 in order to entice White
to clarify the structure in the centre
with 8 dS. In general the knight on e7 is
not well placed and Black's success is
often connected in some way with ac­
tivating this piece. In many of the lines
with 9 tt:Je1, Black brings the knight to
g6. The knight is not necessarily well
placed there, but it may participate in a
kingside attack. Alternatively, the
knight may emerge on fS (after an ex­
change of pawns on fS or e4), and
sometimes Black will redeploy the
The Classic a l Va ria tion
knight by playing ...'it>h8 and ...tt:Jg8.
This looks funny at first, but the knight
can re-emerge on f6 or even h6. There
are also cases where the knight goes to
c8 or even to c6.
White has four basic ways to pro­
ceed in this position. He can play 9
tt:Je1, which can be considered the
'main line'. This move is considered in
Chapters 1-3. A different knight rede­
ployment is 9 tt:Jd2, which is covered in
Chapter 4. The dangerous 'Bayonet', 9
b4, has been White's main weapon for
the last couple of decades. The play
tends to be quite different here, as cen­
tral play is more prominent than it is
after White's knight moves. 9 b4 is cov­
ered in Chapters 5 and 6. Other 9th
moves are less common but they are
not altogether harmless. These lines
will be examined in Chapter 7. Some­
times too White chooses to avoid the
discussion of the Mar del Plata Varia­
tion and avoids the critical 8 dS. These
sidelines are not so dangerous, but
Black should not ignore them, as they
can be tricky to meet for the unpre­
pared. White's 8th-move deviations are
covered in Chapter 8.
11
 Chapt�r1· . ·.
The Mar def Plata Variation
1 d4 lt:Jf6 2 c4 g 6 3 tt:Jc3 .ig7 4 e4 d6 5
tt:Jf3 0-0 6 ii.el es 7 0-0 tt:Jc6 8 ds lt:Je7 9
tt:Je1
This move quickly became White's
main try when the Mar del Plata Varia­
tion first became popular and it re­
mains topical to this day. White pre­
vents Black's knight manoeuvre ...tt:Jhs,
prepares to bolster the centre with f3 if
necessary and avoids blocking in his c1-
bishop. Although the knight is passively
placed at the moment, it can quickly
come to d3, from where it surveys the
important central squares cs, es and
12
. . .
f4 The knight will help force through
White's important c4-cS advance and it
may also drop back to the f2-square,
where it will help prevent Black's ... gs­
g4 pawn break.
One thing we must always be aware
of is the value of Black's c8-bishop.
Without this piece not only is it diffi­
cult to break with ... gS-g4, but the
bishop is often needed to sacrifice itself
on h3 if the kingside becomes blocked.
Sometimes the bishop does not even
move until it can deliver the decisive
blow. So valuable is this bishop to
Black's attack that we will often see
White spending a number of tempi to
hunt down the bishop on its original
square.
9 lt:Jd7
...
Black clears the way for .. .f7-fS and
covers the cs-square at the same time.
Another plan for Black is 9 ...tt:Je8. This
looks less logical, but as we shall see in
the main lines following 9 ... lt:Jd7 10
tt:Jd3 fS 11 .id2 tt:Jf6 12 f3 f4, Black of-
 Th e M a r de/ Plata Varia tio n : 9 tlJe1 tlJ d 7 10 tiJd3
ten drops the knight back to e8 after all
in order to protect the sensitive c7-
square. By playing 9...t2Je8 straight­
away, Black hopes to gain a couple of
tempi. Of course White does not need
to be so cooperative though. For one
White can continue with 10 t2Jd3 fS,
but instead of 11 .id2, as played in this
chapter, White may change the charac­
ter of the position with 11 f4, exploit­
ing the fact that Black's knights are far
away from the es-square. Another
dangerous plan is to continue along
the lines of Chapter 2 by playing 10 f3
fS 11 g4, when the knight on e8 is less
flexible than it would be on d7. Perhaps
the greatest danger, though, comes
from 10 .ie3. In this case White will
have no trouble playing c4-c5 and
plans involving a4-a5 become very
dangerous.
While 9 ... t2Je8 may be playable, the
main move has always been 9 ...l2Jd7
and most experts consider it best.
From here White has three main
schemes to choose between. In this
chapter we look at the flexible 10 t2Jd3.
The next chapter examines the radical
plan 10 f3 fS 11 g4! ?, while Chapter 3
covers White's most aggressive con­
tinuation, 10 .ie3.
10 t2Jd3
White improves the position of his
knight and prepares the c4-c5 advance.
This is a very straightforward continua­
tion and some of the theory goes very
deep in the main lines. The fashion of
such long lines comes and goes, and
eventually White turned to other sys­
tems in the Classical (most notably 10
.ie3 and the Bayonet), while Black also
found ways to deviate. Recently 10 t2Jd3
has become rather fashionable again.
We will generally study the main lines
because that will help to develop a
good general King's Indian under­
standing. There are also some interest­
ing sidelines that will be mentioned
that may be worthy of further research.
10 f s 11 .id2
...
This is the main line. White simply
develops his bishop and prepares !:!.cl.
There are a couple of alternatives.
a) First, let us take a look at 11 f3.
13
A ttacking Ch ess: The King 's In dian, Vo l u m e 1
This is quite a natural reaction to
Black's pawn advance, but sometimes
this move is considered inaccurate.
This is because Black can begin his
kingside play while leaving his knight
on d7 to hold up White's cs-break.
Nevertheless, this move is not bad and
some even consider it to be the most
accurate move because it avoids the
line 11 .id2 fxe4!?, which we discuss in
the notes to Black's 11th move, below.
White may also steer the game into
independent channels by avoiding .id2
altogether. Here Black can play 11 .. .f4
right away, because there is no .ie2-g4
to worry about. 11 . ..tt:Jf6 is possible too,
but Black can try to take advantage of
White's move order by leaving the
knight at d7 to hinder White's cs­
advance. Note that 11 ... 'it>h8?! is not so
good because White can switch gears
with 12 .ie3!, aiming for positions akin
to those in Chapter 3 where 11 ...'it>h8 is
likely to be a wasted tempo.
After 11 .. .f4 White has:
al) 12 .id2 looks like an inferior
version of the main line because White
14
will have to spend an additional tempo
to break with c4-cS. This move does
have some historical significance, how­
ever. Now 12 ... gs 13 b4
13 ...lt:Jf6 14 cs tt:Jg6 1S .l:tc1 .l:tf7
compares favourably to the main line
for Black because White has played b4
(instead of tt:Jbs or cxd6, for example)
which may or may not tum out to be a
useful move.
Black's last few moves seem rather
obvious today, but this was not always
so. An old plan for Black in such posi­
tions is to play ....l:tf6 and ... .rf.h6, with
the idea of directly attacking down the
h-file with ...'1i° e8 and ...'ifh s. If White
plays h2-h3, Black will try to engineer a
....ixh3 sacrifice. Although this plan
can indeed be quite dangerous to the
first player, to me it seems a bit too
simplistic to think that White's position
can just be taken by storm like this.
Often Black will have to worry about
the possibility of White playing tt:Jbs
after the queen has gone to e8. He can
deal with this by throwing caution to
the wind and allowing tt:Jbsxe7. This
 Th e M a r de/ Plata Va ria tion : 9 li.Je1 li.J d 7 10 li.Jd3
sounds like fun, but if White can play
li.Jf2 and h3, the h-file attack will not
work. In this case, Black can change the
direction of the attack by then playing
... .l:tg6 and ... hs.
Another way for White to defend is
by playing 'it>h1 and .ig1. Black does
not really have any way to increase the
pressure on h2, but if he could get a
knight to g 3, it would be mate. This can
be achieved either by moving the black
queen from h S to threaten ...li.Jf6-hS or
by a sacrifice such as ... li.Jf6xdS; exds
li.Je7-fS. This all sounds very nice, but if
White is alert to the possibility it is not
so easy to achieve these tricks.
White may also defend against
mate on h2 in some positions with his
bishop on f2 by playing the ugly h4.
This may look fatally weakening, but it
is not always easy to get through and
Black's queenside will certainly disap­
pear.
Black can also throw in ... a6 before
moving the queen to safeguard against
the knight raid to bS, but this costs a
tempo and also leaves Black vulnerable
to li.Ja4-b6 ideas after an exchange of
pawns on d6. In general I have avoided
lines where Black seeks to attack in this
way, but there are exceptions as we
shall see, most notably in Line C of
Chapter 3.
Another attacking plan, which
seems more natural to me, is to play
... li.Jf6 and ... hs. Gligoric was one of the
first players to develop this concept.
One of his games continued (after
1s ... :tf7 above) 16 li.Jf2 h S 17 cxd6 cxd6
18 a4 .if8! 19 as .l:tg7 20 h3 li.Jh8!.
As we shall see, Black's 18th and
19th moves are extremely common­
place, but the last one requires a few
words of explanation. The g6-knight
often heads to h4 to both unleash the
g7-rook and scare the white king. The
knight manoeuvre to h8 is not so un­
common, though, and the knight will
then come back to f7, from where it
may hop to gs or even to h6, to further
support the ... g4-break. We will even
see positions where Black plays ... 'it>h8
and ... li.Je7-g8 in order to come to h6.
The game continued 21 li.Jbs g4 22 fxg4
hxg4 23 hxg4 a6 24 li.Ja3 .id7 2S li.Jc4
.l:tc8 26 li.Jb6 .l:txc1 27 .ixc1 .ie8 28 .ia3
li.Jf7 29 '1i'c2 li.Jh6 30 gs .l:txgs 31 !:i.c1
.l:tg3 32 .ib2 li.Jfg4 33 li.Jxg4 li.Jxg4 34
.ixg4 .l:txg4 3S '1i'f2 .ig6 36 .l:tc4 '1i'e7 37
.ic3 '1i'h7 38 '1i'e2 .l:Ih4 39 'it>f2 f3 40 '1i'e3
.l:tf4 41 gxf3 '1i'h2+ 42 'it>e1 '1i'h1+ 43
'it>e2 .ihs 44 'it>d2 .:txf3 4S 'iY'gs+ .ig7
46 'it>c2 !:i.f2+ 47 .id2 '1i'd1+ 48 'it>c3
'1i'a1+ 0-1, M.Najdorf-5.Gligoric, Mar del
Plata 19S3. Najdorf learnt a lesson
15
A tt a c king C h e s s : Th e King 's In dian, Vo l u m e 1
from this game as we shall see in Line B
of Chapter 3.
Going back, White probably can re­
frain from 13 b4 and play 13 !:!.cl tt:J g6
14 cs!? when instead of grabbing the
pawn, Black should really just play
14 ...tt:Jf6 to tran spose to the main lines,
while 13 ... hs is also possible.
Another idea is 13 g4 when Black
can pl ay 13 ... hs 14 h3 !:i.f6!? (Black tries
to turn the delay in going ...tt:Jf6 to his
benefit) 1S b4 (White should probably
prefer a quick ii.el and 'it>g 2) 1S ... !:i.h 6
16 'it>g2 tt:Jg6 17 !:i.hl lt:Jh4+ 18 'it>f2 .if8
19 !:!.cl as 20 a3 c6, which gave him
good play in R.Shabtai-D.Komljenovic,
Biel 1989.
a2) 12 g4 is quite a common
counter in general - see, for instance,
Line D of this chapter. Here, however, it
does make a difference that Black has
not played either ...tt:Jf6 or ... gs yet.
White's thrust is well met by
12 ....if6 !. Black's delay in playing ... tt:Jf6
allows this move, seizing control of the
h4-square. After 13 .id2 hS 14 h3 .ih4
(alternatives are 14... �7 and
16
1 4...'it>g 7!?) 1 S ii.el .ixel 16 tt:Jxel 'it>g7
17 h4?! gs (or 17 ... hxg4 18 fxg4 .l:th8 19
tt:Jg2 gs 20 hs tt:Jf6) 18 tt:Jg2 hxg4 19
fxg4 gxh4 20 tt:Jxh4 .l:th8 Black had a
strong attack brewing in S.Karp­
A.Kuzmin, Ostend 1991.
a3) 12 b4 (in the main line we will
see that this preparatory move usually
is not necessary; the only way this
move makes sense for White is if he
intends to deploy the bishop on a3 or
b2) 12 ... gs (Black has also tried to take
advantage of White's move order with
12 ... as, but I think Black should just
leave the queenside alone and get on
with it) 13 cs tt:Jf6 14 a4 and now:
a31) 14...'it>h8 ! ? (here we see an­
other typical plan; Black gears up for a
quick ...g4, rather than play ....l:If7-g 7,
which makes some sense because
White has refrained from .id2 and
!:!.cl, so the c7-square does not really
need additional protection) 1S .ia3
.l:tg8 16 as .if8 17 cxd6 cxd6 18 bS (or
18 h3 hS 19 tt:Jf2 when Black could try
19 ....l:I g7 20 !:!.cl tt:J eg8, with the idea
...tt:J h6 and ...g4) 18 ... g4 and Black had
 Th e M a r de/ Plata Va riatio n : 9 li.J e1 li.Jdl 10 li.J d3
good play in S.Lagrotteria-B.Socko,
Saint Vincent 2001.
a32) The direct 14...h s is also possi­
ble and after 1s li.Jf2 li.Jg6 16 as .l::i. f7 17
c6 .if8 18 cxb7 .ixb7 19 !:i.a3 .l:tg7 20 h3
.ic8 21 bS li.Jh8 22 li.Ja2 li.Jf7 23 li.Jb4
li.Jh6 24 li.Jc6 '1i'e8 2S '1i'c2 g4 Black had
good counterplay in G.Vescovi­
M.Krylov, Moscow 2010.
b) The old move 11 exfs is not seen
too often these days. The simplest solu­
tion is 11 ...li.Jxfs. Usually in the King's
Indian Black prefers to take with the f­
pawn, but here Black immediately
solves the problem of the e7-knight
and has good piece play. That said,
11 ... gxfs 12 f4 li.Jg6 is also possible. Af­
ter 11 ...li.Jxfs 12 f3 li.Jf6 13 li.Jf2 Black
has:
bl) 13 ....ih6 14 .ixh6 li.Jxh6 and
now 1S g4 was suggested by Hort, to
prevent the knight from coming to d4,
but after 1s ...li.Jf7 Black looks to be fine
in any case.
b2) 13 ... c6 was Nunn's recommen­
dation. However, 14 li.Jfe4 li.Jxe4 (in-
stead 14... cxds 1s li.Jxf6+ .ixf6 16 li.Jxds
.igs 17 .id3 gave White a slight edge
in F.Doettling-A.Hunt, Patras 1999) 1S
li.Jxe4 'ii'b6+ (1S ...cxds 16 'iY'xds+) 16
cs!? dxcs 17 .ic4 gives White danger­
ous compensation for the pawn .
b3) 13 ... cs 14li.Jfe4 li.Jd4 1 s .igs .ifs
looked fine for Black in T.Overbeck­
V.Rojicek, Pardubice 2009.
b4) 13 ... li.Jd4 is simple enough: 14
li.Jfe4 c6 1S .ie3 li.Jxe4 16 li.Jxe4 cxds 17
cxds .ifs was equal in R.Markus­
R.Polzin, Austrian League 2006.
Now we finally return to the main
line, 10 .id2.
11 li.Jf6
•••
At the cost of 'letting go' of the cs­
square, Black induces White to play f3
before advancing on the kingside. Note
that the immediate 11 .. .f4? is a big
strategic mistake because after 12 .ig4
Black will be hard pressed to avoid the
exchange of the light-squared bishops.
Black does have a couple of other ideas
here, though:
a) One possibility is 11 ...'it>h8!?.
This is Black's main alternative to
the main lines. If Black wants to avoid
17
A ttac king Ch ess: The King 's In dian, Vo l u m e 1
the long main lines of 'A' and 'B', the
complications of '(' and the blocked
positions of 'D', this is a decent alterna­
tive. Instead of fixing the kingside,
Black makes a useful, semi-waiting
move. Playing 11 ... 'it>h8 does a few
things. The king gets off of the a2-g8
diagonal, which may seem rather
vague, but in the main lines we will see
a few situations where this could be
useful. Black also moves his king off of
the g-file. This could be good or bad -
we will come across positions where
after multiple exchanges on g4, a white
rook or queen ends up sliding over to
h3. Perhaps the biggest advantage to
11 ... 'it>h8 is that the g8-square is
cleared for Black's knight. Playing
... lt:Je7-g8 creates the possibility of play­
ing ....�h6 and clears the h4-d8 diago­
nal for both Black's queen and g7-
bishop. This can be useful if White
adopts a plan involving g4. And, lastly,
by leaving the knight on d7, Black
hopes to make White spend an extra
move preparing the cs-advance.
Perhaps the most natural continua­
tion is 12 !:!.cl. White hopes to get a
favourable version of the main lines by
opening the c-file. Now 12 ... tt:Jf6 13 f3
f4? just allows 14 cs and thus is a
worse version of the main lines, be­
cause Black has played ... 'it>h8 rather
than ... gs. In stead Black usually plays
13 ... cs to try and make the move .l:tc1
look irrelevant, while 13 ...tt:J eg8 and
13 ... c6 are altern atives.
b) Another idea is 11 .. .fxe4!?. This
18
looks too simplistic, but it i s not easy
for White to prove an advantage. Mar­
kos, for one, thinks enough of this reac­
tion to avoid playing 10 tt:Jd3 alto­
gether. After 12 tt:Jxe4 tt:Jfs (12 ... tt:Jf6
looks less accurate because Black
would like to recapture on f6 with the
queen) it seems as though Black has
lost time compared to the line 11 exfs
tt:Jxfs, but there White generally prefers
to play f3, tt:Jd3-f2 and tt:Jfe4. White's
knight moves have also left him with
less control of the dS-square, so a ...c6-
break may be appealing.
Now 13 f3 tt:Jf6 14 tt:Jdf2 c6!
(14... lt:Jd4 is also okay) looks fine for
Black. One example: 1S .id3 cxds 16
cxds 'it>h8 17 tt:Jxf6 .ixf6 18 lt:Je4 '1i°b6+
19 .l:tf2 .ih4 20 g3 .�e7 with a good
game for Black in J.Sofrevski-L.Portisch,
Skopje 1968.
A better idea is 13 .ic3, although
this is not too terrifying either: 13 ... tt:Jf6
14 .if3 lt:Jh4 (14...tt:Jxe4 15 .ixe4 '1i°h4
16 .:tel b6 17 g3 looks a little better for
White) 1S tt:Jxf6+ '1i"xf6 16 .ie4 .�fs 17
'1i°e2 .ixe4 18 '1i"xe4 'iY"fs 19 .l:tael was
 Th e M a r d e/ Plata Va riation: 9 li.J e 1 li.Jd7 10 li.Jd 3
agreed drawn here in P.Lukacs­
T.Radjabov, Budapest 2000. After
19 ... .l:tae8 20 f3 'iY"gs 21 '1i"e2 li.Jf5 22 li.Jf2
ii.. h6 Black should be able to hold the
balance.
Overall, 11 .. .fxe4 looks very solid, al­
though it is a little dull.
These lines are interesting and look
quite playable. I think that after learn­
ing the main lines, these alternatives
would make a useful addition to a
player's repertoire, but we will chiefly
stick to the main line, 11 ...li.Jf6.
12 f3 f4
Here 12 ... 'it>h8?! would make little
sense, as White could proceed with 13
cs immediately.
With the centre closed, the battle
lines are clear - White will play on the
queenside, whereas Black will try to
attack the white king. Here the way
divides:
A: 13 cs gS 14 cxd6 cxd6 1S l:c1
B: 13 cs g s 14 cxd6 cxd6 1 s tDf2
C: 13 cs g5 14 :C1 liJg6 15 %51?
D:Hg41?
Usually White just proceeds with
his own play, and in Lines A and B we
look at White's traditional continua­
tions where we have a typical King's
Indian 'race', with both sides trying to
blow the other away. Line C examines a
radical idea that has become very
popular of late. Finally, Line D looks at a
completely different plan where White
looks to slow Black down on the king­
side before pursuing his queenside
ambitions. The position becomes
blocked and the play greatly slows
down.
A) 13 cs
This is the main line. White imme­
diately initiates queenside play, trust­
ing in his chances in the ensuing race.
He will decimate the black queenside
while Black tries to mate the white
king. Although White's play is usually
faster, Black is playing for higher
stakes.
13 gs
...
There is no need for Black to be sub­
tle at this point. White will now bring a
19
A ttacking Chess: Th e King 's I n dian, Vo l u m e 1

rook to the c-file, but he has a choice as We will see this multi-purpose
to how to go about it. move again and again. Black defends
14 cxd6 the c7-square in advance, and prepares
White immediately opens the c-file. ... 3'.f8 and ... l:tg7. These moves will help
The preparatory 14 l:tcl will be seen in him prepare ... g4. The bishop on f8 is
Line C. often well placed - from there it pro­
14...cxd6 tects the d6-pawn and sometimes the
bishop will come to e7 from where it
could join a kingside attack (with
... 3'.h4 or ...3'.gs), or emerge on the
queenside ( ... 3'.d8).
Note that 16 ... 'ii'b 6+?! doesn't win a
piece; rather it just wastes time be­
cause White is happy to play 17 tllf2
when the bS-knight is suddenly pro­
tected by the e2-bishop.
17 'ikc2

15 !tel
White's main choice, taking the c­
file immediately. Instead 15 tllf2 pre­
pares to bring the king's rook to cl. This
is Line B.

White again threatens tll c 7-e6, so

1s ...tllg6 16 tllbs
The knight threatens to invade on
c7, from where it will go to e6.
16...l:tf7
Black has to retreat his knight, making
the ... g4 advance more difficult to
achieve. Also the tactical shot tllxa7 is
sometimes possible, exploiting the
loose bishop on c8.
17 ...tlle8 18 a4
This move is always useful for
White to include. It gives the bS-knight
extra support and should Black kick the

20
 Th e M a r def Plata Varia tio n : 9 t'De1 t'Dd 7 1 0 t'Dd3
knight back with ... a6, then t'Da3 will be
played when ...bs is prevented, and
White will often thr�aten l'Dc4 and 3'.as
(or a4-a5), with a grip on the b6-square.
If White can play t'Db6, he will be able
to eliminate Black's important light­
squared bishop. Black will often feel
compelled to play ... a6 to drive the
knight away anyway, but usually he
makes some useful moves first
1s hs
...
Black threatens ...g4 again.
19 C'Df2
White prevents ... g4 and increases
the scope of his light-squared bishop. It
is too early for ...a6 because of t'Da3-c4,
followed by as and t'Db6. The e8-knight
has to defend c7, so Black has to make
a choice between two bishop moves.
We have:
A1:19..:..lf8
A2t 19.- .t.d:J'rf
• c-file immediately.
Al) 19 3'.fS
...
This is the traditional move. Black
prepares ... l:tg7, overprotects d6 and
may also improve the bishop by playing
... 3'.e7. For a while 19 ... 3'.d7 supplanted
19 ...3'.f8, but I think the two moves are
of approximately equal value.
20 h3
This prophylactic move is the most
common. Instead 20 b4 is not terribly
useful, but it does set a positional trap.
After 20 ... l:tg7 (Black could consider
20 ... a6) 21 t'Dxa7 Black should avoid
21 ... l:tc7? because of 22 t'Dc6! bxc6 23
dxc6 when White had more than
enough for the piece in N.Rashkovsky­
E.Gufeld, Daugavpils 1978. Instead
Black can play 21 ... 3'.d7! with a good
version of the pawn sacrifice discussed
below.
20 'ikb3 is a tricky move order:
20...l:tg7?! (20 ... 3'.d7 transposes to Line
A2, but Black should consider 20...l'Df6
too) 21 l:tc2! when White can omit the
prophylactic h2-h3 and double on the
White does have an important al­
ternative in 20 t'Dxa7.
This grabs a pawn, but Black has a
counter available which regains the
21
A ttacking C h e s s : Th e King 's I n dian, V o l u m e 1
pawn if he is willing to go into an end-
game. One would think that this would
favour White, but actually Black equal­
izes pretty easily. Still, most black play­
ers have preferred to sacrifice a pawn
here to keep playing for an attack.
Black has:
a) 20.. .lk7 forces White's reply and
mass exchanges follow: 21 .Jtas l::t xc2
22 .Jtxd8 l::t x e2 23 'Llxc8 l:txa4 (worse is
23 .. Axb2 24 as! with hopes of an edge)
24 'Lld3 (24 .Jtxgs l:txb2) 24...g4 2s l:tf2
l::t e3! 26 'Llel l::t a8 27 l:tfc2 l:tb3 28 'iii> f2
l:ta2 29 l::t b1 'iii> f7 30 'ite2 .lte7 31 'Llxe7
'Llxe7 32 'Lld3 l::t a8 33 .Jtxe7 'itxe7 was
completely equal in l.Novikov-1.Glek,
USSR 198S.
b) 20....Jtd7!? leads to a more full­
blooded struggle, but it is of course
riskier: 21 'Libs l:tg7 22 h3 'Llh4 23 'ii b 3
'iii> h 8 sees Black patiently make one
more little preparation for ...g4.
(seefollowing diagram)
The main point of this move is to
get off the a2-g8 diagonal in anticipa-
22
tion of an eventual 'Llc7-e6.
The importance of this detail can be
seen after 23 ... g4?! 24 fxg4 hxg4 2S
hxg4 'Llf6 26 'Lle7! (much better than
26 .Jte1 'Llh S! which gave Black coun­
terplay in L.Ftacnik-A.Sznapik, Baile
Herculane 1982) 26 ... 'Llxg4 27 'Lle6
'Llxf2 28 l;Ixf2 .Jtxe6 29 dxe6 'iii> h 8 30
l:tc3 .lte7 31 l::t h 3 when White was
much better in L.Ftacnik-M.Vokac,
Czech Championship, Frenstat 1982.
After 23 ...'ith8 White has tried both 24
l::t c 3 and 24 as. In both cases Black
should sacrifice a pawn with 24... g4,
when in practice he has been rather
successful. This is all very nice, but I
suspect that 20 ... l::t c7 is sounder.
20 l:tg7
...
Instead 20 ... .Jtd7 21 'ikb3 l:tg7 will
likely transpose to Line A2.
Also possible is 20...'Llh4. Then 21
'Llxa7 .ltd7 (note that 21...l:tc7? is not
good here, because after 22 .Jtas l:txc2
23 .Jtxd8 l::t x e2 24 'Llxc8 l:txa4 2S .Jtxg s !
White wins a pawn and a tempo) 2 2
'Li b s l:tg7 23 'ii b 3 is the pawn sacrifice
in variation 'b' above.
 Th e M a r def P lata Varia tio n : 9 CDe 1 CDd 7 1 0 CDd3
21 'ikb3
White prepares to double rooks on
the c-file. The queen is also useful here
for defence - after mass exchanges on
g4, she may slide along the third rank.
Sometimes White plays l:tc3 (or l:ta3 in
certain positions) to utilize a rook in
this fashion too.
White can also change his mind and
take the pawn after all with 21 ttJxa7,
when 21 ... 3'.d7 22 lLibS is again varia­
tion 'b' above. Sounder is 21 .. Jk7 22
l.as which leads to the same endgame
as before with White having the extra
move h3. This does not necessarily help
him: for example, 22 ...l::t x c2 23 3'.xd8
:xe2 24 ttJxc8 l:txa4 2S ttJd3 g4 26 l::t f2
:e3 27 lLi el g3 was fine for Black in
O.Averkin-G.Kasparov, Moscow 1979.
Another idea is 21 as CDh4
(21 ... 3'.d7!? 22 'ikb3 lLih4 is probably
better) 22 ttJxa7 ! ? l:tc7 (22 ...3'.d7 ! ?) 23
.'Llc6! bxc6 24 dxc6 l:tg7 2 S b4 lLic7 26
bS lLixbs ! ? 27 3'.xbs g4 and now in­
stead of 28 hxg4 CDxf3+! 29 gxf3 't!Vh4,
which was unclear in A.Jakubiec­
K.Chojnacki, Lublin 2009, I suspect that
28 fxg4 hxg4 29 hxg4 3'.xg4 30 ltJxg4
l:txg4 31 l:tf2 would test Black's play.
21...ttJh4 22 l:tc2 a6
Black plays this now that he will be
able to force through ... g4. Alterna­
tively:
a) 22 ... g4?! (the immediate advance
is premature) 23 fxg4 CDf6 24 3'.el hxg4
2S hxg4 ctJhS was played in V.lvanchuk-
1.Cheparinov, Sofia 2008. Here 26 CDdl!
3'.xg4 27 3'.xg4 l:txg4 28 'ikh3 'ikg s 29
ltJc7 f 3 30 3'.xh4 l:txh4 31 'ike6+ 'iii> h8 32
l:txf3 gives White a big advantage ac­
cording to Mikhalevski.
b) 22 ... lLif6 ! ? is possible, though: 23
l:tfcl g4 24 fxg4 hxg4 2 S hxg4 3'.xg4 26
3'.xg4 ttJxg4 27 ttJxg4 l:txg4 has been
play a few times and seems okay for
Black. Then 28 3'.el f3 ! 29 3'.xh4 'ikxh4
30 'ikxf3 l::t xe4 31 l:tc3 (31 g3 'ikg4 32
'ikxg4+ l:txg4 was at least equal for
Black in K.Langeweg-J.Donner, Amster­
dam 1971) 31 ... l::t el+ 32 l:txel 'ikxel+ 33
'ikfl 'ikxfl+ 34 'itxfl a6 3 S lLic7 l::t c 8 36
tLie6 l:txc3 37 bxc3 b6 38 lLic7 as 39
'ite2 3'.e7 1/2-1/2 was J.Knap-V.Tasic, cor­
respondence 2006.
23 ttJa3 lLif6
Another possibility is 23 ...3'.d7,
which actually transposes to the main
line of Line A2.
24 3'.e1
This move sizes up the h4-knight
and clears the second rank for defence.
24...g4 25 hxg4
Instead 2S fxg4 hxg4 26 hxg4 (26
ttJxg4?! ttJxe4) 26 ...ttJxg4 should trans­
pose, but White should certainly avoid
23
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

the temptation of 27 l:txc8?. 32 l::tf3 3'.e7

He hopes to win two pieces for a This position has been reached a
rook, but Black has the trick 27 ...'Llxf2! few of times in practice and Black has
28 3'.xf2 (28 l:txd8 l:txg2 mate) 28 ... l:txc8 enough play. White's knight is far away
29 't!Vh3 'Llxg2 30 3'.g4 'Lle3 31 3'.xe3 and his pawn structure is a little loose.
'ikd7 and 0-1 in V.Talla-M.Szelag, Us­ One game continued 33 l:th3 l:tg8 34
tron 2008. l:txh4+ 'ikxh4 35 l:tc3 'ikel+ 36 'ith2
2s ...hxg4 26 'Llxg4 'Llxg4 l:txg2+ 1/2-1/2, l.Schuett-V.Conti, corre­
The alternative 26 ...'Llhs may be spondence 1999.
playable, but it looks sketchy to me.
Instead of 27 as 'Llg3 28 3'.xg3 fxg3 29 A2) 19...3'.d7
'ikb6 'it e7 30 l:tfc1 3'.xg4 31 fxg4 'Llg6 32
ilf3 'ikh4 33 �1 'Llf4, when Black had
plenty of counterplay in L.Ftacnik­
Z.Polgar, Trencianske Teplice 1985,
White should play 27 'Llc4!. Now
27 ... 'Llg3 28 3'.xg3 fxg3 29 'ikb6 'ike7 30
'Llce3 l:th7 31 l:tfc1 has been played in a
couple of correspondence games. The
position is messy, but I think Black is
better off with the text move.
27 fxg4 3'.xg4 28 3'.xg4 l:txg4 29 'ikh3
This is a good square for defensive This is a more modem move. Black
reasons as well as for a possible coun­ avoids any 'Llxa7 tricks and sometimes
terattack. However, Black can hold the even plays on the queenside himself. Of
balance. course, the main plan is still a kingside
29...'ikgs 30 3'.xh4 l:txh4 31 'ike6+ 'ith8 attack

24
 Th e M a r def Plata Varia t i o n : 9 tDe1 lDd 7 1 0 CDd3
20 'ikb3
White can also play 20 h3 3'.f8 21
'Wb3 which will transpose.
20 3'.fS
...
Black has also tried 20 ...3'.f6 ! ?, which
is not without its logic. Black often
plays ... 3'.f8-e7 and then sometimes
either ... 3'.h4 or ...3'.d8, so it makes
some sense to put the bishop on the
h4-d8 diagonal in one move instead of
two. White has:
a) With 21 3'.b4 White decides to
take advantage of Black's last move to
attack d6 immediately, but he should
probably wait on this, as it runs into
some tactical problems: 21 ... 3'.e7 22
:te2 a6 23 CDa3 g4! 24 fxg4 hxg4 2S
3'.xg4 3'.xg4 26 ttJxg4 'ikb6+ 27 'iii> h l as
28 ctJh6+ 'iii> g7 29 ttJxf7 'ikxb4 with
complications that were not unfavour­
able to Black in V.Neverov-M.Vokac,
Bled Olympiad 2002.
b) 21 l:tc2 'ikb8!? (consistent; Black
wants to play ... 3'.d8 quickly) 22 as 3'.d8
23 l:tfcl (the prophylactic 23 ltJa3! ? is
another idea when 3'.e2-bs is a possi­
bility and there are ideas like CDc4, 3'.b4
and 'ika3, laying siege to the d6-pawn)
23 ... a6 24 CDa3 bS gave Black counter­
play in S.Zilka-R.Ramondino, Vienna
2009.
21J::tc2
This is very natural. White doubles
rooks on the c-file. Others:
a) 21 h3 and now 21 ... l:tg7 will gen­
erally transpose elsewhere, depending
where White moves his rook, but Black
also has the option of making some
manoeuvres on the queenside:
al) 21 ... 'ikb8!? 22 as l:tg7 23 l:tc3?!
(this leaves the as-pawn without sup­
port; 23 l:tc2 is more natural and could
transpose to the next note) 23 ... 3'.e7 24
l:tal (not a nice move to make) 24 ... 3'.d8
2S l:tccl lLih8 ! ? (a typical idea; Black
routes the knight to h6 to support the
... g4-break) 26 ttJa3 a6 27 CDc4 3'.bs 28
'iii> fl CDf7 29 3'.el ltJh6 30 lLib6 3'.xb6 31
axb6 3'.xe2+ 32 'itxe2 'ikd8 33 CDd3 g4
gave Black the initiative in M.Roeder­
M.Hebden, Bern 1992.
a2) 21 ... 3'.e7!? 22 l:tc2 'ikb8 23 l:tfcl
3'.d8 24 as a6 2 s ltJa3 bS 26 3'.e1 (26
3'.b4 l:tg7 27 'iii> fl lLif6 28 'ikd3 lLih8 29
lLibl CDf7 was similar in M.Vukic­
D.Sahovic, Tuzla 1983) 26 ...lLif6 27 'ikd3
l:tg7 28 3'.d1 ctJh8! left ... g4 imminent in
A.Olcayoz-E.Grivas, Mangalia 1992.
b) 21 l:tc3 was played in S.Conquest­
P.Thipsay, British Championship,
Southampton 1986. It guards the third
rank (in the event of exchanges follow­
ing ... g4), and leaves the possibility of
'ikb3-d1 open, but compared to 21 l::t c 2,
White limits his d2-bishop - an a4-aS
25
Attacking C h e s s : The K i n g 's I n dian, Vol u m e 1
advance will not be supported and .Jtb4
is not possible.
c) 21 J::t c4 is a bit strange looking be­
cause there is no 'Lla3-c4 possibility:
21...a6 22 'Lla3 (again, there is no 'Llc4;
perhaps White was hoping to use the
rook on the fourth rank with l:tb4 to
pressure the b7-pawn) 22 ...l:tg7 23 as
'Llf6 24 'ikb6 'ike8 2s h3 (2s 'ikxb7? .Jtbs)
2S ...g4 26 fxg4 hxg4 27 hxg4 'Llh4 28
l::t c7 'ikg6! 29 .Jte1 (or 29 'ikxb7 l::t d 8 30
�xa6 ii.xg4 31 l::t x g7+ .Jtxg7 32 'Llxg4
'Llxg4 33 .Jtxg4 'ikxg4 34 l:tf2 f3 3S �d3
l::t f8 with a decisive attack according to
Gallagher) 29 ...l::th7 30 'ii b3 (30 'ikxb7
'ii h6! 31 'Llh3 'Llxg2 wins) 30 ... 'Llxe4 31
'ikd3? (after 31 'ikxb7 'Llg3! 32 .Jtd3 .Jtfs!
33 l::t xh7 'ikxh7 34 'ikxh7+ .Jtxh7 Black
wins the exchange, but this was proba­
bly the best chance) 31 ...'Llg3 32 l::t xb7
was Z.Kozul-T. Radjabov, Sarajevo 2003.
Here Gallagher points out that
32 ... .Jtfs! 33 l:txh7 (33 gxfs 'Llxe2+ 34
'ikxe2 'ikxg2 mate) 33 ... .Jtxd3 34 .Jtxd3
'itxh7 3S .Jtxg6+ 'Llxg6 gives Black a
clear advantage.
21 a6
...
26
Without White having spent a
tempo on h3, 21 ...'ii b 8?! looks too slow:
22 l::t fcl a6 23 'Lla3 .lte7 24 as .Jtd8 2S
'Llc4 .Jtbs 26 .ltb4 b6? was G.Kacheish­
vili-D.Sharavdorj, Lubbock 2009, and
here White could have played 27 'Llxd6!
'Llxd6 28 .Jtxd6 'ikxd6 29 .Jtxbs axbs 30
l:tc6, winning the knight on g6.
22 'Lla3 l:tg7
22 ...'Llf6 23 h3 l::t g7 is the same.
23 h3 'Llh4
Black has to time his pawn break
carefully: 23 ...'Llf6 24 l:tfc1 g4?!
(24...'Llh4) 2S fxg4 hxg4 26 hxg4 'Llh4
27 l::t c 7! .lte7 28 .Jte1 was much better
for White in L.Ftacnik-G.Ligterink, Am­
sterdam 1977. If 28 ...'Llxg4 then 29
.Jtxg4 .Jtxg4 30 'Llxg4 l::t xg4 31 't!fh3
l::t x g2+ 32 'iii> h l wins for White.
24J::tfc1
24 'Llf6
.•.
Again 24... g4?! is premature: 2S
fxg4 hxg4 26 hxg4 l:tc8 (or 26...'Llf6 27
l:tc7, transposing to 23 ...'Llf6, above) 27
l::t xc8 .Jtxc8 28 'Llc4 'Llf6 29 'ikh3 with a
clear advantage in N.Rashkovsky­
A.Vitolinsh, Daugavpils 1978.
 Th e M a r def Plata Va ria tio n : 9 tDe1 lDd 7 1 0 lDd3
Black also does not need to commit
to 24 ... l:tb8 yet. After 2S CDC4 b6 26 as!
g4 27 fxg4 lLif6 28 ttJxb6 hxg4 29 hxg4
.!Lixg2, as in V.lvanchuk-J.Timman,
Hilversum 1991, White could play 30
'itixg2! ltJxg4 31 3'.xg4 3'.xg4 32 �1
3'.d7 33 'ite2 when he should win.
2S 3'.e1
Alternatively, 2S 'ikxb7? runs into
2s ... ttJxg2!, while 2S l;Ic7 can be met by
2s ... l:tb8 26 as ltJe8 27 l:t7c2 lLif6 28 l:tc7
!lJe8 29 l:t7c2 lLif6 when H.Gruenberg­
P.Hesse, Eilenburg 1984, ended in a
draw.
2s ...g4
Black has made the necessary
preparations and now breaks.
26 fxg4 hxg4 27 hxg4 3'.xg4 28 3'.xg4
!Llxg4 29 ttJxg4 l:txg4 30 'ikh3 'ikgs
This position was agreed drawn in
F.Quiroga-E.Maggiolo, Buenos Aires
2000. White was quite a bit higher­
rated here, which shows the health of
Black's position. The position is very
similar to the one we saw in Line Al,
but here Black has an even better ver­
sion.
B) 13 cs gs 14 cxd6 cxd6 1S ttJf2
Before playing 'ikc2 and �fcl, White
shores up his defence of the g4-square.
1s ...ttJg6
Black could also begin with 1s ...h s
t o try t o force White to play h3. I f 16
'ikc2 (16 h3 ttJg6 17 'ikc2 �f7 18 l:tfcl
transposes the main line), then
16 ... g4! ? (16 ...l:tf7 11 ttJbs ttJg6 18 l:tfc1
ltJe8 19 a4 allows White to delay h3) 17
lLibS g3 is an untried possibility.
16 'ikc2 l:tf7 17 l:tfc1 hs
After 17...ltJe8 18 a4 hS White went
backwards with 19 CDcdl ! ? and won a
long struggle after 19 ... 3'.f8 20 l::t a 3 in
L.Aronian-H.Nakamura, World Team
Championship, Bursa 2010. The text
move avoids this possibility because
then 18 ttJcdl could be met with
18 ... g4!.
Black has also tried 17 ...a6, but after
18 a4 hS 19 h3 3'.f8 20 as we transpose
to 18 ... a6 19 a4 3'.f8 in the note to
Black's next move, which looks unsatis­
factory for him.
18 h3
It is important for White to hold up
27
A ttacking C h e s s : The King 's I n dian, Vo lu m e 1
... g4 as long as he can. He can, though,
also play the immediate 18 'Libs here.
After 18 ...'Lle8 19 a4 3'.f8 or 19 ...3'.d7
the play is similar to the main line and
may even transpose.
18 3'.fS!?
...
Black stays calm. Practice has also
seen him take immediate measures on
either flank:
a) 18 ... a6 stops 'Li bS, but it does
weaken the queenside and White has
other ways into the black position after
19 a4!. Now 19 ...b6 is considered bad
after 20 l:ta3 (with the idea of 'Lla2 and
l:tb3 or l:tac3, while 20 'Lla2 is also pos­
sible) 20...3'.f8 21 'Lla2 ! . The alternatives
are supposed to be better, but it all just
seems rather shaky to me:
al) After 19 ...'Llh4 20 'ikdl g4 21
fxg4 hxg4 22 hxg4 'Llh7 23 as 'Ll gs 24
.i.f1 bS 2S axb6 'ikxb6 26 'Lla4 'ika7 27
3'.as ! l:tb7 28 l:tc6 l:tab8 29 3'.b6! White
was winning in B.Lalic-L.McShane,
Southend 2000.
a2) Instead 19 ...3'. f8 runs into 20 as
(with the idea 'Lla4-b6) 20 ... g4 21 fxg4
hxg4 22 hxg4 bS 23 axb6 'ikxb6 24 'Lla4
28
'ika7 2S 3'.as l:tb8, and now White has a
very pleasant choice between 26 3'.c7
and 26 g s ! 'Llg4 27 3'.xg4 3'.xg4 28 'Li es!
(but less good here is 28 'ikc6 3'.e2 !).
b) 18 ...g4!? is a better try: 19 fxg4
hxg4 20 hxg4
20...'Lle8! (Black intends ...3'.f6-h4; a
useful idea to know) 21 a4 3'.f6 22 l:ta3
3'.h4 23 'Llcdl 3'.g3 24 l:tc3 (24 'Llh3
't!fh4 2s 'Lldf2 'Llf6 26 'ikd1 3'.d7 27 as
l:taf8 28 3'.el f3 29 l:txf3 'Llf4 30 l:txf4
and Yi-1/2 was G.Sosonko-F.Hellers, Wijk
aan Zee 1986) 24 ... 3'.d7 2s 3'.bs?! 3'.xbs
26 axbs 'ikh4 27 'Llh3 'Llf6 28 l:tc8+? (28
gs 'Llg4 29 l:tc8+ l:txc8 30 'ikxc8+ 'Llf8! is
given by Gallagher) 28 ...l:.xc8 29 'ikxc8+
'iii> g7 3o'Lldf2 3'.xf2+ 3 1 'Llxf2 f3 ! 32 'ikfs
'Llxe4 33 'ikxf7+ (33 'ikxe4 'ikg3)
33...'itxf7 34 'Llxe4 'ikxg4 and Black had
a winning attack in G.Andruet­
V.Spasov, Sofia 1990.
19 'Libs 'Li es 20 a4
The thematic 20 'Llxa7?! seems to
work tactically because the c8-bishop is
loose, but it appears too slow: 20 ... 3'.d7
(the counter 20 ...l:tc7? does not work
well here after 21 3'.as l:txc2 22 .i.xd8
 Th e M a r def Plata Va ria tio n : 9 tDe1 lDd 7 1 0 lDd3
.:.Xe2 23 ttJxc8 l:txb2 24 3'.xgs l:taxa2 2 S
.:.Xa2 l:txa2 2 6 l:tb1 and White i s much
better) 21 ttJbs g4! 22 fxg4 hxg4 23
hxg4 f3 ! 24 gxf3 3'.xbs 2s 3'.xbs l:txf3
gives Black excellent play: for example,
26 'ikdl l:txf2 27 'itxf2 'ikb6+ 28 3'.e3
'ikxbs M.Stean-J.Hjartarson, Lucerne
Olympiad 1982, or 26 3'.xe8 't!fh4 27
�gs 'ikg3+ 28 'itf1 l:txe8 29 'ike2 (29
'Wd2 't!Vh 3+ 30 'ite2 'ikg2 31 l:tf1 lLif4+ 32
.i.xf4 exf4 is also winning for Black)
29 ... 3'.e7 0-1, J.Barkhagen-M.Tumer, Sas
van Gent 1992.
20 3'.d7
...
With the bS-square covered, Black
should not ignore White's idea: for ex­
ample, 20...lLih4 21 ttJxa7! 3'.d7 22 ttJbs
g4 23 fxg4 hxg4 24 hxg4 l:tg7 2S 'ikdl
.i. e7 26 l:ta3 was much better for White
in Z.Kozul-A.Sznapik, Tbilisi 1988.
20...l:tg7 is playable: 21 l:ta3 a6 22
:C3 3'.d7 23 ttJa3 lLih4 24 3'.el l:tb8 2 S
'Wdl lLif6 gives Black counterplay.
21 'ikd1
This move tries to clamp down on
the g4-square. Probably White should
look at the alternatives, though. 21 l:ta3
a6! forces the knight back to c3, while
after 21 as!? Black should play 21 ...l:tg7,
which is the same as the Line A2, ex­
cept that White's rook is on al instead
of fl. This probably does not change
much, especially as White usually dou­
bles rooks on the c-file over the next
couple of moves, but White will have
the option of playing a quick 'ikdl or
�el without locking his rook in on fl.
21...l:tg7
Black intends ... lLih4, ... a6 and ...lLif6
to force through ... g4. White should
play 22 as or double rooks with 22 l:ta3
or 22 l:tC3. Instead 22 'itfl?! 'takes ad­
vantage' of the free fl-square, but this
is not a very good place for the king.
After 22 ...lLih4 23 3'.el 3'.e7 24 l:tc3
'ikb8 ! ? (another idea is 24... a6 2 S ttJa3
lLif6 26 l:tacl g4 2 7 hxg4 hxg4 28 fxg4
'ike8!?, hitting the a4-pawn and prepar­
ing ... 'ikg6) 2S as 3'.d8! (here we see one
of the points of 24...'ikb8; from d8 the
bishop can help defend the queenside
and it may even become active itself)
26 l:tb3 a6 27 CDc3 (27 ttJa3 3'.a4!)
27 ...CDC7!? (Black intends to use the bS­
square, but a more thematic plan is
27 ...lLif6, trying to force through ...g4)
28 l:tb4 'ith8 29 l:ta3 'ikc8 30 l:tc4 'ikb8
31 l:tb4 'ikc8 32 l:tab3 ttJbs 33 ttJxbs
axbs 34 3'.xbs 3'.xas 3S 3'.xd7 l:txd7 36
l:ta4 l:tc7! Black suddenly took over the
queenside in A.Gavrilov-A.Shomoev,
Krasnoyarsk 2007.
C) 14 l:tc1
White clearly will need a rook on the
29
A t ta c k i ng Ch ess: Th e King 's Indian, Vo l u m e 1
c-file, so he plays this immediately in­
stead of exchanging on d6.
14 'Llg6 1s 'Libs!?
...
This is a very sharp move which has
become popular recently. Of course lS
cxd6 cxd6 brin gs us back to Line A.
By delaying the exchange on d6
White is able to create some unusual
problems. Black can now force White to
make a positional piece sacrifice or he
can play normally, although this allows
White to play an original manoeuvre to
weaken the black queen side.
1s l:tf7
...
Instead both 1S ... g4? 16 cxd6 cxd6
17 'Llc7 and 1S ...'Lle8? 16 3'.as are
clearly better for White, but practice
has also seen the critical 1S ... a6, trying
to refute White's play.
Here White has:
a) 16 'Lla3 intends 'Llc4 and 3'.as.
However, this move abandon s control
of e4 and Black can play 16 ... g4 imme­
diately. After 17 cxd6 cxd6 18 'Llc4 g3
White has:
al) 19 h3 3'.xh3 20 gxh3 bS (better
than 20 ... 'ikd7 21 'Llf2! gxf2+ 22 l:txf2)
30
2 1 l:te1? ! (better i s 2 1 'Lias 'ikd7 2 2 'Llf2
gxf2+ 23 l:txf2, although Black still has
the initiative after 23 ...'LlhS) 21 ...bxc4
22 l:txc4 'Llh4 was clearly better for
Black in H.Hoeksema-L.Riemersma,
Dutch League 1987.
a2) 19 3'.as gxh2+ (worse is 19 ...'ike7
20 'Llb6 gxh 2+ 21 'itxh2 l:tb8 22 'ikc2! )
20 'itxh2 'ike7 2 1 l:th 1 'Llhs 22 'Llb6 'Llg3
gave Black good play in M.Yudovich ­
E.Arlind, correspondence 1974.
b) 16 cxd6! (this piece sacrifice is
critical) 16 ... axbs 17 dxe7 'it d 7 (worse is
17...'ike8?! 18 'ikb3 g4 19 'Lies 'Llh4 20
fxg4 and White was much better in the
game F.Berkes-M.Pavlovic, Kragujevac
2009) reaches quite a complicated po­
sition.
White's play certainly seems easier,
but Black is up a piece:
bl) 18 'ikb3 'Lle8 with a further di­
vide:
b11) 19 3'.b4 'Lle8 19 'ikb3 'Llxc7 20
3'.xf8 Lf8 looks unclear.
b12) 19 'Li es 'ikd6 20 'Lle6 (after 20
3'.b4 Black can just play 20 ...'ikb6)
20 ... 3'.xe6! (much better than 20 ...'ikb6+
 Th e M a r def Plata Va riatio n : 9 tDe1 lDd 7 1 0 lDd3

21 'iii> h l 3'.xe6 22 dxe6 ttJxc7 23 e7+ l:tf7

24 3'.c4 lLih8 25 l:tfdl when White had a
big advantage in O.Biriukov-S.Soloviov,
Saint Petersburg 1999) 21 dxe6 ttJxc7
22 3'.b4, and now both 22 ... 'ikb6+ and
22 ...'ikd4+ 23 'iii> h l l:tfc8 are possible,
while earlier 21 3'.b4 'ikb6+ 22 3'.c5 'ika5
23 dxe6 ttJxc7 24 3'.xf8 'itxf8 25 l:tfdl
'ikb6+ 26 'iii> h l ttJxe6 27 'ikb4+ 'itg8 is
unclear according to Soloviov.
b2) 18 3'.b4 was Giri's recent try.
Black has: 16 3'.as!
b21) 18 ...ltJe8 19 3'.xf8 (19 'ikb3 is This is the point of White's play - he
'b11' above) 19 ...3'.xf8 with the idea of forces Black to weaken his queenside .
...3'.d6 was given by Nunn, although This is somewhat annoying, but the
Hoeksma points out that White has manoeuvre costs White some time and
some initiative after 20 'ikb3, intending Black is not without practical chances.
l:tc3 and l:tfcl. Instead 16 cxd6 cxd6 17 'ikc2 would
b22) 18...g4?! 19 CDc5 (19 d6! gxf3 20 again take play back into Line A.
gxf3 b6 21 tLif2 'it h8 22 a3 l:tg8 23 'iii> h l
'ike8 24 l:tgl gives White more than
enough compensation according to
Giri) 19 ... 'ikxc7 20 CDe6 and here, in­
stead of 20...'ikf7? 21 3'.xb5! when
White was much better in A.Giri­
F.Nijboer, Haaksbergen 2009, Black
should have tried 20 ... 'ikb6+ 21 3'.c5
'ika5.
b23) 18...l:tf7 19 'ikb3 3'.f8 20 3'.xf8
'itxf8 21 'ikb4+ 'itg8 22 CDf2 CDe8 23
3'.xb5 'ikd6 24 'ikxd6 ttJxd6 25 a4 'itf8 is 16 ... b6
a line given by Giri, who thinks that This weakens the c6-square, but
Black is okay here. there is no choice.
1s ...l::tf1 17 cxd6!
This is a more normal continuation. This is more accurate than 17 cxb6
Black hopes to transpose to the main when 17 ...axb6 may be a playable al­
lines, but White has an independent ternativefor Black.
course. 11...cxd6

31
A tt a c k i ng C h e s s : Th e K i ng 's In dian, Vo l u m e 1
Instead 17 ... bxas? is just bad after
18 dxc7 'ikf8 19 'Lies (also good is 19
'ika4 g4 20 'Lies gxf3 21 3'.xf3 'Llh4 22
'ikxas 'Lle8 23 'Lle6 and Black was
mauled in V.Bhat-D.Aldama, Boca
Raton 2008) 19 ... a6 20 'Lle6 'ikb4 (or 19
'Li es a6 20 'Lle6 'ike7 21 'Llxg7 'itxg7 22
d6 'ike8 23 'Lla3 with a crushing posi­
tion in M.Mchedlishvili-A.Rustamov,
Tashkent 2010) 21 a3 'ikxb2 22 l:tbl
'ika2, as in D.Fridman-M.Szelag, War­
saw (rapid) 2009, and now 23 'Lld6
wins quickly.
18 3'.e1
This is better than 18 3'.b4 3'.f8 19
'Llf2 ?! (Bhat suggests 19 l:tc6 or 19 'ikc2)
19 ... a6 20 'Lla3 hS 21 CDC4 as 22 3'.el
3'.d7 23 a4 l:tg7 24 h3 'Llh8!?, which
gives Black good play. After 2S 'Lla3 'Llf7
26 'Libs l:tc8 2 7 l:txc8 'ikxc8 28 'Lld3 g4
29 3'.h4 'Llgs he scored a notable upset
in V.Bhat-K.Sai, New Delhi 2009.
18 g4!?
...
This is a very principled reply and a
typical reaction to the white knight's
jaunt to bS. Black takes the oppqrtunity
to play this advance immediately be-
32
cause the e4-pawn is not well pro­
tected. Another possibility is 18 ... a6 19
'Llc3 (after 19 'Lla3 !? Black should
probably play 19 ... g4, again exploiting
the fact that the e4-pawn lacks protec­
tion), and here:
a) 19 ... 3'.d7?! 20 'Llb4 bS 21 'Llc6
3'.xc6 22 dxc6 was much better for
White in D.Pergericht-W.Winterstein,
Luxembourg 1987.
b) 19 ...hs?! compares poorly with
the main line: 20 'Llb4 g4 21 'Lla4 l:tb7?
22 3'.xa6 l:txa6 23 'Llxa6 bS 24 3'.as !
'ikxas 2s l:txc8+ 'iii>h 7 26 'Ll4cS l:ta7 27
'ikel 'ikxa2 28 'Llb4 'ikc4 29 b3! 'ikd4+ 30
'iii> hl and 1-0 was G.Meier­
S.Bromberger, Zurich 2009.
c) 19 ... as! is the best move. Black
must cover the b4-square to prevent a
knight invasion.
Here White has tried:
cl) 20 'Libs g4 21 l:tc6 3'.f8 22 3'.f2
l:tb8 23 'ikc2, although Black had coun­
terplay after 23 ... g3 24 3'.xb6 gxh 2+ 2S
'itxh2 l:txb6 26 l:txc8 'ike7 27 l:tc7 'Lld7
28 'Llf2 'Llh4 in D.Cummings-B.Sam­
buev, Toronto 2010.
 Th e M a r def P lata Varia t i o n : 9 tb e 1 tll d 7 1 0 tll d3

c2) 20 tllf2 3'.f8 21 tllb 5 h5 22 l:tc6

l:tg7 23 a4 3'.d7 24 'ith1! ? tllh 8 25 l:tc3
g4 26 l:tgl (winning the exchange with
26 tllc7 leads to an unclear position
after 26 ... 3'.xa4 27 'ikcl 3'.d7 28 tllxa8
'ikxa8) 26 ... g3 27 tllh 3 l;Ic8 28 'ikc2 gxh2
29 'itxh2 tll g 4+! 30 fxg4 l:txc3 31 'ikxc3
hxg4 32 g3 gxh3 33 tllc7 was G.Meier­
J.Lopez Martinez, Pamplona 2009. Now
Black's best looks like 33 ...tll g6, with the
idea 34 tlle 6 3'.xe6 35 dxe6 d5! when
Black has the initiative in a very com­ Black is a pawn down and White
plicated position. has made some inroads into his queen­
c3) 20 3'.f2 l:tb8 21 tllb 5 3'.f8 22 l:tc6 side, but he has very good counterplay.
tll e8 23 a4 3'.d7 24 l:tc3 l:tb7 25 b4 h5 26 2 3 l:tc3
bxa5 bxa5 27 tlla7 tllf6 28 tllc 6 'ika8 29 To stop ...3'.e3+.
'ikc2 g4 30 tllb 2 l:tg7 31 tllc4 g3 32 hxg3 2 3 ...3'.f4
fxg3 33 3'.xg3 h4 34 3'.h2 tllh 5 was There are other tempting possibili­
G.Meier-R.Polzin, Austrian League ties as well, such as 23 .. 11g7 ! ? and
2010. White is probably quite a bit bet­ 23 ... tllh 5 ! ?.
ter here, but as usual there is counter­ 24 3'.xf4
play and Black went on to score a big Not 24 3'.f2? 'ikh6! 25 g3 (25 l:te1 'ikh2+
upset against one of the main propo­ 26 'iii> f l 'ikhl+ 27 3'.gl tllh 4 mates)
nents of White's set-up. 25 ...'ikh3, which is winning for Black
19 tll b4 24 ...'Llxf4 25 g3
This is critical. White pounces on
the weak c6-square. Instead 19 fxg4
tll x e4 is unclear. 19 l:tc6 is also worth
investigation, but after 19 ... 3'.f8 at least
there is no knight coming into c6.
19...g3!
A thematic and strong pawn sacri­
fice. We will now follow the game
R.Rusev-A.Diamant, Malakoff 2010:
20 hxg3
20 tllc 6 just gives Black the extra
option of 20...gxh2+. 2 s...'Llh3+
20...fxg3 21 tll c 6 'ikf8 22 3'.xg3 3'.h6! Also good is 25 ...l;Ig7!? 26 'iii> f2 'Ll4h5

33
A ttacking C h e s s : The King 's I n dian, Vol u m e 1
(or 26 ... 'Llh 3+) 27 'Itel (27 g4? 'Llxe4+)
27 ... 'Llxg3 28 l:tgl 'Llgxe4!? 29 l::t xg7+
'ikxg7 30 fxe4 'Llxe4 31 'ikc2 'ikgl+ 32
3'.fl 'Llxc3 33 'Llxc3 3'.h3 (33...'ike3+ i s
also interesting) 3 4 'ikd3 3'.g4!? which
seems promising for Black.
26 'iii>g2 l:tg7 27 l:th1??
Under pressure White blunders, but
after 27 'ikc2 'Llh s Black has strong
pressure.
27 ...'Llxe4! 28 'Lle7+
As 28 fxe4 'ikf2 is mate, White's po­
sition falls apart.
28 ... l::txe7 29 l:txc8 'Llf4+!
Black does not relax, even in a win­
ning position. 29 ...l:txc8 30 l:txh3 a6 is
still much better for him, but the text is
much stronger.
30 gxf4 l:tg7+ 31 '>W1 l:txc8 32 fxe4
'ikxf4+ 33 'Itel l:tcl 34 'Llxd6 l:txdl+ 35
'itxdl l:tc7 36 l:tgl+ 'ith8 3 7 3'.c4 'ikf3+
38 'Itel l;Ixc4 0-1
In conclusion, lS 'Libs is rather dan­
gerous, but White must take risks too
and not everyone wants to commit to
sacrificing a piece with the white
pieces right out of the opening. If prac­
tice does eventually show this line to be
too dangerous for Black, it will negate
several decades of opening theory!
That said, although we must be pre­
pared for this line, we will not worry
too much just yet. The important thing
is to be aware of the possibility of l S
'Li b s and t o have something i n mind.
D) 13 g4!?
This is a completely different ap-
34
proach. Instead of engaging i n a
straight race, White takes a stance on
the kingside first. If White can com­
pletely block the kingside, he will have
a free hand on the other side of the
board, where he holds a spatial advan­
tage. This plan was very popular for
White in the late 1980s and prompted
Black to look at flexible alternatives to
11 ... 'Llf6 12 f3 f4, such as 11 ... 'ith8.
Nowadays this plan is not considered
to be so alarming to Black, although
the play can become a bit stodgy.
13 ... gs
After 13 .. .fxg3?! 14 hxg3 Black does
not even have an advantage on the
kingside, so he leaves the position
closed and plans to open the h-file. The
immediate 13 ... hs?! is bad because af­
ter 14 g s 'Llh7 lS h4 Black cannot open
the kingside unless he makes an un­
sound piece sacrifice on gs.
By playing 13 ...g s Black prevents h4
and will open the kingside himself with
... h s. He also prepares ... 'Llg6, aiming
for the h4-square. White usually plays
'it g2, so he can contest the h-file, and a
 Th e M a r def Plata Varia tio n : 9 tDe 1 CDd l 1 0 lDd3
quick 3'.e1-f2, when the bishop covers
the h4-square and also takes aim at the
black queenside. Admittedly, Black has
a narrow field to play with, so he needs
to hold things together on the queen­
side while gradually building up on the
kingside, looking for tactical opportu­
nities. This often involves sacrifices on
g4 or e4.
14 3'.e1
Both sides have a lot of flexibility
here. Instead 14 'itg2 hS lS h3 ttJg6 16
3'.el transposes to note 'b' to White's
16th move, below. White can also play
on the queenside immediately:
a) 14 cs hs 1s h3 'iii> f7 (1s ... ttJg6 16
3'.el leads back to the main line) 16
l:tc1 l:th8 17 ttJbs hxg4 18 hxg4 l:th3 19
'itf2 3'.xg4! 20 cxd6 cxd6 21 fxg4 ltJxe4+
22 'Itel 'ikb6 gave Black excellent play
in G.Tallaksen-E.Lie, Gausdal 2007.
b) 14 b4 hS lS h3 'iii>f7 (1S ...ltJg6 16
3'.el l:tf7 is another way) 16 3'.el l:th8
17 'it g2 ltJg6 18 cs hxg4 19 hxg4 lLihS?!
(an amazing offer, which Gelfand de­
clines) 20 l:th1 (20 gxh s l:txhs 21 l:th1
l:txh l 22 'itxhl 'ilih8+ 23 'itg1 tests
Black's idea) 20 ... ttJg3 21 3'.xg3 fxg3 22
'ikd2 l:th4 23 'itxg3 3'.d7 24 l:txh4 gxh4+
2S 'iii> h 2 3'.f6 gave Black dark-squared
compensation for the pawn in
B.Gelfand-G.Kasparov, Reggio Emilia
1991. Kasparov's sacrifice may not have
been completely sound (it is easier to
say this almost 20 years later!), but the
idea is still noteworthy.
14... hs 1s h3 ttJg6
16 cs
White begins his queenside play.
Some alternatives:
a) 16 3'.f2 l:tf7 17 a4 3'.f8 18 'itg2
3'.e7 19 as 'ikf8! 20 ttJbs 3'.d8 21 ttJxa7
hxg4 22 hxg4 3'.xg4! was a typical shot
in H.Fioramonti-0.Cvitan, Geneva 199S .
b) 16 'itg2 l:tf7 (I prefer this regroup­
ing to 16 ...'itf7 because Black keeps the
option of a quick ...'ikf8 available) 17
3'.f2 3'.f8 and here:
bl) 18 b4 l:th7 19 l:th1 3'.e7 20 'ikb3
'itg7 21 cs 3'.d7 22 l:tacl a6 (Black
avoids 22 ... 'ikh8 23 c6! bxc6 24 dxc6
3'.xc6 2S 'ikc4) 23 a4 hxg4 24 hxg4
lLih4+ 2S 3'.xh4 l:txh4 26 l:txh4 gxh4 27
'ikc4 h3+ 28 'iii> hl (28 'itxh3 't!fh8+ 29
35
A ttacking C h e s s : Th e King 's I n dian, Vol u m e 1
'itg2 'ikh4 gives Black a strong attack)
28 ... 't!fh8 29 c6 bxc6 30 dxc6 3'.c8 31
l:tg1 'Llh7 32 'ikb3 'Llgs 33 'Lids 3'.e6 34
bs axbs 3S axbs 3'.xds 36 'ikxds 'ikb8
was good for Black in P.Lukacs-E.Grivas,
Budapest 1993.
b2) 18 l:thl l:th7 19 'ikh3 (19 cs 'Llh4+
20 3'.xh4 gxh4 21 b4 'ike8 22 'ith2 hxg4
23 fxg4!? l:tg7 24 3'.f3 'Llh7 2S l:tcl 3'.d7
26 l:tfl as gave Black good counterplay
in S.Knott-M.Hebden, British Champion­
ship, Scarborough 2004) 19 ...'Llh4+
(Hebden is fond of this knight hop; the
kingside closes up a bit, but White may
miss his dark-squared bishop) 20 3'.xh4
gxh4 21 l:tagl hxg4 22 hxg4 h3+ 23
l:txh3 l;Ixh3 24 'itxh3 'Llhs! ? 2s 3'.d1 3'.e7
26 'itg2 'Llg3 27 'Lle2 'Llxe2 28 3'.xe2 3'.h4
29 l:th1 'ikg s 30 'Llf2 'itf8 31 l:th3 'iii> e7 32
'ikdl 3'.d7 33 'ikhl l:th8 gave Black
enough compensation to draw in
S.Knott-M.Hebden, British League 2009.
16...l:tf7
17 b4
White maintains the tension on the
queenside. He may also exchange on
d6 immediately with 17 cxd6 to avoid
36
the possibility of Black recapturing on
d6 with his bishop. The downside to
this early exchange is that White gives
up some of his queenside space. He
also loses flexibility, since, for example,
cS-c6 is no longer an option. After
17 ... cxd6 18 l:tcl 3'.f8 19 3'.f2 l:th7 20
'ikb3 hxg4 21 hxg4 'Llh4 22 3'.xh4 l:txh4
23 'Llf2 a6 24 a4 l:th7 2S 'itg2 3'.d7 26
'Llb1 bS! 27 axbs axbs 28 'Lla3 'ikb6 29
'LlxbS (after 29 'Llc2 Black would play
29 ... 'iii> g7, with the idea of ... 3'.e7 and
... l:tah8 with an attack) 29 ... 3'.xbs 30
'ikxbs 'ikxbs 31 3'.xbs l;Ib8 32 3'.c6 l:txb2
33 l:tal, as in A.Khalifman-V.Spasov,
Manila lnterzonal 1990, and now
33 ... 'Lld7 34 l:tfb1 l:txb1 3 s l:txb1 'Li es 36
l:tal 'Llb3 with the idea ... 'Lld4 equalizes
according to Khalifman.
11 ...3'.fS 18 'itg2 3'.e7 19 3'.f2 'ikf8!
This is a nice set-up for Black. The
d6-pawn remains well protected, from
f8 the queen can go to the h-file and,
more importantly, there is latent pres­
sure along the f-file, which may create
opportunities for a sacrificial break­
through on g4 or e4.
 Th e M a r def Plata Varia tio n : 9 t'De 1 CDd 7 1 0 CDd3
20 t'D bs .ilds!
This is another typical idea. The
bishop efficiently defends the c7-
square and an exchange of pawns on
d6 will also open a route for Black's
bishop to the queenside.
21 a4 3'.d7
Black has a harmonious position :
a) Grabbing a pawn with 22 cxd6
cxd6 23 t'Dxa7 fails to the shot 23 ... hxg4
24 hxg4 3'.xg4!.
b) Instead 22 t'Dc3 3'.e7 23 t'D bs 3'.d8
24 t'Dc3 repeats moves, but Black chose
to play on with 24... a6 in Z.Gyimesi-
0.Cvitan, Ticino 1994.
c) A good example of Black's possi­
bilities is 22 l:th l as 23 bxas l:txas 24
cxd6 cxd6 25 t'D b2? (White wants to
play l'D c4, but he overlooks a tactic)
2S ... hxg4 26 hxg4 t'Dxg4! (White
thought that everything was covered,
but it turns out that the bS-knight is
not sufficiently supported) 27 fxg4 f3+
28 3'.xf3 3'.xbs 29 3'.g3 (the white
queen is overloaded, since 29 axbs
l:txal 30 'ikxal l:txf3 wins for Black)
29 ...l'D h4+! (now this check is well
timed; the position has opened up and
Black comes in on the dark squares) 30
3'.xh4 gxh4 31 l:th3 3'.d7 (31 ... 3'.a6! pre­
vents l'D c4 and looks even stronger) 32
l'Dc4 l:tcs 33 t'De3 l:tg 7 34 'ith1 'ikf4
when Black was still better and went
on to win in l.Bedgarini-P.Popovic,
Moscow Olympiad 1994.
37
 Chapter 2
The Mar del Plata Variation
9 lbe1 ltJd7 10 f3
1 d4 'Llf6 2 c4 g6 3 'Llc3 3'.g7 4 e4 d6 5
'Llf3 0-0 6 3'.e2 es 7 0-0 'Llc6 8 ds 'Lle7 9
'Lle1 'Lld7 10 f3
This move can just be used for
transpositional purposes, but in this
chapter we will look at an independent
idea.
10...fs 11 g4
Instead 11 'Lld3 transposes to Chap­
ter 1, while 11 3'.e3 transposes to Chap­
ter 3.
This idea was pioneered by the
Hungarian Grandmasters Benko, Pinter
and Lukacs. As in Line D seen in Chap-
38
ter 1, White fights for space on the
king side.
There i s a big difference here, how­
ever, because Black has not resolved
the tension with ...f4. Although in gen­
eral Black should be happy about this,
he must not forget that White may be
able to play on the kingside too. The
famous game Pinter-Nunn did a lot to
dampen enthusiasm for this line from
White's point of view, but what is old is
new and recently there has been some
revived interest in this variation.
11...'ithS
 Th e M a r def Plata Variatio n : 9 tb e 1 tll d 7 1 0 f3
This flexible move is now consid­
ered the main line and will be the only
continuation we examine in detail.
11 .. .f4 is considered to be a mistake
because White can keep the kingside
closed with 12 h4, although after
12 ... cs Black's position is very solid.
Still, because Black cannot open the
kingside, he is basically just trying to
hold the queenside and this way of
playing is too passive. 11 ...tllf6 is the
other ma in continuation. After 12 tlld3
c6 Black increases the tension across
the board. Then 13 .te3 'iii> h8 is similar
to the main lines, except that both
sides have committed themselves -
White with tlld3 and .te3, and Black
with ... tllf6.
It is logical for Black to keep the ten­
sion and the move 11 ... 'ith8 is almost
always useful to him. The main idea is
to improve the placement of Black's
worst piece - the knight on e7. It will
go to g8 from where it can re-emerge
on f6 to pressure the e4- and g4-pawns.
Moving the knight also opens up the
h4-d8 diagonal and 11 ...'ith8 can be
seen too as preparation for an eventual
...f4. This advance will make some
sense if White can be prevented from
blocking the kingside. 11 ...'ith8 helps
because the retreat ...tllg 8 will allow
Black to fight for the h4-d8 diagonal
with both his queen and dark-squared
bishop. Note that the other knight re­
mains on d7 for the moment, not only
to leave the f6-square free for Black's
bishop in some positions, but because
Black may want to play ... as and ... tll c s,
increasing the pressure on the e4-
pawn.
Because 11 g4 is becoming popular
again, rather than just examine the
aforementioned model game Pinter­
Nunn, we will look at all of White's pos­
sibilities in some detail.
A: 12 .te3
B: 12 liJd3
C: 12 tllg2
D: 12 h4
Instead 12 .td2 tllg 8 13 'ikcl pre­
vents any ideas of ... .th6, but is a bit
artificial. After 13 ...f4 Black is ready to
begin his kingside play. Meanwhile 12
gs tll g 8 13 h4 transposes to Line D, as
does 12 'itg2 tll g 8 13 gs f4 14 h4. Fi­
nally, 12 'iii> h l tllg 8 13 l:tgl has been
played several times by the Hungarian
IM Kiss. Black can play 13 ...f4, when 14
tlld3 gs transposes to Line B, or he can
try 13 ... as!?.
A) 12 .te3
39
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

llld 3 l:Ih7 18 .l:thl .Jth4 19 .Jtd1

Statistically this has been the most

popular move, but it is not seen much 19 ...lllfB
nowadays. The bishop goes to an active Often in the Mar del Plata Variation
square, but now ...f4 will come with we see Black play ... llld7-f6 and ... llle7-
tempo. g6. Here Black has already played
12 ...lllgB 13 'ii'd 2 ... lllg 8, so he finds another way to bring
Instead 13 lll d 3 is considered in the a knight to g6.
notes to White's 13th move in Line B. 20 cs .Jtxf2 21 'ii'xf2 gs 22 cxd6 cxd6 23
There is an alternative in 13 lll g 2 f4 l:Ic1 .Jid7 24 .Jta4 .Jtxa4 2s lll xa4 lllg6
14 .Jtf2 hS (Black could also try 14 ... g s Black certainly had no problems
with the idea 1 s h4 h S!) 1s gs (after 1 s here in D.Jacimovic-M.Vukic, Kastel
h3 .Jtf6 16 b4 l:.f7 1 7 cs l:Ih7 18 'ii'd3 Stari 1988.
lllh 6 19 cxd6 cxd6 20 lllb s lllf7 21 h4 a6
22 lllc3 gs Black tore open the kingside B) 12 llld 3
in S.Mirovshchikov-S. Soloviov, St Pe­
tersburg 2002) 1s ...'ii'x gs 16 .Jih4 'ii'h6
11 lllbs gs 18 .Jtf2 llldf6 19 lllxa .Jth3
20 lllxa8 .i::tx a8 was A.Lesiege­
J.Fedorowicz, New York 1993. White
has won the exchange, but Black's
kingside play is still very dangerous.
13 ...f4!
This advance is usually called for
when White cannot play h4. Here Black
has not committed to ... gs yet, so he
can still use the h4-d8 diagonal. 12 ...lllgB
14 .Jtf2 hS 1S h3 l:.f7 16 @g2 .Jtf6! 17 Black continues with the standard

40
 The M a r def Plata Varia t i o n : 9 t:D e 1 t:D d 7 1 0 f3
plan. Another idea is to block the
queenside and activate the d7-knight
by going to cs. After 12 ... as 13 j_e3 b6
14 'ii'd 2 (14 a3 t:Dcs 1s b4? walks into
1s ... axb4 16 axb4 l:Ixa1 17 'ii'xa1 t:Dxd3
18 j_xd3 fxg4) 14 ... t:Dcs 1s t:Dxcs bxcs
16 @h1 j_d7 17 .i:Ig1 f4 18 j_f2 t:Dc8 19
h4 j_f6 20 'ii'e 1 t:Db6 Black was very
solid in E.L'Ami-S. Shyam, Dieren 2009.
Black can also play the immediate
12 .. .f4. Because White's 12th move did
nothing to influence the h4-square,
Black immediately closes the kingside,
after which the e7-knight will often go
to g6. Now 13 h4 t:Dg8 14 'ii'e 1 j_f6 1 S
g s j_xg s (Black could also try 1s ...j_e7!?
with the idea of ...h6) 16 hxgs 'ii'x gS+ is
a draw. More interesting is 13 l:If2!?
j_f6 14 l:Ig2 j_h4 1S b4, which was
played in P.Eljanov-1.Cheparinov, Jer­
muk 2009, and here Golubev suggests
1s ... gs 16 cs t:Df6, with the idea of ... h s
and ...t:Dg6.
13 @h1
White's next couple of moves indi­
cate that he may have ambitions on
the kingside himself. After 13 j_e3 j_h6
(13 .. .f4 14 j_f2 hS is also possible and
then 1S h3 j_f6 is similar to Line A)
Black has scored quite well. Compared
to 12 j_e3 t:Dg8 13 t:Dg2, White does not
control the h4-square and after 14 j_f2
j_gS Black has improved the scope of
his bishop.
13 ...f4
Again the plan with 13 ... as is a solid
alternative. After 14 l:Ig1 f 4 1S j_d2
j_f6 16 a3 b6 17 b4 j_h4 the game
K. S akaev-V.Bologan, European Club
Cup, Ohrid 2009, was soon agreed
drawn.
14 �g1
14...gs!?
White's kingside manoeuvres indi­
cate that Black should switch plans.
After 14...j_f6 1S b4 j_h4 16 cs t:Ddf6 17
j_b2 Black's pieces do not coordinate
well and White can play on all parts of
the board. Then 11 ... j_d7 18 a4 h S?! 19
t:Dxf4! exf4 20 g s was V.Belov­
A.Grischuk, Russian Team Champion­
ship 2009, where White won back the
piece and kept the initiative.
15 j_d2 hs 16 h3 l:If6 17 I:tc1 l:Ih6 18
41
A ttacking C h e s s : Th e King 's I n d i an, Vo l u m e 1
@g2 .Jtf8 19 b4 tfJe7 20 cs tfJg6 21 cxd6
.Jtxd6!?
Perhaps Black was concerned about
21 ... cxd6 22 tfJbs, although this sortie
does not achieve much after 22 ... hxg4
23 hxg4 CiJh4+ 24 @f1 tfJf6 2S CiJC7 .U.b8
because White's remaining pieces are
not very active. Still, the capture with
the bishop is quite viable.
22 tfJbs tfJf6 23 tfJf2 .Jtd7 24 a4
This was 1.Cheparinov-A.Fedorov,
Khanty Mansiysk 20os, and in this bal­
anced yet dynamic position a draw was
agreed.
C) 12 tfJg2
White shores up his kingside, but
this move is somewhat passive.
12 ...tfJgB
Black can also play 12 ... as 13 h4 tfJcs
14 .Jte3 tfJg8 and in fact this was the
actual move order of Pinter-Nunn.
However 12 ...tfJg8 is both more com­
mon and more flexible, so we will con­
sider it the main line.
13 h4
The alternative 13 .Jte3 was consid-
42
ered in the notes to White's 13th move
in Line A.
13 ... as
Because White is making prophylac­
tic kingside measures, Black goes for
the queenside plan. It will not be easy
for White to keep control of Black's
counterplay across the whole board.
14 .Jte3 tfJcs
15 .U.b1
This does not give White anything,
but other moves have also failed to
trouble Black:
a) 1s gs f4 16 .Jtf2 h6 opens the
kingside before White is ready for it.
b) 1S Itel 'ii e7 16 a3 fxg4 17 fxg4
I:txf1+ 18 @xf1 (it is unappealing to
capture this way, but it is usually nec­
essary, as here, to avoid ...tfJf6 forking
White's e4- and g4-pawns) 18...tfJf6 19
.Jtf3 .Jtd7 20 b4 axb4 21 axb4 tfJa4 was
pleasant for Black in T.Markowski­
S .Dolmatov, Polanica Zdroj 1993. White
is beginning to look overextended.
c) 1s 'ii d2 b6 16 exfs gxf s 17 gs f4
(this gives up the e4-square, but Black
will get many dark squares in return)
 Th e M a r def Plata Va riati o n : 9 lb e 1 lb d 7 1 0 f3
18 .Jtxcs bxcs 19 lbe4 lbe7 20 .Jtd3 lbfs
gave good play in A.Chernuschevich­
J.Rowson, Bratislava 1993.
d) 1S 'ii' c 2 .Jtd7 16 .U.ad1 b6 17 .Jtf2
'ii'c8 18 g s h6 19 @h2 'ii'd8 (Black's little
queen moves annoy White; first the g4-
pawn is attacked and then when it ad­
vances, the black queen takes aim at it
again) 20 exfs .Jtxfs 21 'ii'd2 @h7 22
.Jte3 'ii'd7 23 .U.g1 hS was very comfort­
able for Black in B.Rumiancevas-
1. Schutt, correspondence 1996.
e) 1S a3 fxg4 (Black could also try
1s ... a4 16 .Jtxcs dxcs 17 lbxa4 fxg4 18
fxg4 I:txf1+ 19 @xf1 lbh6!? 20 lbc3 lbf7
21 @g1 lbd6 when White's position
looks a bit airy) 16 fxg4 l:txf1+ 17 @xf1
lt:lf6 18 .Jtf3 h S!? (a notable idea; Black
grabs control of g4 for his pieces) 19
gxh s gxh s 20 b4 (20 .Jtxhs lbcxe4)
20... axb4 21 axb4 I:txa1 22 'ii'xa1 lba6
23 bS lbcs 24 'ii'd1 lbg4 1/2-1/2 A.Gipslis­
D.Lapienis, Parnu 1982. White was
much higher-rated here and he surely
realized his position was beginning to
look rather overextended.
1s... .Jtd1 16 b3 b6
17 a3?!
This is a mistake, but it is not clear
how White should continue. For exam­
ple, after 17 'ii'd2 fxg4 (the immediate
17 ... lbf6!? may be even stronger) 18
fxg4 lbf6 White is forced to part with
his dark-squared bishop with 19 .Jtxcs,
when Black can play 19 ...bxcs or even
19 ... dxcs with the idea ... lbe8-d6.
11 ... a4!
This is a useful device to be ac­
quainted with - not only in this varia­
tion, but in the King's Indian in general.
White's knight on g2 is a long way from
controlling d4 and Black takes full ad­
vantage.
18 b4 lbb3 19 lbbs
This prevents ... lbd4, but it leaves
White's pawn structure looking vul­
nerable.
19...lbf6
Also good is 19 _.fxe4 20 fxe4 I:txf1+
21 @xf1 (else 21 ...lbf6) 21 ...lbf6 22 .Jtf3
'ii' c 8! 23 g s lbQ4 24 .Jt g 1 (to cover h2)
24. ..h s and Black's firmly entrenched
knights give him the advantage.
20 exfs gxfs
43
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
This is exactly what Black wants.
The position begins to open up and
White's kingside pawn advances just
look weakening.
21 lllc 3
Instead 21 gs f4! 22 gxf6 j_xf6 23
j_f2 l:Ig8 24 @h2 l:Ixg2+! 2 s @xg2 j_xh4
26 @h1 'ii'f 6! gives Black a winning at­
tack.
21...e4!
This is another typical King's Indian
idea. Black sacrifices a pawn for control
of the dark squares.
22 gs
22 fxe4 lllxg4 is much better for
Black.
22 ...lllh s
This unleashes an attack on the c3-
knight.
23 fxe4 f4
This further pawn sacrifice gives
Black a good position, but 23 ...lllg 3 24
I:tf3 fxe4 may have been even stronger.
24 j_d2 lllxd2 25 'ii'xd2 'ii'e 8 26 j_f3
lllg 3 27 I:tfel j_e5
For the pawn Black has the bishop­
pair and a grip on the dark squares. The
44
rest of the moves of the classic game
J.Pinter-J.Nunn, Thessaloniki Olympiad
1988, were: 28 llle2 lllxe4 29 j_xe4 f3
30 lllef4 fxg2 31 lllx g2 'ii'h s 32 'ii'd 3
j_g4 33 l:Ie3 'ii'f7 34 'ii'd2 'ii'g 7 35 .U.d3
I:tf7 36 l:Ie1 I:taf8 37 llle3 I:tf4 38 lllg 2
l:I4f7 39 llle3 j_h5 40 I:tf1 I:txf1+ 41
lllxf1 I:tf4 42 'ii'e 1 j_d4+ 43 @g2 'ii'e 5 44
lllg 3 j_g4 45 bS j_f2 46 'ii'xf2 j_h3+ 0-1.
D) 12 h4
This committal continuation has re­
cently become popular. White is not
simply trying to block the kingside;
rather he has his own ambitions on
that side of the board.
12 ... lllgB
Black continues his plan. Again try­
ing to hold up the queenside with
12 ... as is also possible. After 13 llld3 b6
14 @g2 lllcs 15 I:th1, rather than the
1s ... j_d7 16 j_e3 lllg 8 of M.Rodshtein­
V.Bologan, Moscow 2008, when 17 h S!
gives White a big advantage, Markos
suggests 1s ... h S!?.
With White making pawn moves on
the kingside, creating more tension
 Th e M a r def Plata Va ri a t i o n : 9 t:D e 1 t:D d 7 1 0 f3
with 12 ... c6!? is also logical. 13 @g2 (13
.!Z)g2 t:Df6 14 a4 stops any ...bs ideas,
but the position became rather sterile
after 14 ... as 1s l::t f2 j_d7 16 j_e3 l::t c 8 17
l:Ia3 cs in V.Laznicka-H.Nakamura,
European Club Cup, Ohrid 2009)
13 ...t:Df6 14 t:Dd3 bS gave rise to im­
mense complications in P.Eljanov­
T.Radjabov, Elista 2008.
13 gs
White initiates kingside play him­
self. Instead 13 t:Dg2 transposes to Line
c.
13.. .f4
This cuts off White's kingside
pawns. 13 ...h6 has also been tried, but I
do not like the look of 14 exfs gxfs 1s
f4 when Black's minor pieces are
cramped.
14 @g2
14...h6
Black nibbles away at White's far­
advanced pawns. Instead 14...l::t f7 1S
.!Z)d3 (or 1S l::t h 1 j_f8 16 t:Dd3) 1s ... j_f8
16 l:Ih1 h6 transposes to the main line.
15 l::t h l l::tf7 16 t:Dd3 j_f8 17 'ii'gll
This is a strong manoeuvre. From g1
the queen may go t o the h-file and she
may support the gs-pawn. With this
move White can also assist his queen­
side play - the queen supports a possi­
ble cs-advance and even attacks the
a7-pawn. White has done well from
this position in practice, but that is
mainly because Black has been ill­
prepared for a thematic sacrifice. De­
spite the closed nature of the position,
Black should play concretely, paying
attention to White's possibilities.
11 ...l::t h 7
Black readies himself for the open­
ing of the h-file. The downside to this
move is that the f4-pawn lacks support.
It is indeed protected already, but often
White will sacrifice a piece on f4, when
his pawn chain can become rather
ominous and smother the cramped
black minor pieces. There are some al­
ternatives, but they do not change the
essence of the position. Black must be
alert to White's sacrificial ideas.
A typical example of what to avoid
went 17 ...hxgs 18 hxgs+ l:Ih7 19 l::t x h7+
@xh7 20 �2 j_e7 21 t:Dxf4! (Black must
always be wary of this type of sacri­
fice!) 2 1 ...exf4 22 j_xf4 .!Z)es?! 23 j_e3
t:Df7 24 f4 with a winning position for
White in L.Polugaevsky-V.Arbakov,
U SS R 1986.
Instead after 17...@g7 18 @fl j_e7
19 j_d2 White is willing to sacrifice a
pawn because he will control the open
lines on the kingside and be able to
play on the queenside as well. Here
19 ... hxgs 20 hxgs j_xgs 21 .!Z)bs and
45
A ttacking C h e s s : Th e King 's I n d i an, Vo l u m e 1
then both 21 ...lbf8 22 cs a6 23 lba3
�6? 24 lbc4 Ith7 2S .Jtc3 .Jih3+ 26 @e1
.U.c8 27 @d2 lbd7 28 lbf2 and 1-0 in
J.Carron-J.Rosenthal, Winterthur 2008,
and 21 ... .Jth6 22 cs a6 23 lba3 lbdf6 24
lbc4 lbe8 2s lbcxes! dxes 26 lbxes I:tf6
27 .Jtc3, 1 .Cheparinov-D. Stellwagen,
Amsterdam 20os, are additional ex­
amples of what Black should avoid.
18 @f1
This is actually a critical position.
Black must be careful or a piece sacri­
fice on f4 could blow him away.
18 ....Jte7!
Black should keep the tension
rather than open lines towards his own
king. 18 ...hxg s 19 hxgs I:txh1 20 'ii'x h1+
@g7 21 'ii'h4 .Jte7 22 lbxf4! exf4 23
.Jtxf4 lbes 24 .Jte3 has been seen sev­
eral times with terrible results for
Black The position is very similar to
Polugaevsky-Arbakov, above.
19 lbxf4
If White does not play this, then
Black will just take on gs, which will
not only gain material, but will protect
46
the f4-pawn as well. After 19 gxh6
I:txh6 the manoeuvre ... ltJgf6-hS is in
the air and Black can pick up the h4-
pawn at his leisure.
19...exf4 20 .Jtxf4 Itf7!
20...lbes? 21 .Jte3 hxgs 22 hxgs
.Jih3+ 23 @e1 was another disaster for
Black in P.Kiss-M.Ricci, Eger 1994.
21 .Jte3 hs!
This move not only keeps the king­
side closed, it secures the g4-square,
which in tum helps Black secure the
es-square. Similar is the immediate
21 ... lbes, as played in F.Khairallah­
T.Cali stri, Cannes 2007. Black has the
same idea of playing ... hs.
22 @e1
22 f4 can still be met by 22 ... lbes
(Black could also try 22 ....Jtf8 23 @e1
.Jt g7) 23 @e1 (2 3 .Jtd4? I:txf4+) 23 ... lbg4.
22 ...lbes 23 'ii'g 3 .JtfB 24 lbbs
This was P.Van Hoolandt-J.Cabrera
Trujillo, Cannes 2007. Here Black
should play 24... c6! to open the posi­
tion when the extra piece should count
for more than White's two pawns.
 Chapter 3
The Mar del Plata Variation
9 lbe1 tt:Jd7 10 ..te3

1 d4 lllf6 2 c4 g6 3 lllc 3 .Jtg7 4 e4 d6 S bishop, not only with .. .f4, but likely too
..'Lif3 o-o 6 .Jte2 es 7 o-o lllc6 8 ds llle7 9 from a ... g4-g3 advance in the form of a
lllel llld7 10 .Jte3 pawn sacrifice. In this variation White's
queenside attack is very straightfor­
ward, but the defence of his king can
be a bit more complicated.
Some consider this line to be a bit of
a suicidal variation for White. Indeed
Black's kingside attack can be very
strong, but both players must develop
a feel for attack and defence to be able
to enjoy success from either side of the
board in these lines.
10...fs 11 f3 f4 12 .Jtf2 gs
Although the 10 llle1 fS 11 llld3 and
10 f3 fS 11 g4 lines of Chapters 1 and 2
respectively have had their own kind of
resurgence in recent years, there is no
doubt that the aggressive 10 .Jte3 is the
main line of the 9 lll e 1 complex. From
e3 the white bishop points directly at
Black's queenside, not only supporting
c4-c5 but also targeting the a7-pawn.
This disadvantage is that Black's pawn­
storm will gain time by attacking the

47
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
The last few moves have been self­
evident, but now White faces a broad
choice.
A� 13 ti4;
: � '· " " '
9tl.3.l£dJ .
C;,t.J. 'llbs
Q::ti l.tt
·E:U U.
Instead 13 g4 is rare, but not neces­
sarily bad. White tries to play alo.n g the
lines of Line D of Chapter 1 and all of
Chapter 2. Now should Black play
13 ... hS 14 h3, his play is slower than in
those similar lines we have examined
because White has got his bishop to
the f2-square very quickly, where it is
useful for both pressuring the queen­
side and defending the h4-square. The
f4-square is weaker than before, how­
ever, and taking on g3 is now quite vi­
able. Indeed, Black can quickly initiate
kingside play with 13 .. .fxg3! 14 hxg3
lllg 6.
Black takes aim at the f4-square.
Now ... hS-h4 is one idea. White has:
48
a) 1 S 'ii'd2 h S 1 6 lllg 2 (16 j_e3 j_f6
with the idea of ... h4 is given by Nunn),
and now rather than 16 ... as 17 llla 4!?
b6 18 a3 h4 19 g4 lllf4 20 @h2 lllf6 21
llle3 with an edge for White in
S . S avchenko-A.Fedorov, Nikolaev 1993,
Black could play 16 ...h4 17 g4 lllf4 18
@h2 lllx g2 19 @xg2 l:.f7 20 b4 lllf8,
which looked okay for him in M.Socko­
K. Kachiani Gersinska, Wuppertal 1998.
Here Bologan's suggestion of 16 ...j_h6
17 j_e3 @h7 18 b4 as 19 a3 h4 with
counterplay also looks sufficient.
b) With 1S lllg 2 White would like to
bring the knight to fs. Now:
bl) 1s ... hs 16 llle3 as (or 16 ... h4 17
g4 lllf4 18 @h2 as 19 lllfs j_f6 20 'ii'd2
b6 21 b3 lllcs 22 .i::tab1 j_d7 23 j_d1
'ii'c8 24 a3 'ii'a6 with an unclear posi­
tion in D.Ruzele-R. Speckner, Regens­
burg 1998) 17 @h2 lllc s 18 llla 4 b6 19
lllxcs bxcs 20 'ii'c2 h4 21 g4 j_d7 22
lllfs j_f6 was V.Dydyshko-L.Van Wely,
Yerevan Olympiad 1996. The position is
messy, but looks satisfactory for Black.
b2) 1s ... as (Black secures the cs­
square for his knight) 16 llle 3 (White
can get the knight to fS, but this is not
fatal; Black has squares too) 16 ... lllcs 17
b3?! (better was 17 'ii'c2 because now
the c3-knight is loose and Black makes
use of this) 17 ... lllf4! 18 j_e1 (instead
18 gxf4 exf4 19 'ii'd2 fxe3 20 j_xe3
j_es ! is nice for Black - if 21 j_xgs?
j_xc3 ! ) 18.. ll.f6 ! ? 19 gxf4 gxf4 20 lllg 2
l:Ig6 21 I:.f2 @h8! and Black had a
strong attack in N. Stanec-G.Timo­
shenko, Vienna 1998.
 Th e M a r def Plata Varia tio n : 9 lbe1 lb d 7 1 0 iL e 3
A) 1 3 b4
This obvious preparation for cs has
fallen out of favour since the classic
game Piket-Kasparov. However, it is not
completely harmless and Black should
not take it too lightly.
13 lbf6 14 cs lbg6 15 cxd6
...
After 1S .l::t c1 .l::tf7 16 cxd6 play sim­
ply transposes, while 16 a4 gives Black
the opportunity to capture on d6 with
his bishop after 16 ...il..f8 17 cxd6
il..xd6! ?. This has been tried a few
times, but actually I prefer 17 ...cxd6,
transposing to the main line.
Instead 1S a4 is important. Black
has to be careful here or else he could
be tricked into a line that he may not
be prepared to play.
(seefollowing diagram)
Gallagher points out that the natu­
ral 1 s ... .l::tf7 may well be unnecessary
when White has not played .l::t c 1 be­
cause the c7-square does not need ad­
ditional defence. Indeed, after 16 as we
have transposed to the risky line 13 a4
lbg6 14 as l:If7 1s b4 lbf6 16 cs which is
discussed in the notes to Black's 13th
move in Line E.
Black should probably prefer 1s ...h s
and then:
a) 16 as (16 cxd6 cxd6 17 as g4 is
the same) 16 ... g4 17 cxd6 cxd6 18 lbbs
g3 19 hxg3 (19 il..x a7! ?) 19 ...fxg3 20
il.. x g3 a6 21 lba3 h4 22 il..f2 lbhs 23 f4
lbhxf4 24 lbc4 'ifgs gave Black an at­
tack in L.Ftacnik-1. Smirin, Biel Inter­
zonal 1993.
b) 16 h3 l:If7 17 as il..f8 18 c6 .l::th 7!
(Black has plenty of support for ...g4, so
the rook is better off on the-soon-to-be­
opened h-file) 19 cxb7 (19 bS b6! 20
axb6 cxb6 21 .l::ta3 'ii' c7 22 lbd3 g4 23
fxg4 hxg4 24 hxg4 'ii' g 7 2S il..xb6 'ii'h 6
gave Black a strong attack in l.Hausner­
S .Dolmatov, German League 1993. de­
spite his ugly queenside) 19 ...il..xb7 20
bS (or 20 a6 il.. c8 21 lbbs g4) 20...il.. c8
with counterplay.
c) 16 c6 il..h 6 (Black could also try
16 ...aS!?) 17 bS b6 18 as g4 (unfortu­
nately 18 ... .l::tb8 19 axb6 axb6 20 'ii'a4
g4 2 1 'ii'a7 gxf3 22 'ii'xb8 fxe2 23 lbxe2
49
A ttacking C h e s s : Th e Kin g 's I n d i an, Vol u m e 1
lllxe4 24 l::t a7 does not look sufficient)
19 axb6 cxb6 20 fxg4 hxg4 21 g3 l::t f7
was unclear in V.Korchnoi-Xie Jun,
Marbella 1999.
1s...cxd6 16 Itel l::tf7 17 a4
17 ...j_fB!
This is Kasparov's clever idea. For
the moment Black leaves the h S-square
open for his knight. The alternative is
the immediate 17 ... h s which also
seems playable. After 18 as Black has:
a) 18... g4 19 lllb s g3 20 j_xa7 lllh 7
21 @h1 (to make room for j_g1)
21 ...l::t xa7 22 .U.xc8 'ii'xc8 23 lllxa7 'ii'd 8
24 h3 lll g s (White was much better
after 24 ...'ii'M 2s 'ii'c2 lt:lgs 26 j_bs lllf8
27 @g1! in L.Bass-K. Spraggett, New
York 1983) 2s 'ii'a4 'ii' b 8 26 lllbs lllx h3
27 llld3 (27 gxh3? 'ii'c8 wins) 27 ...lll g s
28 l:Ic1 'ii'd8 29 'if c2 lllf8 30 'ii'c8 'ii'f6 31
@g1 l::t d7 and Black developed coun­
terplay in L.Ftacnik-T.Oral, Slovakian
League 199S.
b) 18...j_d7 19 lllb s j_xbs 20 j_xbs
g4 21 @h1 g3 22 j_g1 gxh2 23 j_f2 a6
(23 ...h4 24 @xh2 lllh s 2s �g1 lllg 3 26
a6 bxa6 27 j_xa6 j_f8 28 llld3 l::t h 7 29
50
j_e1 h3 3 0 gxh 3 'ii'h4 31 lllf2 @h8 -
31 ... 'ii'h s runs into 32 j_c8! - 32 �c8!
.U.xc8 33 j_xc8 j_e7 34 lllh 1! saw White
defend in J.Piket-P. Paneque, Adelaide
1988) 24 j_b6?! 'ii'f8 2S j_e2 h4 26
@xh2 lllh s 27 llld3 j_f6 28 .U.g 1 lllg 3 29
'ii'd2 j_d8! gave Black good play in
G.Burgess-W.Watson, British Champi­
onship, Plymouth 1989.
18 as j_d7 19 lllb s
This allows Black to execute his
main idea. Instead 19 j_bS j_c8 20 j_e2
repeats (if this is too off-putting then
there is 17 ... h s), while White has also
tried 19 @h1 l::t g 7 (19 ...'ii'e8 was sug­
gested by Kasparov), and here:
a) 20 lllb s g4 21 lllxa7 g3 22 j_b6
'ii'e 8! 23 .i::te7 gxh2 24 l::t x b7 lllh s gave
Black some attacking chances in
G.Burgess-B.Badea, Prestwich 1990.
b) 20 j_bS g4 21 j_xd7 'ii'xd7 22 fxg4
lllxg4 23 lllf3 j_e7 24 j_g1 lllh 4 was
good for Black in D.Gurevich­
H.Gruenberg, New York 1991.
19...g4
20 lllc 7?!
This appears to be too greedy, but
 The M a r def Plata Variati o n : 9 t:De1 tiJd7 1 0 iLe3

as Bareev once pointed out, the pieces B) 13 CiJd3

move differently when one is playing
against Kasparov.
White can try to bail out with 20
fxg4, although Kasparov's suggestion
20 ... t:Dxe4 21 t:Dc7 il..a4 22 'ii'xa4 �xc7
looks at least equal for Black.
Instead 20 t:Dxa7 g3 21 il.. b6 'ii'e 7 22
'it>h 1 t:Dh 5 23 t:Db5 'ii'h 4 (23 ...gxh2 24
fi.f2 t:Dg3+ 25 il.. x g3 fxg3 with a clear
advantage for Black is given by Kas­
parov) 24 il.. g 1 I:tf6 25 t:Dc7 t:De7 26
t:Dxa8 ?! gxh 2 27 il..f2 'ii'xf2! 0-1 was This old move seems to mix up sys­
K.Friesen-M.Lomineis hvili, Rotterdam tems, but t:Dd3 is almost always a use­
1998. Here White could look into 26 ful move, so it can hardly be that bad.
CiJd3, which was suggested by Galla­ 13 ... t:Df6 14 cs t:Dg6 15 a4
gher. The idea is to meet the move Instead 15 cxd6 cxd6 could easily
26 ...I:th6 with 27 'ifel or 26 ... gxh2 with transpose just about anywhere, but
27 fi.f2. there is no benefit for White in giving
20 ... g3! 21 t:Dxa8? up all of his flexibility - see too note 'a',
This just carries things too far. In­ below.
stead 21 hxg3 fxg3 22 il.. xg 3 il..h 6 15 Itel is a valid alternative, how­
(22 ... t:Dh 5) 23 t:Dxa8 t:Dh 5 2 4 fi.f2 t:Dgf4 ever. Black should reply 15 ...Itf7.
25 t:Dd3 Itg7 26 t:Dxf4 fi.xf4 27 g4 il.. xcl
28 'ii'xcl t:Df4 29 'ii' e3 h5 with an attack
is given by Kasparov, but this was still a
better try.
21 .t:Dhs 22 @h1
..

22 fi.xa7 'ii'h 4 23 h3 fi.xh 3 24 gxh3

'ii'xh 3 25 .U.f2 gxf2+ 26 @xf2 t:Dh4 27
ii.fl (27 t:Dd3 'ii'g 3+ 28 @f1 t:Dg2!)
27...'ii'h 2+ 28 t:Dg2 Itg7 looks very
strong.
22...gxf2 23 I:txf2 t:Dg3+ 24 @g1 'ii'xa8
Black is much better. The game Now 16 'iih 3 allows 16 ... g4 immedi­
J.Piket-G.Kasparov, Tilburg 1989, fin­ ately, so White might prefer:
ished 25 il..c4 a6! 26 'ii'd 3?! 'ii' a 7 27 b5 a) 16 cxd6 is the most common sta­
axb5 28 il..x b5 t:Dh1 ! 0-1. tistically, but it is never played by

51
A ttacking C h e s s : Th e King 's I n dian, Vol u m e 1

Grandmasters. When White gives up nov-M.Najdorf, Zurich 19S3.

his queenside flexibility, it makes i t
easier for Black to formulate h i s plans.
After 16 ...cxd6, 17 lllb s, for example,
allows 17 ... g4 with good play.
b) White has also tried 16 a4 .Jtf8 17
as hS 18 h3 which can be met by
18 ...l:Ig7 or even 18 ...g4.
c) The immediate 16 lllb S can, of
course, be met with 16 ... g4. By now we
can see a kind of relationship between
the bS- and g4-squares. If White plays
lllbs too early, Black can often play ... g4 e) 16 @h1 is often a useful prophy­
without further preparation because lactic move, but here 16 ....Jtf8 17 a4 h S
the e4-pawn is loose. After 17 a4 g3 18 1 8 a s g4 1 9 cxd6 cxd6 2 0 lllb s g3 21
hxg3 fxg3 19 .Jtxg3 lllh s 20 .Jth2 .Jth6 .Jtxa7 (21 hxg3 fxg3 22 .Jtxg3 is met by
21 .U.c3 a6 22 llla3 .Jte3+ 23 lllf2 .Jtf4 22 ...h4 and 23 ... lllh s) 21 ... lllh7
Black had a winning attack in Y.Du (21 ...'ii'xas!?) 22 .Jtb6 (or 22 .Jt g 1 h4 23
Bois-V.Bologan, Bern 1997. h3 I:txas with the idea of ... .Jtxh3 and
d) 16 .U.c2 intends to double rooks . ..'ii'd7, which gives Black an attack)
on the c-file. This move was played in 22 ... 'ii'h4 23 .Jtg1 lllg s 24 llle 1
one of the most famous King's Indian
'prototype' games: 16.. ..i.f8 17 cxd6
cxd6 18 'ii'd2 g4 19 .i::tfc1 g 3 ! (the the­
matic sacrifice) 20 hxg3 fxg3 21 .Jtxg3
lllh s 22 .Jth2 .Jte7 23 lll b 1?! .Jt d7 24
'ii'e 1 .Jigs 2s lll d 2 .Jte3+ 26 @h1 'ii'g s 27
.Jtf1 I:taf8 28 .U.d1 bS 29 a4 a6 30 axbs
axbs 31 Ite7 Itg7 32 lllb3 lllh4 33 I:tc2
.Jth3! (every one of Black's pieces par­
ticipates in the attack)

(seefollowing diagram) 24 ... lllh 3! 2S gxh3 was A. Shirov­

34 'ii'e2 lll xg2 3 S .Jtxg2 .Jtxg2+ 36
'ii'xg2 'ii'h 4 37 'ii'x g7+ @xg7 38 .U.g2+
@h8 39 llle 1 lllf4 40 Itg3 .Jtf2 41 l:tg4
'ii' h3 42 llld2 h s 43 I:tgs 0-1, M.Taima-
T.Radjabov, Baku (rapid) 2009, and now
Golubev suggests 2s ...'ii'xh3 26 .Jtd3
lllh4 27 .i::t c2 .i::t g7 with good attacking
chances.
We now return to 1S a4:

52
 Th e M a r def Plata Varia tion : 9 t:De1 t:D d 7 1 0 iL e 3
1s ... hs
This is the most obvious choice, but
there are other ideas:
a) 1s ...I:tf7 16 as il..f8 17 cxd6 il.. xd6
(I would prefer 17 ... cxd6) 18 t:Dcs t:Df8
19 'ii b3 'ii e 7 20 t:De6?! t:Dxe6 21 dxe6
il..xe6 22 'ii xb7 'ii f 8! gave Black coun­
terplay in M.Golubev-S.Ovsejevitsch,
Alushta 1999. However, 20 t:Dxb7! looks
strong.
b) 1s ...@h8! ? introduces a different
plan and might even be Black's best.
Because the move ...@h8 is useful and
in this position White is not targeting
C7, Black plays ...@h8, ... l:Ig8 and ... il..f8,
instead of ...Itf7, ... il..f8 and ... Itg7.
After 16 as Black has:
bl) Following 16 ...a6! ? 17 cxd6 (17
@h1 l:Ig8 18 cxd6 cxd6 19 b4 il.. d 7 20
.U.a3 il..f8 gave Black counterplay in
Z.Kozul-Z.Lanka, Batumi 1999) 17 ...cxd6
18 t:Da4 g4 19 t:Db6 .U.b8 20 t:Dxc8 I:txc8
White is probably a bit better, although
Black has counterplay. This could also
come about after 13 a4 t:Dg6 14 as a6
1S CiJd3 @h8 ! ? in the notes to Black's
13th move in Line E.
b2) 16 ... .U.g8 17 cxd6 cxd6 18 t:Dbs g4
19 fxg4 t:Dxe4 20 il..xa7 il.. d7 21 il..b6
'ii e7 22 t:Da I:taf8 is very messy. A cou­
ple of practical examples:
b21) 23 I:ta3 t:Dgs 24 b4 e4 2s t:De1
was V.Korchnoi-Z.Lanka, Linz 1997.
Now Black should play 2S ...t:Des with
the idea of .. .f3 according to Lanka.
b22) 23 t:De6 iLxe6 24 dxe6 t:Dgs 2s
t:Db4 e4 26 t:Dds (better is 26 il..c4 with
an unclear position in T.Chmielewski­
D.Lybin, correspondence 199S)
26 ...'ii xe6 27 t:Da 'fies and Black had
the initiative in P.Kiriakov-Y.Zimmer­
man, Sochi 2004.
16 a 5
Instead 16 cxd6 cxd6 17 as g4 1 8
t:Dbs g3 transposes t o note 'a' to
White's next move.
Instructive is 16 h3 Itf7 17 c6 as!.
(seefollowing diagram)
This fine move stops the advance of
White's a-pawn and also slows down
White's t:Db4-c6 manoeuvre:
53
A ttacking C h e s s : Th e King 's I n dia n , Vol u m e 1
18 cxb7 .i.xb7 19 b4 .i.c8 20 bxas
�h6 21 lllb 4?! (better is 21 a6 .i.xa6 22
lllb4 .i.c8 with an unclear position ac­
cording to Kasparov) 21 ...g4 22 lllc6
'ii'f8 23 fxg4 hxg4 24 hxg4 .i.gs 2 5 �f3
'ii'h6 26 .i::te1 lllh4 27 .i.xh4 .i.xh4 28 g s
'ii'xgs 2 9 �e2 lllg4 30 I:tb1 .i.g3 31 �d3
'ii'h4 0-1, V.Korchnoi-G.Kasparov, Am­
sterdam 1991.
16...g4
Black has easily achieved his king­
s ide break and White must decide how
to continue on the queenside.
17 c6
Others:
a) 17 cxd6 cxd6 18 lllb s g3 19 .i.xa7
54
(19 hxg3 fxg3 2 0 .i.xg3 h 4 with the idea
... lllh s is the usual theme) 19 ... lllh7
(19 ...llle 8!? is also worth considering)
20 h3 'ii'h4 21 .i.b6 .i.xh3 22 gxh3 'ii'x h3
23 I:tf2 lllh4! 24 'ii'f1 gxf2+ 25 lllxf2
'ii'g 3+ 26 @h1 was B.Larsen-E. Torre,
Bauang 1973. Now both 26 ... lllg s and
26 ... lllxf3 should win for Black.
b) 17 a6! ? is probably the most test­
ing. Black has:
bl) 17 ...bxa6 18 lllb4 g3 19 hxg3
fxg3 20 .i.xg3 h4 21 lllc6 is annoying
for Black.
b2) 17 ...b6 18 cxd6 cxd6 19 lllb4 g3
20 hxg 3 fxg3 21 .i.xg3 h4 and here, in­
stead of 22 .i. h2 lllh s 23 lll c6 'ifgs, the
immediate 22 lll c 6! is again problem­
atic.
b3) 17 ... g3 18 hxg3 fxg3 19 .i.xg3 h4
20 .i.h2 lllh s 21 axb7 .i.xb7 22 f4
lllhxf4 23 l2Jxf4 lllxf4 24 �xf4 exf4 25
�g4 was W.Harper-K.Waidyaratne,
Connecticut 2007. This is a favourable
version of the main line for White.
11 ... g3 18 hxg3 fxg3 19 .i.xg3 h4 20
.i.h2 lllh s 21 cxb7 .i.xb7
Black's kingside play looks ominous,
 Th e M a r def Plata Varia tio n : 9 lb e 1 lb d 7 1 0 iL e 3
but White has a way to fight for space
on that flank by sacrificing the pawn
back
22 f4!
Both sides must always keep this
idea in mind. White returns the pawn
in order to fight on the kingside.
22...lbhxf4 23 lbxf4 ibxf4 24 il..xf4 exf4
2s �g4 'if gs
26 �e6+ @h8 27 I:tf3 il..c8 28 Ith3 �xe6
29 dxe6 �ae8
Here Black retained sufficient play
in V.Korchnoi-Xie Jun, Amsterdam
2001.
C) 13 lbbs
Korchnoi's idea looks odd at first,
but it is actually quite thematic. White
immediately goes after the valuable c8-
bishop. Eventually Korchnoi gave it up,
but the American GM Dmitry Gurevich
still frequently plays this line.
13 ...b6
Instead 13 ...a6 falls in with White's
plans: 14 lba7 l:Ixa7 (otherwise White
will just take Black's precious bishop on
c8) 1s il.. xa7 b6 16 b4 il.. b7 17 cs dxcs
18 Itel! and White maintained the ini­
tiative in V.Korchnoi-K.Hulak, Zagreb
lnterzonal 1987.
The pawn sacrifice 13 ...lbf6!? 14
lbxa7 i..d7 is interesting:
a) 1s lbbs lbg6 16 'ii'c2 hs 17 cs g4
18 c6 bxc6 19 dxc6 il..c8 20 il..c4+ @h8
21 lbd3 g3 was D.Gurevich-M.Golubev,
Biel 1992, when White should probably
test Black's idea with 22 hxg3 fxg3 23
il..x g3 h4, although Black certainly has
counterplay.
b) 1s cs l:Ixa7 16 cxd6 lbc8 17 dxe7
'ii'xe7 18 il.. xa7 lbxa7 19 'ii'b3 'ii'cs+ 20
@h1 lbhs 21 lbd3 and here, instead of
21 ...'ii'e 3 22 lbf2 'ii'x e2 23 'ii'xb7 l:If7 24
55
A ttacking C h e s s : Th e King 's I n d ian, Vol u m e 1
'ii'xa7 g4 2s l:tfe1 'ii'bs 26 fxg4 with a
big advantage in J.Benjamin-J.Nunn,
Hastings 1987/88, Black should play
'ii'd6! (with the idea of ... lllg 3) 22 @g1
lllb s with counterplay.
14 b4
After 14 a4?! a6 1S lllc3 as White
has lost two tempi and Black only one.
Thus White is just a tempo down on
Line E.
14 a6
...
The knight has to get booted at
some point. After 14... lllf6 1S cs g4?!
(better is 1S ... a6 16 cxd6 cxd6 17 lllc3
hS, heading for the main line) 16 cxd6
cxd6 17 Itel White has the initiative.
1s lllc3
The alternative is 1S llla3 and here
1S ... as 16 cs axb4 17 cxd6 cxd6! ? with
the idea ... lllcs is interesting, while
practice has seen:
a) 1S ...h s 16 cs bS 17 llla c2 lllf6 18
a4 bxa4 19 .l::txa4 lllg 6 20 bS g4 21 lllb4
g3 22 hxg3 fxg3 23 .Jtxg3 h4 24 lllc6
'ii'd7 2S .Jth2 .Jth6 was a complete mess
in A.Huzman-1. S mirin, Sverdlovsk 1987.
b) 1S ...l:If6 ! ? 16 llld 3 l:Ih6 17 cs 'ii'e8
56
18 @h1 bS 19 cxd6 cxd6 20 lllb1 lllf6 21
llld2 'iVhs 22 .Jt g1 'ii'g 6 (with the idea of
...g4) 23 Itel lllh S! and White was in big
trouble in M.Gavilan Diaz­
C.Matamoros Franco, Malaga 2010.
1s l:If6!?
...
Black goes for the attack along the
h-file. Usually I am sceptical about this
idea, but here I think there is consider­
able justification in playing this way.
This is about as good a version as Black
will get of this plan because he has
played ... a6 for free. This fits in with the
straightforward plan to play ... .l::th 6 and
...'ii'e8 because White can no longer
attack c7 with lllb s.
The other logical approach is 1S ...hs.
Here White's queenside play is not ob­
vious, but Dmitry Gurevich has demon­
strated a dangerous positional concept
after 16 cs (White has tried other
moves such as 16 Itel and 16 a4, while
one high-level example continued 16
@h1 lllf6 17 cs g4 18 cxb6 cxb6 19 Itel
g3 20 .Jtg1 gxh2 21 .Jtf2 h4 22 llla 4 .l:Ib8
23 bs axbs 24 .Jtxbs lllhs 2s @xh2 lllg 3
26 .l::t g 1 lllg 6 and Black had counterplay
 Th e M a r def Plata Varia t i o n : 9 t:De1 tiJ d 7 1 0 iL e3
in V.Korchnoi-Ye Jiangchuan, Novi Sad
Olympiad 1990) 16 ... t:Df6 17 cxb6 cxb6.
It looks like White will have trouble
breaching the queen side, but in fact his
play is very fast The b6-pawn will come
under fire and White can also exploit
the weakened c6-square: 18 t:Da4! l::t b8
19 b S ! (the key move) 19 ... as 20 l:Icl g4
21 'ii'd2 (now Black has to watch out for
ideas of capturing on b6 followed by
'ii'x as) 21 ...il.. d7 (after 21 ... t:Dg6 the blow
22 l::t c6 is even stronger) 22 l::t c6! .
Now instead of 22 ...il.. xc6 23 dxc6 g3
24 hxg3 fxg 3 2S il.. x g3 t:Dg6 26 'ii'c2
which was very good for White in
D.Gurevich-E.Yanayt, Las Vegas 2006,
Black should play 22 ...t:Dxc6 23 dxc6
il.. e 6 24 il..xb6 (24 t:Dxb6 fails to 24... g3)
24...l:Ixb6 2s 'ii'xas CiJd7 when White
has the initiative, but at least the posi­
tion remains complicated.
16 c5
After 16 t:Dd3 l::t h6 17 cs 'ii'e8 18 @h 1
bxcs 19 bxcs t:Df6, the idea of ...'ii' hs
and ... g4 gave Black an attack in J.Kiltti­
V.Maki, Tampere 1998
16... l::t h 6 17 cxd6 cxd6 18 CiJd3 'ii'e 8 19
ii.el
Instead 19 @h 1 'ii'h s 20 il.. g 1 t:Df6 21
t:Df2 'ii'e 8 22 t:Dg4 iLxg4 23 fxg4 l::t c8 24
Itel l::t xc3 2S .U.xc3 t:Dxe4 gave Black
good compensation in D.Khismatullin­
F.Amonatov, Moscow 2009.
19...t:Df6 20 t:Df2 l:Ig6
Compared to the positions arising
after lS ...h S, White's bishop has been
diverted from attacking the b6-pawn.
After 21 Itel hS 22 h3 Black played
22 ...il..h 6 in J.Federau-K.Klundt, German
League 1988, while 22 ...il.. d7 was seen
in X.Sole Fabregat-A.Pablo Marin, Cata­
lonian League 1998. Both moves look
good - Black has a strong attack
57
A ttacking Chess: The King 's I n d i a n, Vo l u m e 1

D) 13 Itel practical try, but I have my doubts

about its true worth. White can pre­
pare the CS-break OT play it straight­
away as a pawn sacrifice:

This is a very natural and obvious

move, and it may transpose to other
lines if White follows up with a quick
b4 or llld 3. However, there is one im­ a) 14 b4 has scored badly for White
portant independent idea connected overall, but matters are not so clear.
with a pawn sacrifice. 14... .l:th6 1S cS! 1S ... a6 (1S ...'ii'e 8 16 lllbS
13 lllg6
... 'ii'h S 17 h4 lllg6 18 g4 fxg3 19 .Jt xg3
This is the main line. Black has also lllf4 20 lllx c7 may look risky for White,
recently tried 13 ... a6, which seems in­ but it is hard to find anything convinc­
sufficient to me: 14 llld3 b6 (14....l:tf6 1S ing for Black) 16 cxd6 cxd6 17 llld3 lllf6
cs .l:th6 transposes to 13 ....l:tf6, below) 18 llla4! 'ii'e8 19 lllb6 �S 20 h4 was
1S b4 .l:tf6 16 cs .l:th6 17 @h1 ! ? 'ii'e 8 18 S.Atalik-1. Nikolaidis, Halkida 1997.
cxd6 cxd6 19 bS lllf6 20 bxa6 'ii'h s 21
.Jt g 1 'ii'g6 22 lllf2 lllh s 23 lllh 3 lll g 3+ 24
hxg3 .Jtxh3 2S gxh 3 fxg 3 ? (this is bad,
but both 2S ...�xh3+ 26 @g2 and
2s ...'ii'h s 26 I:tf2 fxg3 27 .i:Ig2 'ii'xh 3+ 28
.Jth2 lll g 6 29 'ii' e1 gxh 2 30 .l:txgs also
look insufficient) 26 f4! exf4 27 .Jtg4
with a clear advantage to White in
J.Berkvens-E.lnarkiev, Kerner 2007.
The main alternative is 13 ....l:tf6.
Black intends to play ....l:th6 and
...'ii'e8 when ... 'ii'h s could be a big This is a good example of White's
threat. This is a popular way of meet­ 'desperate' h4 defence, leaving Black's
ing 13 Itel and may well be a good attack looking too speculative. Now

58
 Th e M a r de/ Plata Va ria t io n : 9 lbe1 lb d 7 1 0 iL e 3
2 0 ... l::t b8 21 lbxc8 .U.xc8 22 .U.xc8+ lbxc8
23 'ii'c2 lbe7 24 'ii'c7 lbg6 2S 'ii'xd6
lbxh4 26 'ii'e6+ @h8 27 il..xh4 'ii'xh4 28
'ii' h3 'ii'xh3 29 gxh3 l::t xh3 30 Itel hS
was suggested by Atalik, but this looks
in sufficient for Black.
b) 14 lbd3 also has scored badly
overall, but looks promising for White
as well after 14 ...l::t h 6 1S cs when again
Black has to decide whether or not to
prevent lbbS:
bl) 1S ...'ii'e 8 still seems too optimis­
tic after 16 lbbs 'ii'h s 17 h4 lbg6 18 g4
fxg 3 19 il.. x g3, which looks ugly, but
Black's queen side is melting away and
19 ...lbf4 20 lbxc7 is almost identical to
variation 'a' above. Even though
White's king position looks bad, it is
not clear how Black should continue as
his pieces are awkwardly placed.
b2) 1S ... a6 and now White has two
promising moves:
b21) 16 cxd6 cxd6 17 a4 'ii'e8 18
ii.el! lbf6 19 ibf2 l::t g6 20 as il..d7 21
lbb1 (with the idea lba3-c4-b6) 21 ...hS
22 h3 g4 23 fxg4 hxg4 24 hxg4 il..h 6 2S
l::t c7 l::t c8 26 .i::txc8 lbxc8 27 lbd2 f3 28
gxf3 lbhS 29 lbh3 gave White a win­
ning position in M.Dambacher­
R.Ponomariov, European Club Cup, Kal­
lithea 2008, which was a game with a
large rating disparity.
b22) 16 c6 bxc6 17 dxc6 lbf8 18 lbb4
'ii'e8 19 @h1 lbe6 20 lbcdS @h8
(20 ... lbxdS 21 lbxdS is similar) 21 lbxe7
'ii'xe7 22 lbds 'ii'f7 23 il..g 1 lbd4 24 il..c4
il.. e6 2S 'ii'd3 as 26 b3 l::t b8 27 'ii'd2 l::t a8
28 a3 l::t h4 29 b4 axb4 30 axb4 il..xdS 31
il.. xds 'ii'h s 32 .U.a1 l::tf8 33 l::t a7 g4 34
fxg4 'ii'xg4 3S il..xd4 exd4 36 l::t f3 with a
clear advantage for White in
A.Lenderman-R.Barcenilla, Mesa 2009.
c) With 14 cs White continues his
plan and sacrifices a pawn to open the
c-file. Then 14 ... lbxcs 1S b4 lba6
reach es a key theoretical position for
the 1 3 - Af6 sub-variation .
White has:
cl) 16 lbbs l::t h 6 17 lbxa7 il..d7 18
il.. xa6 (White goes after the c-pawn;
instead 18 il.. bs lbxb4 19 'ii'a4 il.. xbS 20
'ii'xbS b6 21 a3 lba2 ?! 22 lbc6 lbxc6 23
.U.xc6 .U.xa3 24 'ii'b2 l::t a7 2S lbd3 g4 26
fxg4 'ii'g S was l.Rajlich-M.Al Sayed, Bu­
dapest 2001, and here 27 h3 would
give White a clear advantage) 18 ...bxa6
19 'ii' c2 g4 20 'ii'xc7 'ii'e8 and now:
c11) 21 g3? (this turns out badly)
21... @h8 22 .U.c2 lbg8 23 fxg4 il..xg4 24
il..b6 'ii'g 6 2S lbc6 fxg 3 26 lbd8 l::t xd8 27
'ii'xd8 gxh2+ and White had seen
enough in V.Korchnoi-1.Cheparinov,
Amsterdam 2008. After 28 l::t xh2 il.. f3+
29 .U.g2 il.. xg2 30 lbxg2 'ii'xe4 Black has
the initiative and is up material.
59
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
c12) 21 fxg4 il.xg4 and Black is not
threatened by either 22 tllc6 tllxc6 23
dxc6 'it'h5 24 h4 il.f6 (24..Jk8!? was
suggested by Mikhalevski) 25 'it'b7 nf8
26 nC3 il.xh4 27 il.xh4 'it'xh4 28 tllf3
hf3 29 ncxf3 'it'h2+ 30 �2 ng6 3 1
ng1 ng3 32 �f1 �h8 33 'it'e7 nfg8 34
'it'f6+ n8g7 35 'it'f8+ ng8 36 'iVf6+ n8g7
37 'iVf8+ V2-V2, V.lotov-V.Saravanan,
Kalamaria 2008, or 22 il.b6 'i!Vh5 (if
Black wants to play for a win then there
is Mikhalevski's suggestion of 22 ...tllc 8!
23 tllc6 tllxb6 24 'it'xb6 'i!Vh 5 25 tllf3
iLxf3 26 nxf3 'it'xh2+ 27 �fl il.f6 and
Black retains chances to attack) 23 tllf3
il.xf3 24 nxf3 'it'xh2+ 2 5 �fl ng6 26
nc2 ng3 ! 27 il.gl nxf3+ 28 gxf3 'ilr'h3+
29 �e2 'it'g2+ (or 29 ...tllf5 immediately)
30 il.f2 tllf5 3 1 exf5 e4, which leads to
perpetual check, S.Arounopoulos­
M.Ghinda, correspondence 2007.
c2) The assessment of 16 tlld3! ? has
gone back and forth a bit. 16...nh6 17
a4 'it'e8 18 �hl and now:
c21) 18... 'it'h5 19 il.g1 il.d7 20 tll a2
nc8 21 tllf 2! 'it'e8 22 tllg4 ng6 23 b5
tll c 5 24 il.xc5 dxc5 25 nxc5 gave White
60
the advantage i n R.Vera Gonzalez-
1.Nataf, Havana 2002.
c22) 18 ...il.d7 was Natafs improve­
ment against the same opponent, but
as Golubev points out, this may still be
insufficient: 19 'it'b3 �h8 20 tll b 2?!
(better is 20 il.gl, as there is no ...tllh 5
to worry about and 20 ... 'it'h5 is met by
21 tllf2) 20 ... 'it'h5 21 il.g1 g4 22 fxg4
il.xg4 23 il.xg4 'it'xg4 24 tll b 5 il.f6 2 5
tll x e7? tllxe7 26 nxa 'it'e2 27 nd1 'it'xe4
28 tllc4 tllf5 29 il.f2 il.h4 30 il.xh4 was
R.Vera Gonzalez-1.Nataf, Montreal
2003, and now the easiest win is
30...tllxh4 31 'it'h3 ng8 32 nc8 tllf5, as
pointed out by Gallagher. Instead 19
b5?! 'it'h5 20 il.g1 tll c 5 is good for Black,
but Golubev points out that 19 il.gl!
threatens b5 when 19...'ilr'h 5 just trans­
poses to variation 'c21'.
Now we finally return to 13 ...tllg 6.
Here 14 b4 and 14 tlld3 will trans­
pose to Lines A and B respectively, but
White has another idea.
14 cs!?
Kozul's pawn sacrifice rekindled in­
terest in 13 nc1. 1t is clear that 14... dxc5
 Th e M a r def Plata Va riatio n : 9 tD e1 tb d 7 1 0 ii. e3
lS b 4 is no fun after either 1S...cxb4 16
tDbs or 1S ...b6 16 t:Dd3, so Black has to
take the pawn with his knight.
14 tDxcs 15 b4 tDa6
•••
Not 1S ...tDd7? 16 tDbs. This is the point
of White's sacrifice - the black knight
has been sidelined for the foreseeable
future. Now White has a choice: he can
immediately regain the pawn, or he
can play slowly and count on gradual
pressure against Black's queenside.
16 tDbs
This is the most forcing continua­
tion.
White can also play 16 tDd3. Then
16 ...l:tf7 17 t:Dbs 1'..d 7 (17 ... hs 18 a4 ii..h6
19 t:Dxa7 1'..d 7 would be the same) 18
a4 hS (18 ...'it'e8 has also been played,
but it is probably too artificial) 19 t:Dxa7
ii..h6 20 l:tc3 �g7 21 t:Dbs tDf8 ! ? 22 h3
tDh7 23 ii.el tDf6 2 4 tDf2 (otherwise
... g4 comes) 24...tDxb4 2S �xc7 tDa6! 26
�xb7 tDcs 27 �c7 g4! 28 hxg4 hxg4 29
fxg4 tDfxe4 30 tDxe4 tDxe4 31 'it'b3 tDg3
32 ii..x g3 fxg3 33 'it'xg3 ii..xbs 34 �xg7+
�xg7 3S gs (3S ii..xbs? 'it'b6+ 36 �f2
ii..e3 wins for Black) 3 S ... ii..x gs 36 ii..x bs
'it'b6+ 37 �f2 'it'e3 was level in
V.Korchnoi-0.Cvitan, Pula 1997.
16 ii..d 7 17 tDxa7
•••
Instead 17 a4 l:tf7 18 t:Dd3 would
lead us back to 16 t:Dd3, above, but
Black could also try 17 ... h s or 17 ... 'it'b8.
11 h s
•••
This looks like the most accurate
move. Instead 17 ...'it'b8! ? 18 tDbs (or 18
ii..bs l:tf7 19 ii..xd7 �xd7 20 'it'a4 �f7)
18 ....l:tf7 19 a4 h S 20 tbd3 ii..h6 21 �c4
�g7 22 �hl 'it'd8 was the actual move
order of Atalik-Gufeld below, which
leads us back to the main line.
Black has also tried 17....l:tf7 and
here:
a) 18 ii..xa6 ! ? looks a bit better for
White: 18 ... bxa6 19 'it'C2 (19 tDc6 ! ?)
19 ... ii..e8 20 'it'e2 h S 21 'it'xa6 g4 22 tDc6
ii..xc6 23 'it'xc6 �xa2 24 �c2 �a3 2 S bS
ii..f8 26 b6 g3 27 ii.cs (27 hxg3 fxg3 28
ii..x g3 cxb6 29 ii..f2 tDf4 30 ii..xb6 'ifgs
31 'it'c8 was suggested in New in Chess
and looks strong) 27 ...�b3 28 'it'a4 �bl
29 'it'a2 �bS 30 'it'a6 �bl and Yi-Y2 was
D.Gurevich-J.Becerra Rivero, Las Vegas
2007.
b) 18 a4 h s 19 tDbs ii..f 8 (19 ...ii..h6 20
�c4 is the main line after 17...h s) 20
t:Dd3 (also good is 20 tDc3 �g7 21 tDd3
�h8 22 �hl 'it'f6 23 as 1'..e 7 24 bS tDb8
2S �al with some advantage in
A.Yermolinsky-J.Fedorowicz, North Bay
1998) 20 ... �g7 21 �hl lDM (insuffi­
cient is 21 ... tDh8 22 g3 fxg3 23 ii..xg3
tDf7 24 �gl �c8 2S ii..f2 'iVf6 26 ii..e 3
�g6 27 'it'd2 1'..e 7 28 tDc3 �h7 29 bS
tDcs 30 t:Dxcs dxcs which occurred in
61
A ttacking C h e s s : The King 's I n dia n, Vol u m e 1

A.Yermolinsky-H.Yoshiharu, Chicago g3 22 'i:Vxb7 gxf2+ 23 .llxf2 f3 24 tllxf3

2000, as 31 as looks much better for il.h6 2 5 l:tc3 'it'b8 26 'it'xb8 .llfxb8 27
White) 22 g3 fxg3 23 il.xg3 tllg 6 24 l:tgl tllc6 l:tbl+ 28 l:tfl l:txfl+ 29 �xfl l:txa4
'it'f6 25 'it'd2 l:tc8 26 il.f2 il.e7 27 il.e3 30 il.c4 il.e3 in V.Erdos-P.Acs, Hungar­
and here, instead of the 27 ...tllh 4? 28 ian League 2005.
il.xg S ! of A.Yermolinsky-J.Becerra 19 .llc4 l::t f7 20 tllbs l:tg7 21 �h 1 tllf8
Rivero, Chicago 2008, Black should play
21 ...tllf4 28 tllxf4 gxf4 29 il.f2 l:txg1+ 30
l:txgl+ �h8 31 tlla7 l:tg8 32 .llx g8+
�xg8 33 il.xa6 bxa6 34 tllc6 when
White is better, but Black is still in the
game.
18 a4
Capturing on a6 is still an option: 18
il.xa6 bxa6 19 tllc6 'it'f6 20 tlla s 'it'd8 21
tll c 6 'iVf6 22 tlla7 'it'd8 23 tll c 6 was
drawn in D.Gurevich-J.Becerra Rivero,
US Championship, Stillwater 2007. Black intends to play ... g4 with an
Here too White might have tried 19 attack However, because one of Black's
'i:Vc2!?. knights is so far away, White can afford
to fight on the kingside directly.
22 g3! fxg3 23 il.xg3 tllg6 24 tlld 3 tllf4
This is better than 24...�h7 2 5 l:tgl
.llf7 26 il.f2 with an edge in
A.Yermolinsky-S.Kindermann, Gronin­
gen 1997.

18...il.h6
Instead 18 ....llf7 brings us back to
17 ... .llf7, but an interesting idea of
Golubev's is 18 ...tllx b4!? with the idea
19 'it'b3 g4!. This has been seen in prac­
tice and the complications led to a
draw following 20 fxg4 hxg4 21 'i:Vxb4 2 5 il.xf4

62
 Th e M a r def Plata Va riati o n : 9 llJe1 lLl d 7 1 0 ii.e3
This is the most common, but prac­
tice has also seen 2S ll:lxf4 gxf4 26 ii..f2
�h8 (Black could try 26 ... M with the
idea of ... h3) 27 l:tg1 ii.gs 28 1'.. f l nc8 29
'it'e1 'it'e7 30 ii.. g 2 ii..f6 31 'if cl 1'.. h 4 32
ii..xh4 'it'xh4 33 ll:lxd6! cxd6? (a better
chance was 33 ... ncg8 34 ll:lfs ii..xfs 3 S
exfs 'it'f2 3 6 nc2 'it'b6) 3 4 nxc8+ ii..xc8
3S 'it'xc8+ �h7 36 ii.fl and White was
wining in l.Rajlich-A.Brustman, Ostrava
1999.
2s ...gxf4 26 ng1 ii.gs!?
This looks better than 26 ...nxgl+ 27
"i' xgl+ �h7 28 ii.fl 'it'e7 (Black could
try 28 ... ii..g s or 28 ... nc8) 29 ll:lxc7! l::tc8
30 'it'b6 ii..xa4 31 ll:lb2 (this leads no­
where, so 31 b S ! is better) 31 ...ii..b3 32
ll:lxa6 ii..xc4 33 ll:lxc4 bxa6 34 'it'xa6 ng8
3S 'it'b6? (3S 'it'xd6 'it'xd6 36 ll:lxd6 ii..f8
is about even) 3 S ... 'it'g7? (3S ... 'it'h4! with
the idea of ...'if el wins) 36 'it'f2 (36
ll:lxd6 ! ?) 36 ...'it'b7 was drawn in
S.Atalik-E.Gufeld, Honolulu 1997.
27 ii.fl 1'.. h 4 28 nxg7+ �xg7
29 'it'c2
Instead Atalik suggests 29 'if cl nc8
30 nc2, while here Timoshenko gives
30 ii.. g2 �h6 31 'it'f1 'it'e7 32 ii..h 3 l:tg8
33 ii..x d7 'it'xd7 34 lLlC 3. Both feel White
is a little better, but the position is still
very complicated.
29... �h6! 30 ll:lc3 'it'e7 31 ll:le2 b6 32
nc6
We've followed the game S.Atalik­
G.Timoshen ko, Timisu de Sus 1998.
Here Atalik suggests 32 ... ng8! with an
attack
E) 13 a4
This thoroughly modern continua­
tion is the main line and has been
played more than all the other 13th
moves combined. The move a4 is al­
ways useful and White may be able to
play the cs-break without any prepara­
tion, either through tactical means or
after Black moves the d7-knight to
support ... g4. The a-pawn may advance
to as, which often helps White by con­
trolling the b6-square if he manages to
play ll:lbs followed by capturing on a7,
especially because the f2-bishop will
have the b6-square to go to in case of
an eventual ... g3 by Black. The possibil-
63
A ttacking C h e s s : Th e K i n g 's I n dian, Volu m e 1
ity of a quick tllbs remains, and this
sortie could give White an improved
version of Line C.
13 as
...
Eventually this positional move has
come around to establishing itself as
the main line. Black prevents White's
further advance of the a-pawn and
fights for control of the dark squares on
the queenside. White will still break
through of course, but Black hopes that
the time White spends breaking down
his queenside pawn structure can be
used to create counterplay.
It is too early to play 13 ... h s because,
as we shall see, sometimes Black can do
without this move. There is another
plan with 13 ...�f6 14 as a6, with the
familiar idea of ... �h6 and ...'it'e8-hS.
The play is similar to that discussed in
the notes to Black's 13th move in Line
D, but here Black has even scored
worse, so we will not spend time look­
ing at it.
Black's most principled alternative
is 13 ... tll g 6, which is the move perhaps
most in the King's Indian spirit.
64
Here Black ignores White's play for
the moment and continues his own
campaign. It is very risky, but it is excit­
ing. After 14 as (White can also lunge
forth with 14 tllb s ! ?) Black has:
a) 14... tllf6 is too compliant, as it al­
lows 15 cs immediately. Following
1s ... hs 16 cxd6 cxd6 17 tllbs g4 both 18
tllxa7 and 18 ii..x a7 have scored heavily
for White.
b) 14...a6!? is not so bad and could
use further tests: 15 tlld3 tllf6 16 cs
�h8 17 cxd6 cxd6 18 tlla4 g4 19 tllb6
�b8 20 tllxc8 �xc8 is probably better
for White, although the position is not
so clear. We also considered this posi­
tion via a different move order in the
notes to Black's 15th move in Line B.
c) 14...h S is usually a signal for
White to play tllb s. This is because if
White plays tllb s too early (as in Piket­
Kasparov), Black may be able to get ... g4
in without preparing it with ...h s be­
cause the e4-pawn is not defended.
This is usually an achievement for
Black, not only because he saves a
tempo on ... h s, but also because the
move ...tllh s remains a possibility.
However, with Black spending time on
...h s, White should always consider this
lunge. Indeed, here 15 tllb S! tllf6 (or
1s ...a6 16 tlla7) 16 tllxa7 has scored
heavily in White's favour.
d) 14...�h8 intends ...�g8 and ... ii..f8.
Markos considers this to be Black's best
try, before recommending for White 15
tlld3 �g8 16 cs tllf6 17 cxd6 cxd6 18
tll b s g4 19 fxg4! tllxe4 20 'it'c2. After
 Th e M a r def Plata Va riatio n : 9 tDe1 tbd 7 1 0 ii. e 3
Golubev's 20...t:Dxf2, 21 t:Dxf2 controls
e4 and gives White an edge according.
Instead 20.. Ji.d7? runs into 21 t:Dxes
and 20 ...'it'e7 21 tDc7 tDxf2 22 t:Dxf2 nb8
23 tDe4 1'..d 7 24 'it'b3 ! (with the idea
tDe6) 24 ...ii..xg4 2S ii..xg4 'it'xc7 26 ii.. e 6
is also good for White according to
Markos. The best try may be giving up
the exchange with 20... tDf6 !? 21 tDc7
ii.xg4 22 t:Dxa8 ii..x e2 23 'it'xe2 e4 24
lDb4 'it'xa8 2S 1'..d4, but again Markos
thinks White is somewhat better and I
tend to agree.
e) The main line has always been
14...nt1.
Now White has:
el) Just like in Line D, Kozul has
gone lS cs here. I have actually tried
this several times with White, but now
I suspect lS b4 is just stronger. Follow­
ing lS ...tDxcs 16 ii..x cs dxcs 17 1'..c4 (af­
ter 17 a6 b6 18 1'..c4 ii..f 8! 19 d6 c6! 20
'it'b3 tDh8!, as in Z.Kozul-A.Kuzmin,
Oberwart 199S, Black will pick up the
d6-pawn and have two good pawns for
the exchange) 17 ... �h8 (17...ii..f8 18 d6
c6 19 tDd3 ii..xd6 20 tDa4 is much better
for White) 18 a6 a critical position is
reached.
Black has:
ell) 18...b6?! 19 d6 nd7 (or 19 ...nf8
20 ii..d s nb8 21 t:Dbs cxd6 22 t:Dxa7) 20
'it'ds c6 (forced, because 20 ... nb8 21
dxe7 'it'xc7 22 'it'g8 is mate!) 21 'it'xc6
nb8 22 lDbS ii..f8 was D.Vigorito­
E.Yanayt, Philadelphia 2003. Now 23
'it'c7! ii..xd6 24 'it'xd8+ nxd8 2S t:Dxa7
gives White a big advantage.
e12) 18...nf6?! 19 axb7 ii..xb7 20
tDd3 ii..f8 21 nt2 (also possible is 21
nas ii..c 8 22 t:Dxcs, as in D.Gurevich­
A.Sherzer, Durango 1992) 21 ... ii..d6 22
lDa4 'it'e7 23 nc2 ng8 24 tDf2 hS 2S
ii..e 2 ii..c 8 occurred in D.Vigorito­
J.Fedorowicz, US Championship, San
Diego 2006, and here 26 nacl! looks
good for White.
e13) 18 ... bxa6 19 tDd3 ii..f8 20 tDa4
(20 Ji.xa6 Ji.xa6 21 nxa6 C4 22 lDf2 Ji.CS
23 'it'e2 g4! gives Black counterplay ac­
cording to Bologan) 20 ... g4 (20 .. Ab8
has played by Popovic, and after 21 nt2
or 21 b3 I would suggest 21 ... c6! ?) 21
fxg4 'if gs 22 tDf2 ii..d6 23 na3 nb8 24
65
A ttacking C h e s s : The King 's I n dia n, Vo l u m e 1
nc3 nb4 was M.Llopis de Aysa-K.Volke,
Biel 1993, which is unclear according to
Bologan.
e2) The calm move lS b4 may be
best.
Here Black has:
e21) 1s ...il.f8 16 cs dxcs! was sug­
gested by Markos, leading to 17 d6
il.xd6 18 il.c4 cxb4 and now we have
the following:
e211) 19 'llt'b 3 tllh 8! 20 tllb s il.cs! 21
il.xcs tll x cs 22 nd1 'iVf8 23 il.xf7+ tllxf7
24 'it'xb4 tlle6 2S 'it'xf8+ �xf8 26 l:tf2
gives Black compensation according to
Markos.
e212) 19 tllb s tllcs (the 19 ... tlldf8 20
'it'b3 'it'e7 of P.H.Nielsen-M.Solleveld,
German League 2002, can be met with
21 tllxd6) 20 tllx d6! (20 il.xcs il.xcs+ is
check!) 20...'it'xd6 21 'it'xd6 cxd6 22
il.xcs dxcs 23 tll d 3 looks a little better
for White.
e22) 1s ...tllf6 is queried by Markos,
but this is the only move Bologan con­
siders: 16 cs il.f8 17 c6! (17 cxd6 is also
dangerous) 17 ... h s (Markos suggests
11 ... ng1 18 cxb7 il.xb7 19 a6 il.c8 20
66
�hl! il.d7 2 1 il.bs g4 2 2 il.xd7 'it'xd7 23
fxg4 tll x g4 24 tllf3 tllxf2+ 2S nxf2 il.e7,
as in P.H.Nielsen-Y.Kruppa, Minsk 1996,
although White looks somewhat better
here as well, and Markos's suggestion
of 22 il.c6 may be even better) 18 �hl
and White looks faster, but Bologan
has some ideas.
Black has:
e221) 18...nh1 19 tll b s g4 20 tllxa7
g3 21 tllxc8 bxc6 22 il.g1 gxh2 23 il.f2
'it'xc8 24 dxc6 h4 2 S il.c4+ �h8 26 'llt'b 3
was good for White in P.Kiriakov-M.Al
Sayed, Dubai 2002.
e222) Bologan's idea is to play on
the g-file after 18...ng1!?. His main line
goes 19 cxb7 il.xb7 20 a6 il.c8 21 tll bs
g4 22 tll x a7 g3 23 il.g1 l:txa7 24 il.xa7
cs 2s il.xcs dxcs 26 d6 tllh 7 27 il.c4+
�h8 28 na2 'i(h4 29 h3 il.xh3 30 gxh3
'it'xh3+ 31 �gl tllh4 32 'it'e2 cxb4 33 a7
nxa7 34 nxa7 il.xd6 3S na8+ �g7 36
nas 'it'c8 when Black is two rooks for a
piece down, but has tremendous coun­
terplay.
Now we return to the relatively­
safer 13 ... as:
 Th e M a r def Pla ta Varia tio n : 9 tDe1 tb d 7 1 0 ii.e3
14 lDd3 b6
Black must not allow cs. Now White
can regroup his pieces for both attack
and defence, or he can strike at once on
the queenside.
El: 15 .i.e1
E2: 15 b4
El) 15 ii.el
This is a flexible, yet somewhat slow
continuation. White wants to bring the
d3-knight to f2 in order to hold up
Black's ...g4-break. He also prepares to
recapture with his bishop after playing
b4 which would then support the as-
and cs-breaks.
15 ...tDf6 16 tDf2
The immediate 16 b4 axb4 17 t:Db s
transposes to Line E21, below.
16... h 5 17 h3
11 ...�hS!
This is a very important move. The
natural 17... tDg6?! 18 t:Dbs leaves Black
struggling to achieve ... g4, whereas
White is ready to break open the
queenside with b4. Thus the text in­
tends to regroup with ...tDeg8, ...tDh6,
....llf7, ... ii..f8 and ....ll g7 when ... g4 will
come with great force.
18 t:Db5
Both sides can employ a few differ­
ent move orders here. One example is
18 b4 tDeg8 19 bxas bxas 20 cs tDh6 21
t:Dbs llf7 22 cxd6 cxd6 23 .ll c l ii..f8,
transposing to the main line.
18...tDeg8 19 b4 .llf7!
It is important for Black to avoid
capturing on b4, because then White's
dark-squared bishop would spring to
life. After 19 ... axb4?! 20 ii..xb4 the aS­
break is coming and Black's play is too
slow. Instead 19 ...tDh6 is possible,
67
A ttacking Chess: Th e Kin g 's I n dian, Vo l u m e 1
though, and will lead to the main line
after 20 bxas (or 20 cs bxcs 21 bxcs)
20...bxas 21 cs nf7 22 cxd6 cxd6.
20 bxas
Again 20 cs bxcs 21 bxcs makes no
difference.
20...bxas 21 cs il.f8 22 cxd6
White should not become too fancy:
a) 22 tlla 3?! tllh 6 23 tllc4 ng1 24
na3 �g8 2S tllx as g4 26 fxg4 hxg4 27
hxg4 tllh xg4 28 tllx g4 tllxg4 29 tllc4
'it'g s 30 il.xg4 il.xg4 31 'it'c2 il.e7 32 nf2
'ikh S 33 g3 il.dl! 34 'it'b2 nf8 and Black
had a strong attack in S.Ghane Gardeh­
V.Spasov, Izmir 2003.
b) 22 c6?! tllh 6 23 na3 ng1 24 il.d2
g4 2S fxg4 hxg4 26 hxg4 tllh xg4 2 7
tllxg4 tllxg4 28 'it' el (this prevents
... 'it'h4 and also eyes the a7-pawn)
28 ... �g8! (after breaking with ... g4, this
is a typical move to avoid any na3-h3+
business) 29 nff3 (29 il.xg4 il.xg4 30
il.xas 'it'g s 31 nf2 nh1 32 tllx c7 nxas
33 'it'xas 'it'h4 34 nf1 'it'h2+ 3S �f2 f3
gives Black a winning attack according
to Mikhalevski) 29 ... il.e7 30 nh3 tllf6 (or
30 ... tlle 3 31 nhxe3 il.h4 32 g3 fxe3 33
nxe3 'it'g s 34 nf3 'it'g6 with a clear ad­
vantage - Mikhalevski) 3 1 iLfl ! ? il.xh3
32 nxh3 'it'c8 33 'it'h4?! �f7 was much
better for Black in S.Krivoshey­
G.Schebler, German League 2004.
22 ... cxd6 23 nc1 tllh 6 24 nc4
After 24 'it'c2 tlle 8 2 S 'it'b3 (instead
2S 'it'd3 ng1 26 nc2 �g8 27 'iVdl looked
very strange in A.Bachmann-A.Frolov,
Berlin 1994, and here Black should
probably continue 27 ... g4 28 hxg4 hxg4
68
2 9 fxg4 tllf6) 2 s ...ng1 2 6 nc6 il.d7 27
il.d2 il.xc6?! 28 dxc6 'it'b6 29 nc1 g4 30
fxg4 il.e7 31 'it'e6 was good for White
in P.H.Nielsen-M.Golubev, Internet
(blitz) 2004. Instead Golubev suggests
27 ... g4, while I think Black could also
investigate 27 ...tllf6!?.
The text reaches a critical position.
24...g4!
This is better than 24...ng7?! 2s 'it'c2
g4 which runs into 26 il.xas!. After
26 ... 'it'e8 (even worse is 26 ...'it'xas 27
nxc8 gxh 3 28 tllxh3, J.Fang-B.Dean
Kawamura, Parsippany 2008), White
should play 27 fxg4! hxg4 28 tllc 7 'it'g6
(or 28 ... nxa 29 il.xc7 gxh 3 30 tllx h3
tllh g4 31 il.b6) 29 tllx a8 gxh 3 30 il.f3!
when he survives the attack and keeps
a decisive material advantage.
Instead 27 h4? gxf3 28 il.xf3 il.g4 29
il.xg4 tllfxg4! 30 il.b6 'it'e7 31 tllxg4
tllx g4 32 il.f2 f3 33 g3? (this loses
quickly, but Black is still much better
after 3 3 'JJ..C7 fxg2 34 ne1 'iVf6 3 S nxg7
il.xg7 36 �Xg2 il.h6 OT 33 nc3 fxg2 34
na1 ng8 3S il.g3 il.h6 36 naa3 il.f4)
33 ... tlle 3 ! 34 il.xe3 nxg3+ 3 S �2 ng2+
 The M a r def Plata Va ria tio n : 9 tDe1 tb d 7 1 0 ii.e3

36 �Xf3 nxc2 37 nxc2 'ifxh4 38 ngl Aguirre, Sort 2006. Black should now
nxa4 39 tDc3 nc4 40 �e2 ii..h 6! 41 ii..b6 play 30 ... t:Dxfl 31 �xfl ng1 with good
nxe4+ and 0-1 was the well-known chances, such as after 32 t:Dxa8 ii.. a 6.
game P.H.Nielsen-V.Kotronias, Hastings
2003/04. 82) 15 b4
25 fxg4
After 2S 'it'c2 1'..d 7 26 tDC7 Black has
the shot 26 ...ii..x a4!.
25 ...hxg4 26 hxg4 ng1

This is a more direct approach from

White.
1s ...axb4
Black should certainly take now, be­
21 gs!? cause otherwise White will either cap­
Worse is 2 7 'it'c2 t:Dfxg4 28 t:Dxg4 ture on as himself or else even break
tDxg4 29 ii..x g4 ii.xg4 when Black has with cs immediately, when the f2-
good attacking chances. bishop remains very active. Now White
21.. Jbgs 28 'it'c2 has a further choice:
After 28 tDc7 na1 29 'it'c2 (or 29 tDe6
ii.xe6 30 dxe6 nag7 31 ii..f3 tDhg4 with E21: 16 "llbS
counterplay) 29 ...1'..d 7 30 tDe6 ii.xe6 31 E22: 16 "bxb4
dxe6 ng8 both sides have chances ac­
cording to Avrukh. E21) 16 t:Dbs
28...tDhs White is hoping to obtain a favour­
All of 28...t:Dfg4, 28...tDhg4 and able version of Line El by playing ii.el
28...ng8 deserve attention as well. and ii..xb4, but the early knight hop
29 ii..f3? gives Black an opportunity to acceler­
29 ii..xhs nxhs 30 na 1'..d 7 is un­ ate his attack and this method of play
clear. for White has a bit of a suicidal reputa-
29... tDg3 30 tDc7 tion. In this variation both sides must
This was S.Krivoshey-N .Ortiz try to break through quickly.

69
A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1
16...tllf6! 17 il.e1
Instead 17 tllxb4 looks misguided,
but matters are not so clear after
17 ... g4 18 il.h4 g3 19 hxg3 (after the 19
h3 of P.H.Nielsen-H.Harestad, Gausdal
1996, Black could consider 19 ...'it'd7! ?,
with the idea of ...tll g 6, or Gallagher's
19 ... il.d7) 19 ... tllg6 20 tllc6 �d7, which
is murky, E.Kobylkin-S.Soloviov, Alushta
2001.
17 ...g4!
Because the e4-pawn is loose, Black
does not need any further preparation
for his key kingside break and can ad­
vance at once.
18 il.xb4
70
18 fxg4 smacks of desperation, but
really it is just unclear: 18 ...tllxe4 19
il.xb4 il.d7 (or 19 ...tll g6 20 as tllcs 21
tllx cs bxcs 22 il.c3 e4 23 il.xg 7 �xg7 24
'it'c2 'it'e7 2S 'it'c3+ tll e s with the initia­
tive, A.Nikitin-S.Soloviov, Alushta 2002)
20 'it'c2 tll gs 21 h4 tlle4 22 il.e1 tllg 6 23
h s tll g 3 24 hxg6 'it'h4 2s tllf2 il.xbs 26
gxh7+ (26 cxbs e4 with an attack)
26 ... �h8 (26 ...'it'xh7 wins immediately
after 27 �xh7+ �xh7 28 il.d3+ e4 29
tllxe4 tllxfl, but the text is certainly
good enough) 27 cxbs e4 28 nd1 f3 29
gxf3 tllx e2+ 30 'it'xe2 exf3 0-1,
J.Chabanon-J.Degraeve, French League
1999.
18...g3
Another idea is 18 ...tll g 6 as sug­
gested by Golubev. Indeed, the further
19 as bxas 20 nxas l:tb8 21 l:ta7 g3 22
'it'c2 nf7 (better than 22 ...tllh s 23 nxa
gxh2+ 24 �xh2 'it'h4+ 2S �gl tll g 3 26
il.xd6) 23 nb1 gxh2+ (23 ... tllxds 24
exds e4 2S fxe4 nxbs 26 na8!) 24 �xh2
tllxdS !? 2S exds e4 certainly deserves
attention.
19 h3
Instead 19 as should be met with
19 ...bxas! (19 ...gxh2+ 20 �xh2 bxas 21
.l:txas nxas 22 il.xas tlle8 23 cs is better
for White according to Markos) 20
nxas (20 il.xas nxas 21 nxas c6)
20 ...l:txas 21 il.xas tlle 8 (with the idea
of ...tll g 6) 22 cs tll g6 23 'it'c2 'i!fh4 24 h 3
il.xh3 2S gxh3 'it'xh3 2 6 il.d1 tllh4 2 7
ne1 dxcs, which i s a little better for
Black according to lkonnikov in New in
Chess.
 Th e M a r def Plata Varia t i o n : 9 tDe 1 tb d 7 1 0 ii.e3
The text reaches a controversial po­
sition. Black's play is very far advanced,
but with the kingside closed he will
have to sacrifice to break through.
19 ii..x h3!
...
The assessment of this move has
gone back and forth from '?!' to '!'.
Black will clearly obtain an attack, but
will White be able to defend? For a
while I believed that this sacrifice was
not quite good enough, but upon fur­
ther investigation I believe it is indeed
the best, both from a practical and an
analytical point of view. We will see,
however, that Black's order of moves is
extremely important.
If Black does not sacrifice immedi­
ately, White may be able to hold the
kingside together: for example,
19 ... tDg6 20 ne1 tDe8 21 ii.fl tDh4 22
'it'e2 nf6 23 nebl ng6 24 tDel ii..f8 25
as gave White all the play in
D.Vigorito-S.Muhammad, Chicago
2007. Instead 19 ... tDe8 was praised by
Markos, but 20 'it'd2 (20 ne1 with the
idea of j,.f1, ne2, tDel and nea2 can be
met with 20 ...cs!?, trying to hold White
up on the queenside) 20 ... tDg6 (20 ... cs
was suggested by Markos) 21 ntc1 lDh4
22 tDel nf6 23 ii.fl should be some­
what better for White, even though
Black eventually made a sacrificial
breakthrough in N.Sulava-V.Neved­
nichy, Bad Worishofen 2000.
20 gxh3 'it'd7 21 'it'c2
This is the best defence. White will
try to defend along the second rank.
Instead 21 �g2? gets crushed: 21 ... tDg6
22 nh1 tDh4+ 23 �g1
23 ...t:Dxe4! 24 fxe4 f3 25 ii.. d 2 (or 25
ii..xf3 nxf3 26 tDel 'it'f7 27 t:Dxf3 t:Dxf3+
0-1, S.Ghane Gardeh-S.Krivoshey, Dubai
2006) 2s .. .f2+ 26 t:Dxf2 nxf2 27 l:th2
gxh2+! 28 �xf2 'it'xh3 29 'ikh1 nf8+ 30
�el 'it'g2 31 ii..f3 t:Dxf3+ and 0-1 in
R.Pogorelov-C.Matamoros Franco, Dos
Hermanas 2003.
21 'it'd2 is better, but unimpressive:
21 ... t:Dg6 (or perhaps even 21 ... 'it'xh3) 22
ii..dl ii..h6 23 tDel �h8 24 na2 ng8 25
'it'd3? (25 'it'g2 tDh4 26 'it'h1 is unclear ­
Gallagher) 2s ... 'it'xh3 26 .i:tg2 tDh4 27
'it'e2 ngs was winning for Black in
S.Krivoshey-Xie Jun, Linares 1997.
71
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
21...tllg 6!
This move is very important. The
obvious 21 ...'it'xh3 is not as good, but it
is worth looking at it to see some typi­
cal ideas for both sides. After 22 il.d1
tllg 6 23 'it'g2 'it'h6 practice has seen:
a) 24 'iVhl tllh4 25 tlle l?! (25 ne1!)
25...tllxe4! 26 tllg 2 was S.Koutsin­
A.Frolov, Kiev 1995, and now
26 ...tllxf3+! 27 il.xf3 'it'xhl+ 28 �xhl
tllf2+ 29 nxf2 (29 �gl e4) 29 ... gxf2 30
nt1 (30 il.e4 nxa4!) 30 ...nxa4 31 il.a3
nta8! (this seems even better than Gal­
lagher's 31 ..Axc4 32 nxf2) 32 il.b2
nxc4 33 nxf2 na2 should win for Black.
b) 24 tll x c7! tllh4 25 'it'd2! (but los­
ing is 25 'it'e2? tllxe4! 26 'it'xe4 tllg 2! as
given by Gallagher) is a key idea -
White lines up against Black's queen to
avoid the ... tllxe4 and .. .f3 bomb.
Indeed, 25 ... g2 (after 25 ... nac8 26
tlle6 g2 27 ne1 nxc4 both 28 tllxf8 and
28 tllf2 give White a big advantage) 26
ne1 tllxf3+ 2 7 il.xf3 'it'hl+ 28 �2 'it'h4+
29 �e2 tllxe4 30 il.xe4 f3+ 31 �dl f2
32 il.xg2 fxel'iV+ 33 tllx el 'it'g4+ 34 il.f3
�f3 3 5 tllxf3 nt8 36 na3 �f3 37 nxf3
72
'it'xf3+ 38 �c2 was winning for White
in C.Persson-G.Calzolari, correspon­
dence 1998.
Now back to 21...tllg 6!:
22 ntb1
This move is often recommended.
Instead:
a) 22 il.d1? loses to 22 ...tllM 23 tll e1
'it'xh3 24 tllxe7 (24 na2 tllh 5) 24...tllh 5!
25 tllxa8 g2 2 6 tllx g2 tllg 3 and 0-1 in
M.Dziuba-J.Czakon, Koszalin 2005.
b) 22 nfdl is similar to OUT main
line: 22 ...'it'xh3 ! (22 ...tllh4 23 tll e1 'it'xh3
24 iLfl at least staves off the first wave
of the attack) 23 il.f1 'it'h5 24 tllxa (or
24 tlle1 tllxe4 25 fxe4 f3) 24...tllxe4! 25
tllxa8 tllf2 wins, since after, for exam­
ple, 26 il.g2 'it'h2+ 27 �fl tllh4 28 tll e1
e4 White will be crushed.
22 .. 'it'xh3!
.
This move was mentioned by
Matamoros back in 2003 in New in
Chess, but it seems to have gone unno­
ticed. Weaker is 22 ...tllM 23 tll e 1 with a
clear advantage for White as given by
Markos.
23 il.f1
 Th e M a r def Plata Vari a t i o n : 9 llJ e 1 lLl d 7 1 0 ii. e 3
After 23 ii..d 1 Black has 23...ll:lh4!
(23 ... ll:lxe4 24 'it'g2) 24 ll:le1 ll:lxe4! with
familiar devastation.
23... 'it'hs 24 ll:le1
No better is 24 ll:lxc7 ll:lxe4! 2S ii..g 2
(otherwise, 2S ll:lxa8 ll:lgs 26 ll:lel e4
wins, but 2 s ... ll:lf2 may be even better:
for example, 26 ii..g 2 'it'h2+ 27 �fl ll:lh4
28 ll:le1 e4) 2s ... ll:lgs 26 ll:le1 e4 27 l:ta3
ll:lh4 and White cannot defend.
24...ll:lxe41 25 fxe4
2 S ii..g 2 ll:lf2 26 ll:lxe7 ll:lh4 winning.
2s ...f3 26 ll:lxf3 �xf3 27 'it'g2 ll:lf4 28
'it'h1 'it'g4 29 �b2 �f8 30 ll:lxc7 �f2
Black has a decisive attack - Mata­
moros.
E22) 16 ll:lxb4
While both lS ii.el and lS b4 axb4
16 ll:lbs are very ambitious, the simple
recapture with 16 ll:lxb4, while very
obvious, almost looks naive. The knight
cannot defend the kingside and it is
also seems harder for White to support
the advances a4-aS and c4-cS. How­
ever, as is often still playable and the
b4-knight may hop to c6. As always, a
lot will come down to specifics.
16...ll:lf6
Instead 16 ...ll:lcs 17 as! bxas
(17 ...�xas 18 �xas bxas 19 ll:ld3 is
similar) 18 ll:ld3 ll:lxd3 19 'it'xd3 gives
White a slight but clear edge according
to Markos. White will double rooks on
the a-file to win back the pawn and
Black's counterplay is not so easy.
17 ll:ld3
White has several alternatives:
a) 17 ll:lc6 is the most direct:
17 ...ll:lxc6 18 dxc6 'it'e8 (18 ...ii..e 6 is also
possible) 19 ll:lds �f7 20 as bxas 21 'it'a4
g4 22 'iYbs (or 22 1'..h4 ll:lxds 23 cxds g3
24 hxg3 fxg3 2S ii..x g3. as in P.Kiriakov­
D.Lobzhanidze, Groningen 1996, and
73
A ttacking Chess: Th e K i n g 's I n dian, Vo l u m e 1
now Black should go 2S ... il.h6) 22 ...tllxds
23 cxds g3 24 hxg3 fxg3 2s il.xg3 il.h6
gives Black counterplay.
b) 17 as bxas 18 tllc6 tllxc6 19 dxc6
'it'e8 20 cs dxcs 21 il.xcs nf7 22 'it'b3?!
'it'xc6 23 'it'bs 'it'xbs 24 il.xbs il.f8 was
better for Black in O.Wilgenhof­
F.Nijboer, Dieren 2009.
c) 17 il.el and then:
cl) 17... hs 18 h3 nf7 (worse is
18 ...�h8?! 19 tlld 3! tlle g8 20 tllf2 tll h6
21 tllbs .l:tf7 22 as bxas 23 nxas nxas
24 il.xas il.f8 2S cs!) 19 tll d3 il.f8 20
tllf2 is a little better for White.
c2) 17 ...'it'd7 18 tlld3 tllg 6 19 tllf2 h S
2 0 h3 tllh4 21 a s ! nf7 (or 21 ...nxas 2 2
nxas bxas 23 tllb s) 22 axb6 nxal 23
'it'xal cxb6 24 'it'b2 'it'd8 2S tlla4 looks
shaky for Black, although he drummed
up some counterplay after 2S ...g4 26
fxg4 f3 27 gxf3 tll h7 28 'it'xb6 'iVf6 in
O.Gladyszev-S.Soloviov, Soc hi 200S.
d) 17 na3 il.d7 18 tll b s was Yusu­
pov's choice and has been endorsed by
Markos.
Now Black has:
dl) 18 ...�h8 19 'it'al! (19 il.el ng8
74
gave Black counterplay in A.Yusupov­
G.Kasparov, Yerevan Olympiad 1996)
19 ...nas (or 19 ... g4 20 as gxf3 21 il.xf3
bxas 22 nxas nb8 23 l:ta7) 20 tlld3 g4
21 il.el na6 22 as is a line given by
Markos. White's play looks faster.
d2) 18 ... g4!? and then:
d21) 19 fxg4 tllxe4 20 il.d3 tllxf2 21
l:txf2 il.f6 is level according to Yusupov.
d22) 19 il.el gxf3 20 il.xf3 �h8 21
tlld3 tllg 6 22 as bxas 23 il.xas il.xbs 24
cxbs 'it'd7 2S 'it'b3 nfb8 26 nb1 tllh4
gave Black counterplay in M.Bley-
1.Novak, correspondence 200S.
d23) 19 il.h4 g 3 ! ? (Black could also try
19 ... gxf3 20 il.xf3 tll g 6 21 il.f2 'it'c8 or
19 ...nf7!? with the idea 20 il.xf6 gxf3
21 il.xe7 fxe2 22 il.xd8 exdl'it' 23 nxdl
nxd8 24 tllxe7 nc8) 20 hxg3 tllg 6 was
suggested by Yusupov. Then 21 tllc6
il.xc6 22 dxc6 tllxh4 23 gxh4 tllh s 24 as
is given by Markos, but after 24 ... bxas
2S cs tllg3 I think I would prefer Black.
11 hs
...
Black should also consider
17 ... tllg 6!? with the idea 18 tllbs g4!.
1s tt::i bs tllg6
After 18 ...il.d7 19 il.el 'it'b8 20 �hl
g4 21 as bxas 22 il.xas 'it'b7, as in
S.Ghane Gardeh-V.Kotronias, Cappelle
la Grande 2008, Markos suggests 23
'it'c2 ntc8 24 nfbl when White is way
ahead. Black could, though, consider
18 ...g4 19 il.h4 gxf3 (after 19 ... g3 20
hxg3 tllg6 21 as tllxh4 22 gxh4 Black
does not have ...tllh s available, so it is
not so easy to attack) 20 il.xf3 tllg4
with counterplay.
 Th e M a r def Plata Va ria t i o n : 9 tDe1 tb d 7 1 0 ii.e3
19 a 5
Instead 19 'if el is interesting, taking
a look at both the as- and h4-squares:
19...g4 20 as h4! ? 21 ii..xh4 (or 21 axb6
nxal 22 'it'xal g3 23 ii.el h3 21 fxg4
tDxe4 22 axb6 .l:txal 23 'it'xal cxb6 with
counterplay - Golubev) 21... tDxM 22
'it'xh4 nxas 23 lDb4 was E.Guseinov­
M.Golubev, Moscow 2006. Now 23 ...1'..d7
24 t:Dc6 ii..xc6 2S dxc6 'it'a8 26 tDxe7 (or
26 nxas bxas with the idea of 27 cs?!
'it'xc6 28 1'..c4+? dS 29 exds 'it'xcs+, win­
ning) 26 ...'it'xc6 27 tDe6 ne8!
(E.Guseinov) 28 t:Dxg7 �xg7 29 fxg4?!
t:Dxe4 gives Black the initiative. The
threat is 30 ... 'it'cs 31 �hl tDg3!.
19. bxas 20 cs
••
20 ii.el g4 21 nxas .l:tb8 22 lDb4 nb7
23 tDc6 'it'e8 24 na8 gave White a big
advantage in Z.Kozul-D.Rogic, Bled
1997, but Black should try 20...a4 or
21 ...1'..d 7.
20 g4 21 'it'c2 g3! 22 hxg3
•••
After 22 ii.el t:De8 23 hxg3 fxg3 24
ii..x g3 h4 2S ii..f2 'it'gs Black has attack­
ing chances.
22 fxg3 23 ii..x g3 h4 24 ii.el tDhs
•••
Dembo also mentions 24...h 3 ! ?.
2 s nxas?!
After 2S cxd6! cxd6 26 tDc7 nb8 27
tDe6 ii.xe6 28 dxe6 tDg3 the position is
very messy.
2s nxas 26 ii.. x as ii..h 6!
•••
Black has excellent attacking chances.
27 ii..d 2
Dembo also gives the lines 27 t:Df2
h3, 27 t:Dxd6 h3!, 27 tDxe7 ii..e 3+ 28 �h2
'it'gs, 27 cxd6 ii.. e 3+ 28 �h2 'it'gs and 27
ii.xc7 ii..e 3+ 28 nf2 (or 28 tDf2 'it'gs 29
ii.. x d6 tDgf4 30 g4 hxg3) 28...'it'gs. In all
cases Black retains a strong attack
27 ii..x d2 28 'it'xd2 t:Dgf4 29 tDxf4
•••
If 29 cxd6? 'it'gs 30 t:Dxf4 t:Dxf4 threat­
ens both mate on g2 and 31...tDh3+. After
31 �2 'it'xg2+ 32 �e3 (or 32 �el ii..h 3,
opening the way for Black's rook to a8)
32 ...cxd6 Black's attack is decisive.
29 tDxf4
•••
Black is much better. Indeed, after
30 .l:tdl 'it'gs 31 ii.fl 'it'g3 32 nc1 tDh3+
33 �hl t:Df2+ 34 �gl h3 White re­
signed in K.Nikolaidis-Y.Dembo, Corfu
2007, in view of 3S 'ikh6 hxg2 36 ii..x g2
Ji.h3 OT 3S 'ifxf2 h2+.
75
 Chapter 4
The Mar del Plata Variation
9 '2Jd2
1 d4 ll:lf6 2 c4 g6 3 ll:lc3 il.g7 4 e4 d6 5
ll:lf3 o-o 6 il.e2 es 7 o-o ll:lc6 8 ds ll:le7 9
ll:ld2
By playing 9 ll:ld2 White takes a very
optimistic view towards his queenside
play. This line was the main battle­
ground in the early 1990s until 9 b4
experienced a tremendous boom in
popularity. Like 9 ll:lel, 9 ll:ld2 prevents
Black's ... ll:lhs and at the same time
prepares to bolster his centre with f2-
f3. In this case, however, instead of re­
routing the knight to help support the
c4-c5 break (for example with ll:le1-d3),
76
White intends to force through c4-c5,
even as a pawn sacrifice, after which
the d2-knight will take up a tremen­
dous post at c4. From there it will pres­
sure d6, while its control of b6 often
ends up being important as well.
At first 9 ll:ld2 looks more active
than 9 ll:lel, but if White cannot effec­
tively reposition the knight, White's
development can become a bit clogged
up. Logically Black should look to pre­
vent White from playing cs and ll:lc4
effectively. In that sense 9 ... cs might be
considered Black's most natural re­
sponse, blocking the position. At one
point this was, indeed, the main line,
but after 10 �bl ll:le8 11 b4 b6 12 bxcs
bxcs 13 ll:lb3 White's knight finds use­
ful employment via another route. For
example, 13 .. .fs 14 il.g s! h6 15 il.xe7
'Wlxe7 16 ll:las and the knight becomes
a menace on c6.
Another idea is to play a 'straight
race' with 9 ...ll:ld7 (or 9 ...ll:le8 - the two
moves will generally transpose after

the knight returns to f6) 10 b4 fS 11 cs
�f6 12 f3 f4 13 tllc4 gs 14 a4 ( 14 il.a3 is
also possible) 14...tll g 6 lS il.a3 l:tf7.
Here both Beliavsky's 16 as and the
main line, 16 bS, are theoretically very
dangerous, although Black has had
some recent successes, especially in the
later variation after 16 ... dxcs! 17 il.xcs
hS 18 as g4 when he has dangerous
counterplay.
9 as
...
This is established as the main line.
Black prevents b4 and in a sense he
gains a tempo, because White will have
to play both a3 and l:tbl in order to
achieve the b4-advance. In the 9 tllel
line, this move would be ineffective be­
cause White could play something like
a3 and il.d2 when the al-rook is pro­
tected by very natural means. After 9
tlld2, however, the d2-knight prevents
the cl-bishop from moving, so White
does not have this plan available.
10 a3
White can also play 10 l::tb l, when we
can still play 10 ... il.d7. After 11 b3 (11 a3
a4) Black can play 11 ...tll c 8, 11 ...tll e 8 or
The M a r def Plata Vari a t i o n : 9 tll d 2
11...c6 when in each case White does
not really have anything better than 12
a3, which will just transpose back into
our main lines, below.
After 10 a3 Black faces a very impor­
tant strategic decision. First of all, he
must avoid the race. Even though 9 ... as
has gained a tempo of sorts, it has
weakened Black's queenside and after
the analogous 10 a3 tlld7 11 l::tb l fS 12
b4 tllf6 13 f3 f4 14 cs axb4 lS axb4 gs
16 tllc4 tllg 6 17 bS! (worse is 17 il.a3
�f7) 17...dxcs 18 b6! Black's queenside
is destroyed before he could even
dream of a kingside attack. The block­
ing plan with 10 ... cs is more respect­
able, although after 11 l:tbl tlle 8 12 b4
axb4 13 axb4 b6 14 bxcs bxcs lS tllb3
White again does pretty well. Black
should still consider this structure in
some cases, though, and we will allow
it sometimes if we can obtain a rela­
tively favourable version.
Traditionally the big choice has
been between 10...tlld7 and 10 ... il.d7.
Originally I had hoped to include both
lines, but then decided that, with 9
tlld2 being considered less critical than
9 tllel and 9 b4 these days, one main
line should suffice. I've always kind of
liked the move 10...il.d7, because Black
may actually play on either side of the
board. Instead 10 ...tlld7 actually mir­
rors White's developmental logjam, but
because 10 ...tlld7 is twice as popular
and a more traditional 'attacking'
move, I began to delve into its intrica­
cies first.
77
A ttacking C h e s s : Th e King 's I n dia n , Vo l u m e 1
I decided to avoid 10 ... ll:ld7 because
of 11 nb1 fS 12 b4 �h8
13 f3 (after 13 'it'c2, I do not like
13 ... tllg 8 14 exfs gxfs 1s f4, but
13 ... tllf6! 14 f3 axb4 l S axb4 c6! is an
effective way for Black to deal with
White's threat to play cs and tllc4, and
16 dxc6 ll:lxc6 has proved to be satisfac­
tory for him) 13 ...f4 (after 13 ... axb4 14
axb4 c6 lS dxc6! ? ll:lxc6 White has
gained as his queen has not gone to c2
and Black's knight is still on d7, so he is
quicker to attack the d6-pawn) 13 .. .f4
(after 13 ...tll g 8 lvanchuk's 14 'it'c2 ll:lgf6
l S il.d3 ! forces 1S ...f4 when Black's
knights are suddenly misplaced).
78
This line was Gallagher's recom­
mendation back in 2004 and it was
played successfully by Kasparov. How­
ever, the more I delved into White's
possibilities, the more disheartened I
became. White can prepare the cs­
break in various ways, such as with 14
tllb3, 14 tlla4 and 14 tll bs, and in fact
even the immediate 14 cs!? is danger­
ous. As it turns out, even if Black tries
to prevent this break with ...b6, White
can almost always sacrifice a pawn
with c4-cS anyway. Once that knight
gets to c4, White obtains excellent play.
Although Black has not done so
badly in practice with 10 ... tlld7, while
studying the relevant games I found it
much easier to find improvements for
White than for Black Black's pieces all
seem to be a bit jumbled up. I once
came across a comment by Golubev
that said something to the effect that
he was not convinced that combining
... as and ... tlld7 had to be logical, and I
have come to understand this reason­
ing. After all, the 'logical' 9 ... cs and
9 ... as 10 a3 cs do not stop White's ini­
tiative even though they prevent the
c4-cS advance. While 9 ... as and
10 ...tlld7 seems to discourage White's
breakthrough, they do not prevent it.
So after a lot of time spent on
10 ... tlld7, I felt I had to abandon it, al­
though I think an understanding of
these positions is still beneficial, espe­
cially the idea of playing a pre-emptive
strike with ... c6. However, when I came
around to investigating 10 ... il.d7, I was

quite pleased with the play that arose.
10 il.d7
...
This move not only develops a piece,
but in effect it gains another tempo
because White usually chooses to pre­
vent Black from damaging his pawn
structure with ... a4.
11 b3
This by far the most common move,
preventing Black from locking down
the queenside with ... a4. It is also pos­
sible to ignore Black's idea:
a) 11 .l:tbl a4 12 b4 axb3 13 tllxb3
(after 13 .llxb3 b6 14 a4 Black should
not bother with 14 ... il.xa4 lS tllxa4
.llxa4 16 .l:txb6!, but instead just play
14 ... tlle8 lS .lla3 fS with counterplay)
13 ...b6 (alternatives are 13 ...il.a4, 13 ... c6
and 13 ... cs 14 dxc6 il.xc6 lS 'it'd3 tll e 8!
16 .lld l fS) 14 .ll al (White is ready now
to push the a-pawn, but he has wasted
a lot of time) 14 ... tlle 8 (stopping the a­
pawn with 14...'it'e8!? is also an attrac­
tive idea) lS a4 fS and now:
al) 16 as bxas 17 tllxas (after 17
.llxas .llx as 18 tll x as 'it'a8 19 tll b 3 tllf6
20 il.d3 .llb8 Black already had a nice
Th e M a r def Pla ta Varia tio n : 9 tll d 2
initiative i n L.Brunner-J.Nunn, Nurem­
berg (rapid) 1990) 17 ...tllf6 is fine for
Black. White's queenside play has not
really gone anywhere.
a2) 16 f3 tllf6 (alternatives are
16 ...�h8 and 16 ... cs) 17 il.e3 cs!? (or
17 ... tllh s ! ?) 18 dxc6 il.xc6 was fairly
level in E.Gleizerov-1.Nataf, Montecatini
Terme 1997.
b) 11 .ll a2 hopes to obtain an im­
proved version of variation 'a' by main­
taining the rook on the a-file: 11 ... cs (I
still think 11 ...a4 is playable and after
12 b4 axb3 13 tllxb3 Black could inves­
tigate 13 ... c6, 13 ... cs. and 13 ... b6) 12
dxc6 (after 12 b3 tlle 8 13 .llb2 fS 14 b4
axb4 lS axb4 b6 White has wasted a lot
of time, so Black should be fine)
12 ...bxc6! (this is an ideal circumstance
for this recapture, as White won't re­
gain control of the dS-square) 13 b4
axb4 14 axb4 .llxa2 lS tllxa2 'it'a8 (also
possible are 1s ...'it'b6!? and 1s .....lfi.e6!?)
16 tllc3 dS and Black had the initiative
in G.Flear-M.Gurevich, Tel Aviv 1989.
Instead after 11 b3 White is all set
to play 12 llbl and 13 b4.
79
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
One thing I like about the 10 ... 1'.d7
line is the fact that Black's play is flexi­
ble and there are a few decent options
here:
A: 11 2t)c8
...
8: 11-.lhel
C: 11 c6...
It is also possible to play 11 ... cs, be­
cause Black will end up a tempo ahead
of the line 10...as 11 a3 cs as White will
have to spend two moves with his b­
pawn. Still, after 12 hi.bl ll:ie8 13 b4
axb4 14 axb4 b6 lS bxcs bxcs 16 ll:ib3
(16 'i¥b3 ll:ic8! exploits Black's extra
tempo by covering b6) 16 ...fs 17 f3 lLif6
18 ii.d2 f4 19 hi.al hi.xal 20 'ifxal g s 2 1
Va s White has the initiative. I n general
Black should be willing to accept this
structure, but only on his own terms.
Here even the extra tempo does not
guarantee Black equality.
A) 11 ll:icS
...
I have always been attracted to this
funny-looking move. The idea is to play
80
...ll:ib6 which will allow Black to meet
b4 with ...ll:ia4. Exchanging this knight
not only provides Black with some re­
lief from his spatial disadvantage, but
he may sometimes even take the initia­
tive on the queenside. We will see a
similar plan frequently in the Petrosian
Variation, although there Black plays
...ll:ia6-cS-a4. Black may also initiate
play with ... c6. If White is not careful he
can very easily lose the initiative on the
queen side.
12 hi.bl
This is the natural move, but there
are others:
a) 12 l:ta2 always looks strange to
me. If White plays b4 soon he will end
up with a knight on a2, which hardly
seems ideal for him. After 12 ...ii.h6
(this is a typical idea for Black in the 9
ll:id2 variation) 13 'ifc2 ll:ib6 14 htdl
'i¥e7 lS 'ifd3 ll:ie8 16 h!.c2 fs Black is al­
ready comfortable. White felt the need
to complicate with 17 cs dxcs 18 ll:ic4
ii.xcl 19 d6 ll:ixd6 20 ll:ixb6 fxe4 21
ll:ixe4 ii.ts 22 'ifds+ �g7 23 l:tcxcl
ii.xe4 24 'ifxcs cxb6 2S 'i¥xd6 'i¥xd6 26
h!.xd6 ii.c6, but then Black was just up a
pawn in S.Skembris-V.Kotronias, Xanthi
1991.
b) 12 ii.b2 ll:ib6 (12 ... cs and 12 ...ii.h6
have also been tried) 13 'ifc2 'ife7
(again 13 ... ii.h6 is possible) 14 b4 ii.h6
(or 14... axb4 lS axb4 ii.h6 16 hi.fdl c6!?)
1S ll:ib3 axb4 16 axb4 ll:ia4 17 ll:ias (17
ll:ixa4 ii.xa4 is fine for Black) 17...ll:ixb2
18 'i¥xb2 h!.ab8 19 bs ii.e8! (Black is
ready to block the queenside with

...ll:id7 and this causes White to panic)
20 b6?! cxb6 21 ll:ib3 (21 'i¥xb6 ii.d2
wins the e-pawn) 21 ... 'ifc7 left Black
with an extra pawn and the bishop­
pair in S.Kishnev-B.Damljanovic Sibenik
1990.
c) 12 'i¥c2 ll:ib6 13 a4! ? sees White
switch plans, hoping to make the b6-
knight look silly. He may continue with
ii.a3 and b4. Still, this is time­
consuming and freezing the queenside
looks strange.
Black has:
cl) 13 .. lte8 14 ii.a3 ii.f8 1s ll:i bs cs?!
(1s...ii.xbs 16 cxbs ll:ibd7 is better, but
after 17 b4 White still has an edge) 16
dxc6 ii.xc6 17 Itfdl 'i¥b8 18 ll:ic3 ll:ibd7
19 ll:if1 ll:ics 20 ii.xcs! dxcs 21 ll:ie3 gave
White a positional advantage in
D.Komarov-Kr.Georgiev, Lyon 199S.
c2) With 13 ...ll:ic8 Black is satisfied
that the knight has done its job and
now prepares to block the queenside:
14 ii.a3 b6 lS b4 axb4 16 ii.xb4 cs 17
dxc6 ii.xc6 18 ll:ids ll:id 7 19 1'.g4 ll:ics 20
ii.xcs bxcs 21 ll:ib1 ll:ia7 (or 21 ...ll:ib6! ?
when 22 ll:ibc3 fails to 22 ...ll:ixc4 and 2 2
Th e M a r d e f Plata Va ria t io n : 9 lLi d2
ll:ixb6 'ifxb6 23 ll:ic3 Vas is fine for
Black) 22 ll:ibc3 ii.e8 23 ll:ibs ll:ic6 24
ll:idc7 Itb8 2S ll:ixe8 Itxe8 26 'i¥d3 ll:id4
27 ll:ixd4 exd4 and Black was at least
equal in l.Cosma-B.Damljanovic, Bel­
grade 199S.
c3) 13 ...ii.h6 ! ? is untried, but 14 1'.a3
'i¥e7 lS b4? would run into 1S ... ll:ixa4!,
exploiting the loose knight on d2.
12 ll:ib6
...
13 'i¥c2
After 13 b4 axb4 14 axb4 ll:ia4 lS
'i¥c2 Black has:
a) 1s ...ii.h6 16 ll:ibs ii.xbs (16...'ife7?!
17 cs! is variation 'b21' to Black's 13th,
below) 17 cxbs ll:ib6 18 ll:ib3 ii.xcl 19
Itbxcl ll:ie8 20 ll:ias Itb8 was perhaps
slightly better for White in
D.A nagnostopoulos-1.Nikolaidis, A no
Liosia 199S, but Black is very solid.
b) 1s ... ll:ixc3 16 'ifxc3 ii.h6 (this is
logical, but Black could consider 16 ...c6,
16 ...h!.a2 or 16 ...'ife7) 17 ii.b2 ! ? (White
intends f4, so Black is obliged to grab a
pawn) 17 ... ii.xd2 18 'i¥xd2 ll:ixe4 19 'i¥e3
fs 20 f4 exf4 21 h!.xf4 gives White com­
pensation for t he pawn.
81
A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1
13 'i¥e7
•..
Black connects his rooks and covers
the cs-square. He also has:
a) 13 ... ll:ie8!? 14 b4 axb4 lS axb4
ll:ia4 16 ll:ixa4 ii.xa4 17 'if c3 ii.d7 18 cs
fs 19 f3 ii.h6 (better is 19 ...ll:if6) 20 c6
and White was ahead in K.Lerner­
G.Pieterse, Amsterdam 1988.
b) 13 ... ..ih6 and here:
bl) 14 b4 axb4 lS axb4 ll:ia4
(1s ... 'i¥e7 16 c s ! ) 16 ll:ibs leads back to
13 b4, above, while avoiding 1S ... ll:ixc3,
although this is no special accom­
plishment for White.
b2) 14 ll:ibs 'i¥e7 lS b4 (White can­
not take on c7 because of ... l:tfc8, so he
just continues his queenside play)
1S ... axb4 16 axb4 and here:
b21) 16 ... ll:ia4?! 17 cs! is a trick Black
must watch out for. After 17 ... dxcs 18
ll:ixc7 l:tac8 19 ll:ic4! he was clearly
worse in G.Kamsky-V.Spasov, Tilburg
1992.
b22) 16 ...h!.fc8 17 cs ll:ia4 was
L.Psakhis-V.Spasov, European Team
Championship, Debrecen 1992, and
now the simple 18 lLif3 ii.xcl 19 l:tfxcl
gives White the initiative.
b23) Best is 16 ... c6!, a typical pre­
emptive strike: 17 dxc6 ii.xc6 18 ll:ic3
l:tfc8 19 ll:ib3 ii.xcl 20 hi.fxcl ll:ia4 was
fine for Black in Wu Shaobin-M.Al Mo­
diahki, Beijing 199S.
14 b4
This is consistent. Instead 14 ii.b2
ii.h6 lS l:tbel 1'.g4!? is possible, which
is an idea we will see too in the Petro­
sian Variation. After 16 ii.xg4 ll:ixg4 17
82
a4!? ll:id7 18 ll:if3 ll:ics 19 ii.a3 b6 20
ii.xcs bxcs 21 g3 ll:if6 the position was
fairly level in M.Gurevich-1.Smirin,
Haifa 199S.
14...axb4 15 axb4
1s ...ll:ia4
Or 1S ... l:tfc8 16 ll:ib3 ll:ia4 17 ii.d2
ll:ixc3 18 ii.xc3 c6 19 dxc6 ii.xc6
(19 ...bxc6 !?) 20 f3 dS! 21 cxds ii.xds 22
'i¥b2 ii.e6 23 hi.al ll:id7 with equality in
A.Beliavsky-Ki.Georgiev, Groningen
1994.
16 ll:ixa4
White acquiesces to the exchange
of knights. Instead 16 ll:id1?! is too pas­
sive. After 16 ...ll:ig4!? (16 ... c6 also looks
good) 17 f3 (or 17 h3 ll:ih6 with the idea
of ... fs) 17 ...ll:if6! 18 ll:ib3 ll:ihs (the point
of Black's manoeuvres) 19 ll:ias ll:if4!
Black had the initiative in
R.Kasimdzh anov-1.Smirin, Elista Olym­
piad 1998.
After 16 cs Black should avoid
16 ... ii.h6 17 ll:ibs! and play either
16...h!.fc8 or 16 ... ll:ixc3 17 'it'xc3 c6!?,
with the idea of 18 cxd6 'it'xd6 19 lLic4?!

16 ... ii.xa4 17 'ifc3
After 17 'it'd3 the moves 17 ... c6,
17 ...l:tfc8, and 17 ...1'.d7 all look reason­
able.
11...ii.h6
Pressuring the e4-pawn. 17... c6!? is
another idea.
18 'i¥d3
Instead 18 f3 is met by 18 ...ll:ih5!.
White can again consider sacrificing
a pawn with 18 ii.b2 ii.xd2 19 'i¥xd2
ll:ixe4 20 'it'e3 f5 21 f4, although this is
somewhat speculative.
18 ... c6 19 l2Jf3
19 ...1'.g7!?
Black keeps pieces on the board. Af­
ter 19 ...ii.xcl 20 l:tfxcl Mikhail Gure­
vich believes White is better, but Black
looks okay to me.
20 ii.gs cxds 21 cxds h6 22 ii.d2 1'.d7
Black could also consider playing
22 .. JUc8 immediately.
23 h!.bcl 'ifdB 24 'it'b1
24 b5 can be met by 24...h!.a4! when
the e4-pawn is in trouble.
24... ll:ihs 2s g3 1'.g4
Not 25 ... f5? 26 ll:ih4.
Th e M a r def Plata Va riati o n : 9 lLid2
26 �g2 'ifd7 21 ll:ig1 ii.xe2 28 ll:ixe2 fs
Trading light-squared bishops is not
always fatal in the King's Indian if Black
can generate counterplay. Here Black
could also consider 28 ... 'ifb5 29 ll:ic3
'i¥c4 with equal chances.
29 ll:ic3 ll:if6 30 f3 l:tfc8 31 'i¥d3 l:ic7 32
ll:ie2 h!.xcl 33 h!.xcl h!.a2 34 h!.c2
After 34 ll:ic3 fxe4! 35 fxe4 l:tb2 36
l:tbl l:txbl 37 'ii' x bl 'i¥c7 38 'it'd3 'ifb6
Black has counterplay. His queen is ac­
tive and the e4-pawn will always need
attention. Given time, Black can also
consider something like ...h 5, ...�h7 and
...ii.h6.
34...fxe4 35 fxe4 'if a4
Black had sufficient play in
M.Gurevich-1.Smirin, Elenite 1994.
B) 11...ll:ieS
This typical King's Indian move has
recently garnered some attention.
Black simply prepares .. .f5.
12 hi.bl f5 13 b4
Instead 13 f3 is well met by
13 ... ii.h6, while Black could also con­
sider 13 .. .f4.
83
A ttacking C h e s s : The King 's I n dian, Vo l u m e 1
13 ...axb4 14 axb4 ll:if6
This is by far the commonest move,
but Black could also consider 14...�h8
1s cs ll:ig8 16 f3 ll:igf6 17 ll:ic4 ll:ihs or
14... ii.h6 lS cs fxe4 16 ll:idxe4 ii.xcl 17
'ifxcl ll:ifs.
15 cs
The most direct. Instead lS f3 c6!?
(or just 1 s...ii.h6), and here:
a) 16 cs!? cxds 17 exds ll:iexds 18
ll:ixds ll:ixds 19 1'.c4 ii.e6 20 ll:ie4 (this is
the point of White's play, but it does not
lead to an advantage) 20...fxe4 21 ii.xds
'ifd7 22 'tlfb3 ii.xds 23 'ifxds+ �h8 24
'ifxe4 ds 2 S htdl l:tad8 was level in
C.Bemard-A.David, Andorra 1996.
b) 16 dxc6 ll:ixc6 17 ll:ibs (17 ll:ib3
ii.e6 18 ll:ibs fxe4 19 ll:ixd6 exf3 was
unclear in E.Gleizerov-1.Zaitsev, Podolsk
1992) 17 ... ii.e6 18 'it'c2 'ife7 (or 18 ... l:tc8,
as in R.Scherbakov-J. Nunn, H astings
1993/94) 19 ii.d3 ll:ihs (both 19 .. .f4!?
and 19 ...ll:ie8 ! ? are possible too) 20 cs
dxcs 21 'ifxcs was V.Chuchelov-J.Nunn,
Leeuwarden 199S, and now 21 ...ll:if4
would give Black a slight initiative.
c) 16 �h1 �h8 17 ii.b2 h s 18 c s ! ?
84
(the complications begin) 18...fxe4 19
fxe4 dxcs 20 d6 cxb4 21 dxe7 'ifxe7 22
ll:ia4 ll:ig4 23 ll:ib6 ll:if2+ 24 h!.xf2 h!.xf2
2 s 'it'e1 (2s ll:ixa8 'ii'g s!) 2 s ...l:taf8 26
ll:ixd7 'ii'g s ! 27 g3 hi.8f7 28 ll:ics 'it'e3 29
1'.c4 was R.Nolte-W.So, Tagaytay City
2010. Now 29 ... 'ifxcs 30 ii.xf7 hi.xf7
would be unclear.
15 ...fxe4
Black gives up the e4-square in or­
der to generate active piece play him­
self. There are a couple of alternatives:
a) 1s ...�h8 16 f3 ll:ihs 17 ll:ic4 ll:if4
18 ii.e3 ll:ic8 (Black covers the d6-pawn
and intends ...'if gs with counterplay)
19 �hl! g s 20 ll:ias (20 b s also looks
good) 20...'ife8 21 ll:ixb7 h!.a3 was
M.Gurevich-B.Gelfand, Munich 1993.
Now 22 l:tcl seems to refute Black's
play because after 22 ... h!.xc3 23 h!.xc3
ll:ixe2 24 'ifxe2 ii.bs 2S h!.c4 Black can­
not exploit the pin and White will re­
main a good pawn up.
b) 1s ... ii.h6!? is Black's latest at­
tempt to fight for the initiative: 16 ll:ic4
(16 ii.b2 and 16 f3 are possible as well)
16 ... ii.xcl (perhaps Black should prefer

16 ... ll:ixe4!? 17 ll:ixe4 1'.a4 18 'ifel ii.xcl
19 l:txcl fxe4) 17 'ifxcl (after 17 h!.xcl
ll:ixe4 18 ll:ixe4 fxe4 19 cxd6 cxd6 20
ll:ixd6? is not possible because of
20 ... ii.a4) 17 _.fxe4 18 cxd6 and here:
bl) 18 ...ll:iexds 19 ll:ixds ll:ixds 20
ll:ixes ii.e6 21 dxe7 'ifd6 (no better is
21...'ifxe7 22 'i¥xc7 ll:ixc7 23 l:tbcl) 22
'i¥b2 and White is more comfortable.
b2) 18 ...ll:ifs! ? and then:
b21) 19 ll:ixes?! cxd6 20 ll:ic4 l:tc8
(20... bs!?) 21 'if d2 hi.e8 22 l:tfel and
now 22 ... e3! 23 ll:ixe3 ll:ixe3 24 fxe3
'ifb6 2 S ii.f3 hi.c4 gave Black excellent
play for the pawn in A.Khalifman-
1.Cheparinov, Amsterdam 2007.
b22) 19 dxe7 'i¥xc7 20 d6! ? (20 l:tdl
e3! 21 ll:ixe3 hi.fc8 22 ll:ic4 bS is not
clear, but both 20 'i¥d2 and 20 'i¥b2 look
somewhat more pleasant for White)
20 ... ll:ixd6 21 ll:ixes ii.ts? (21 ... hi.ac8) 22
ll:ixe4! won a pawn in V.Chuchelov­
J.Degraeve, German League 2008.
16 ll:idxe4 ll:ifs 17 ii.gs
White is willing give up his dark­
squared bishop to consolidate his grip
on e4.
Th e M a r def Plata Va ria t io n : 9 lLid2
11...h6
This natural move has always been
played, even though 17 ... 'i¥e7 and
17 ... 'i¥e8 look reasonable. In fact I be­
lieve the strongest move is 17 ... ll:ixe4!
18 ll:ixe4 (18 ii.xd8 ll:ixc3 19 'if el ll:ixe2+
20 'i¥xe2 l:taxd8 gives Black three won­
derful minor pieces for the queen)
18 ... 'i¥e8, which looks fine for Black,
while here 18 ... 1'.a4!? may be possible
as well.
1s ii.xf6 ii.xf6 19 bs
White's most recent try was 19 hi.al
and 19 ... l:txal 20 'ifxal ii.gs?! 21 'ifa7
was much better for him in A.Beliavsky­
Z.Lanka, Austrian League 2007. Black
can improve with 20 ... ll:id4 21 ii.d3 ii.el
when White has a small edge at best
after 22 'i¥a2.
19...b6 20 c6 ii.cs 21 ii.d3 ii.gs 22 l:ta1
l:txal 23 'ifxal
23 ... hs
This is Nunn's recommendation. In­
stead 23 ...1'.f4?! 24 ll:ie2 'ifh4 was
G.Kamsky-J.Nunn, Monaco (blindfold)
199S, when Nunn gives 2S g 3 ! 'ifh3 26
ll:ixf4 exf4 27 hi.el! ll:ih4 28 gxh4 'ifxd3
85
A ttacking Chess: Th e Kin g 's I n dian, Vo l u m e 1

29 ll:if6+ �f7 30 'i¥a7! and White is allows Black t o show another idea be­
winning. hind 11 ... c6.
After 23 ... hs Lukacs and Hazai give
24 ll:ixgs 'ifxgs 2s f4 (?!) exf4 26 ll:ie4
'i¥e7 27 h!.xf4 as slightly better for
White, but Black can play the more ac­
tive 26 ...'t!Vh4! 27 'ii'a1 ll:ie3 28 'if xe7 ii.ts
with the idea ... 'ifg4. After 29 g3 fxg3
30 hxg3 'i¥g4 Black has the initiative.

C) 11...c6

12 ... bs1
This opportunistic move gives Black
good counterplay. He can also play
more quietly with 12 ...'ifb8 13 b4 axb4
(after 13 ... cxds 14 cxds hi.c8 White can
play lS ll:ic4! targeting b6 and as, as in
D.Giacomazzi-A.Shchekachev, Metz
199S) 14 axb4 (White may be better off
with 14 l:txb4 as in M.Notkin­
This move was once rather uncom­ V.Bologan, Saint Petersburg 199S,
mon, but has become established as when Black should probably play
the main line. Black creates more ten- 14 ... l:ta7) 14... cxds 1s cxds l:tc8 (or the
sion in the centre and keeps his options immediate 1s ... bs) 16 ii.b2 bs
open on both sides of the board. White (16 ... ii.h6!?) 17 ii.d3 'ifb6 18 ll:ib3 ii.h6
now has a choice: 19 ll:ias hi.Cl gave Black good play in
l.Nemet-J.Gallagher, Swiss League
C1: U .:.%b1 1994.
C2: 12 lla2 13 dxc6
C3: 12 1'.b2 Taking the pawn with 13 cxbs gives
Black excellent chances after 13 ...cxds
Cl) 12 l:tb1 14 exds ii.ts! (but not 14... ll:iexds? 1s
ll:ixds ll:ixds 16 ll:ie4) 1s l:tb2 ll:ifxds 16
(seefollowing diagram) ll:ixds ll:ixds.
Ignoring Black's play with 13 b4 also
This is the most natural move, but it comes to nothing after 13 ... axb4 14

86

axb4 bxc4 15 l2Jxc4 cxds 16 exds ii.ts
17 hi.b3, which was V.Chuchelov­
J.Filipek, Leuven 1999, and here Black
has 17 ... l2Je4!, with the idea of 18 ltJas
'if c8! when he enjoys excellent play.
13 . b4!
..
White's jumbled queenside pieces
allow Black to fight for the initiative.
14 ctJdS
Instead 14 cxd7 bxc3 15 l2Jf3 l2Jxe4
is the main tactical justification of
Black's play. White does have a few
other tries:
a) 14 C7 looks too fancy after
14 ... 'ifxe7 1 5 ltJds ltJexds 16 cxds a4!
(Black's better development allows him
to take over the initiative on the
queenside) 17 axb4 h!.fb8 18 hi.al (or 18
bxa4 ii.xa4 19 'ifel ii.c2 20 h!.b2 l2Jxe4
when Black is much better) 18 ... 'i¥c3 19
ii.a3 axb3 20 l:tcl 'i¥d4 and White posi­
tion fell apart in G.Grigore­
V.Nevednichy, Calarasi 1995.
b) 14 ltJbs is fairly safe. After
14 ... ii.xc6 15 ii.f3 'tlfb8 16 a4 l2Jd7 17
Itel ltJcs 18 l2Jf1 l:td8 the position was
level in G.Grigore-J.Baron Rodriguez,
Th e M a r def Plata Va ria tio n : 9 l2Jd2
Benasque 1999.
c) 14 axb4 axb4 and now:
cl) 15 cxd7 fails again after
1 5 ... bxc3 16 l2Jf3 l2Jxe4.
c2) 15 ltJds l2Jxc6 16 l2Jxf6+ ii.xf6 is
certainly fine for Black
c3) 15 C7 'ifxe7 16 ltJbs (after 16
ltJds ltJexds 17 cxds l:ta2 18 ii.d3 hi.b8
19 Itb2 Itxb2 20 ii.xb2 ii.e8! 21 l2Jc4
l2Jd7 22 ii.cl ltJcs Black was totally
equal in J.Gokhale-P.Mahmoud, Cal­
cutta 1996), and here 16 ... ii.xbs 17
cxbs ds gives Black counterplay accord­
ing to Bologan, but 16...'ifcs ! looks even
better.
14...l2Jxc6 1s l2Jxf6+
Instead 1 5 ii.b2 ii.e6 16 l:tal l2Jd7 17
f4?! exf4 18 ii.xg7 �xg7 19 hi.xf4 ltJcs
20 l2Jf1 'ifgs was much better for Black
in L.Ftacnik-R.Gadjily, Moscow Olym­
piad 1994, and Black indeed managed
to score a notable upset.
15 ...ii.xf6 16 l2Jf3 'i¥e 7 17 a4
This structure is fine for Black He
has good squares for his pieces and can
easily cover his only weakness on d6.
17 ...l2Jd8!
87
A ttacking C h e s s : Th e King 's I n dian, Vol u m e 1
The knight belongs on cs.
18 ii.h6 l:te8 19 'ifd2 ii.c6 20 ii.d3 ll:ie6
Black had an excellent position in
the game P.Lukacs-W.Uhlmann, Vienna
199S. He intends moves like ...'ifb7,
...ii.d8 and ...ll:ics.
C2) 12 l:ta2
This move always looks strange, but
it does prepare the b4-advance without
allowing Black the counterplay we saw
in Line Cl.
12...'ifbS
Here 12 ... bs? fails to 13 dxc6 b4 14
cxd7 bxc3 because White has lS ll:ibl!
when 1S ... ll:ixe4 loses to 16 ii.f3 fs 17
ii.xe4 fxe4 18 'i¥xd6.
Instead 12 ... cs is as usual solid but
stodgy after 13 h!.b2 ll:ie8 14 b4 axb4 lS
axb4 b6 16 bxcs bxcs. Others:
a) 12 ... ii.h6 13 h!.c2 cs 14 ll:idbl! 1'.g7
(14...ii.xcl lS l:txcl with the idea of f4
gives White the initiative) 1 S l:tb2 ll:ie8
16 b4 was M.Gurevich-R.Kasim­
dzhanov, Hoogeveen 1999. and here
16 ...axb4 17 axb4 b6 18 bxcs bxcs 19
ll:ibs fs 20 ll:i1c3 looks better for White.
88
b ) 12 .. JWb6 13 hi.c2 cs 14 l:tb2 l:tfb8 lS
ll:idbl 'i¥d8 16 b4 axb4 17 axb4 b6 was
very solid for Black in M.Gurevich­
A.Shirov, Prague (rapid) 2002. Black has
at least managed to connect his rooks
here and can hold firm on the queen­
side.
c) 12 ...ll:ie8 is very sensible and looks
fine. After 13 b4 axb4 14 axb4 l:txa2 l S
ll:ixa2 cxds 16 cxds f s Black had good
play in M.Gurevich-V.Babula, German
League 2000.
13 hi.c2 hi.cs 14 ll:idb1
A good example of Black's play was
14 ii.d3 cxds 1 S cxds bs 16 b4 axb4 17
axb4 ll:ihs 18 hi.el ii.h6 19 ll:idbl ii.xcl
20 'ifxcl hi.c7 21 'i¥e3 fs 22 f3 f 4 23 'i¥e2
gs 24 l:tec1 l:tb7 2s ll:id1 ll:if6 26 g4 h s
27 h 3 h!.a4 2 8 l:tb2 'i¥a7+ 2 9 'iff2 'i¥d4
with good play in V.Chuchelov­
V.Bologan, Istanbul 2003.
14 ... cxds
Black can also play 14 ... bs when lS
cxbs cxbs leads to the same position.
15 cxds bs
Black has a piece formation that is
very typical of the Ruy Lopez or the Old

Indian Defence. With such good devel­
opment, Black is not afraid to initiate
play on the queenside.
16 b4
16...ll:ieS
This is a very flexible move, not
rushing to determine the queenside
pawn structure. Black has also resolved
the tension immediately:
a) 16 ... a4 is probably a little prema­
ture, as White is not badly placed to
fight on the kingside: for example, 17
ii.e3 l:tc7 18 l:tc1 'i¥e8 19 'if d3 l:tb7 20 f4
and White had the initiative in
M.Gurevich-Y.Vovk, Cappelle la Grande
2010.
b) 16 ... axb4 is not so bad. After 17
axb4 l:ta1 18 ii.d3 ll:ihs 19 g3 l:ta7 20
ii.gs f6 21 ii.e3 hi.ac7 22 f3 fs Black had
counterplay in D.Sharavdorj-M.Al Mo­
diahki, Yangon 1999.
17 ii.e3 f5 18 f3 ll:if6 19 'ifd2 a4
Now Black decides to close off the
queenside because he will have good
play on the other side of the board.
20 h!.fcl h!.c7 21 'if d3 htb7 22 ll:id1 f4 23
ii.f2 gs
Th e M a r def Plata Varia tio n : 9 lLid2
Black had good counterplay in
A.Beliavsky-D.Stellwagen, Amsterdam
2006.
C3) 12 ii.b2
This move makes a lot of sense.
White completes his development and
does not declare his intentions just yet.
12 ...ii.h6
This is a typical deployment, al­
though Black must mind the long di­
agonal that White has just occupied.
Others:
a) 12 ... 'ifb8 is the thematic move,
but White is better developed here and
after 13 b4 axb4 14 axb4 h!.xa1 1s 'if xal
cxds (Bologan recommends Black try to
draw by sacrificing a pawn starting
with 1s ... bs!?) 16 cxds hi.c8 17 Vas
White has maintained control of the
queenside and enjoys a slight edge.
b) 12 ... 'ifb6 13 dxc6 (13 l:tcl cs is
solid) 13 ... bxc6 (Black could consider
recapturing with a piece) 14 ll:ia4 'ifc7
1s cs ds 16 ll:ib6 l:tad8 17 ii.c3 with
complications in V.Malakhov­
V.Zvjaginsev, Poikovsky 2004. Here
89
A tt a c k i n g C h e s s : The King 's In dian, Vo l u m e 1
Black could consider 11 ... ii.e6 or 17 ... a4.
c) 12 ... cs is common, but after 13
ll:ibs ll:ie8 14 b4 axb4 lS axb4 l:txal 16
'ii'xal ii.h6 (16 ... b6 17 bxcs bxcS 18 ii.c3
gives White a pleasant edge), 17 ll:ixd6 !
is a strong sacrifice, M.Ulibin-A.Galkin,
Russian Championship, Elista 1996.
13 'it>hl
White takes a moment to make a
prophylactic move and again does not
disclose his intentions. This is a patient
approach, but many other moves have
been tried too.
a) 13 dxc6 ll:ixc6 (this is fine, but
both 13 ... ii.xc6 and 13 ... bxc6 are possi­
ble too) 14 ll:ibs ii.e6 1s ii.d3 ll:id7 16
ll:ib1 ll:ics 17 ll:i1c3 ll:ia7 18 ll:ixa7 l:txa7
19 ii.c2 h!.a6 20 'ii'e2 h!.c6 21 l:tadl 'ii'b6
22 ll:ibs a4! 23 b4 ll:ib3 gave Black good
counterplay against the c4-pawn in
Shen Yang-L.Van Wely, Moscow 2009.
b) 13 .rf.bl ll:ic8!? (again we see this
manoeuvre) 14 dxc6 ii.xc6 lS ii.d3 ll:ib6
16 b4 axb4 17 axb4 ll:ihs (after 17 ... 1'.d7
18 ll:ib3 ll:ia4 19 ll:ixa4 .Jha4 20 ii.c3
bS!? 21 'ii' e2 - or 21 cxbs 'ii' b6 -
21 ...bxc4 22 ii.xc4 ii.d7 23 hi.fdl 'ii'b6
Black was also doing well in Li Wenli­
ang-F.Nijboer, Groningen 2003) 18 g3
ii.d7 19 'ii'e2 l:tc8 20 ll:ib3 ll:ia4 21 ll:ixa4
ii.xa4 22 ii.cl ii.xcl 23 l:tfxcl ii.xb3! 24
Itxb3 ll:ig 7 and with the knight bound
for e6, Black had good play in L.Ftacnik­
V.Topalov, Polanica Zdroj 199S.
c) 13 cs!? is a sharp try: 13 ...dxcs
(better than 13 ... cxds?! 14 cxd6 ll:ic8 l S
ll:ixds ll:ixds 16 exds f6 17 a4 ll:ixd6 1 8
ii.a3 ii.ts 19 ll:ic4 ll:ixc4 20 ii.xc4 when
90
White had the initiative in
R.Sherbakov-V.Syrtlanov, Koszalin
1996) 14 ll:ic4 cxds 1s exds e4! (after
1s ... ll:ifs 16 ll:ixes ll:id4 17 1'.c4 ii.ts 18
ll:ie2 ii.c2 19 'ifel 1'.g 7 20 ll:ixd4 cxd4 21
ii.xd4 ll:ixds 22 ll:ixf7! White was doing
well in R.Sherbakov-A. Korotylev, Par­
dubice 1996) 16 ll:ia4 (after 16 ll:id6
Black played 16 ... ii.g7 in F.Atakis­
T.Remmel, correspondence 2003, but
16 ... ii.fs, with the idea of 17 ll:ixb7?
'ii' b6, looks even better) 16 ... ll:iexds 17
ll:ixcs ii.c6 18 'it'd4 ll:if4 19 'it'xd8 ll:ixe2+
20 �hl l:tfxd8 21 ii.xf6 .rf.e8! ? 22 ll:id6
ii.f8 23 ll:ixe8 ii.xcs gave Black compen­
sation for the exchange in R.Jansen­
F.Nijboer, Dutch Championship, Rot­
terdam 1999, and 21 ... l:tds 22 ll:ixe4
l:tbs with a strong initiative is a good
alternative.
13 cs
...
White's slow play finally convinces
Black to acquiesce to the blocking of
the queenside.
14 ii.d3
After 14 'it'c2 ll:ie8 1s ll:ibs fs 16 ii.d3
fxe4 the game was agreed drawn in
 Th e M a r def Plata Va ria tio n : 9 lLid2

K.Sakaev-F.Amonatov, Dagomys 2009. nate the d4-knight, but this allows

If we continue with 17 ll:ixe4 ii.xb5 18 Black to contest White's domination of
cxb5 both 18 ...ll:if5 and 18...ll:if6 look the e4-square.
fine for Black.
White can also swing back his
knight with 14 ll:if3, but 14...ll:ih5 (more
active than 14 ... ll:ie8 15 ii.cl ii.xcl 16
'it'xcl ll:ic8 17 ll:ig1 ll:ig7 18 f4! with the
initiative in A.Beliavsky-M.Erdogdu,
European Championship, Budva 2009)
15 g3 f5 gave Black good play in Zhao
Xue-Wenjun Ju, Jiangsu Wuxi 2008.
14...ll:ihs 15 g3 1'.h3 16 hi.el fs 11 exfs
ll:ixf5 18 ll:ide4
Of course not 18 g4? ii.xd2. 19 ll:ibs ll:ixbs 20 cxbs ll:if6
18 ...ll:id4 By now Black already had a slight
Black has active pieces and no struc­ initiative in V.Kramnik-L.Van Wely, Nice
tural problems. White decides to elimi- (blindfold) 2008.

91
 Ch,a,pte'(c 5
Tfte Mar· def Plata Variation
1 d4 ll:if6 2 c4 g6 3 ll:ic3 1'.g7 4 e4 d6 5
l2Jf3 0-0 6 ii.e2 es 7 0-0 ll:ic6 8 ds ll:ie7 9
b4
The Bayonet. White had tremen­
dous success in the second half of the
1990s with this move which forced
Black to really shore up his defences.
White's king is usually (but not al­
ways!) pretty safe and despite the rela­
tively closed nature of the position,
there is often a lot of play in the centre.
9 ll:ihs
...
This is the challenging main line,
which will be the variation to which we
92
will devote our attention. Black takes
advantage of the possibility to jump to
a more active square than d7 or e8,
eyes the f4-square and prepares .. .fs.
It is easy to forget that the critical
10 hi.el of our next chapter is a rela­
tively new move. Even though it only
came to prominence as late as 1995, it
has been played more than all the
other moves combined and by quite a
wide margin. Nevertheless, White has
other continuations on move 10 that
Black cannot ignore and it is these lines
thatare examined in this chapter.
 Th e M a r def Plata Vari a t i o n : 9 b4 tl'i h s with o u t 1 0 hi. e l
A: 10 �2
8.: 10 tlt'c2
C: 10 C5
0; 10 g3
A) 10 tl'id2
White forces Black's knight to f4.
This old move has been considered in­
ferior since the well-known game Pet­
rosian-Gligoric, Zagreb 1970, but it is
likely that this attractive game does
not represent best play for either col­
our. Nowadays this move is very rare
and this is understandable because
Black should not really experience any
problems.
10...tl'if4 11 ..if3
White can also play 11 a4 fS 12 ii.f3,
transposing to the main line. This
avoids the possibility given in the note
to Black's 11th move, but gives Black
the option of playing 11 ... tl'ixe2+ 12
'ifxe2 fs.
11...fs
Black can also play 11 ...tl'id3 12 ii.a3
as 13 bxas hi.xas, as in V.Anand-
B.Gelfand, Dortmund 1997. As the
main line is fine for Black, we will focus
our attention there.
12 a4 gs
This is commonest and quite sound,
but Black can already consider experi­
menting:
a) 12 ...�h8!? 13 tl'ib3 g s 14 exfs
tl'ixfs 1s g3 ll'ih3+ 16 �g2 tl'ih6!? 17
1'.e4 ll'ig4 18 h!.a2 'ife8 19 f3 ll'if6 20
ii.bl 'ifhs was unclear in G.Sosonko­
F.Nijboer, Dutch Championship 1996.
b) 12 ...ii.d7 13 cs (13 ii.a3 gs 14 exfs
tl'ixfs 1s g3?! ll'ih3+ 16 �g2 'ifc8! is
good for Black) 13 ...gs 14 exfs tl'ixfs 1s
g3 ll'ih3+ 16 �g2 'ifc8 17 1'.e4 g4 was
also messy in D.Zagorskis-G.Beckhuis,
Muenster 1994.
13 exf5 ll'ixfs 14 g3
14... ll'ih3+
This looks sounder than 14 ... ll'id4!?
lS gxf4 tl'ixf3+ (1s ...exf4!? has been
suggested as a better move order, be­
cause 16 tl'ide4 tl'ixf3+ 17 'ifxf3 g4 leads
to variation 'a', below, but White could
try 16 l:ta3 ! ?) and here:
a) 16 'ifxf3 g4 17 't!fh1? exf4 18 ii.b2
93
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
ii.ts 19 l:tfel f3 20 ll:ide4 'ifh4 21 h3
ii.es gave Black a strong attack in
T.Petrosian-5.Gligoric, Zagreb 1970.
b) However, 16 ll:ixf3 ! gives Black
problems: 16 ... e4? (this loses, but
16 ... exf4 17 ii.b2 g4 18 �hl! gxf3 19
l:tg l also gives Black big problems) 17
ll:ixg s ! ii.xc3 18 l:ta3 1'.g7 19 htg3 and,
facing an overwhelming attack, Black
resigned in A.Martin-R.Britton, British
Championship, Eastbourne 1991.
Black could, however, throw in the
move 14 ... as!? to disrupt White's
queenside. In fact, I think this move
would improve Black's cause in almost
any position that follows.
15 '>t>g2 'it'd7!
Instead 1 S...g4?! 16 ii.xg4 ll:ixf2
backfired after 17 'it>xf2 ll:ih6+ 18 ii.f3
ll:ig4+ 19 '>t>gl ll:ie3 20 'it'e2 ll:ixfl 21
ll:ide4 1'.h3 22 ii.gs 'i¥d7 23 ii.g2 with a
clear advantage for White in J.Parker­
J.Gallagher, Hastings 1991/92, but here
too 1s ... as!? is possible.
16 1'.e4
Other moves are likely to bring
White trouble:
94
a) 16 1'.g4? is bad because 16 ...ll:ixf2!
does work now. After 17 �xf2 ll:id4+ 18
ii.f3 g4 Black is winning.
b) 16 ll:ide4 is natural, but White's
light-squared bishop becomes a target
after 16 ... l2Jd4 when 17 ii.hs ? ll:if4+ 18
gxf4 'ifh3+ 19 �hl g4 gives Black a
winning position.
c) 16 ll:ib3 ll:id4 (16 ... as!?) 17 ll:ixd4
exd4 18 ll:ibs c6 19 ll:ia3 (instead 19
ll:ixd4 l:txf3 20 �xf3 'if g4+ 21 �e3 looks
crazy, but 21...cS 22 bxcs dxcs 23 ii.b2
is actually not so clear) 19 ... h!.xf3 (sim­
pler is 19 ...ii.es with an excellent posi­
tion) 20 'ifxf3 g4 with compensation in
R.Keene-L.Kavalek, Teeside 197S.
16...g4
Yet again 16 ... as looks good, while
16 ... 'iff7 is also interesting. Instead
16 ...ll:ixf2 17 �xf2 ll:ixg3+ 18 '>t>gl ll:ixfl
19 ll:ixfl was unclear in D.Ruzele­
D.Lapienis, Lithuanian Championship
1993, but I would rather not trade
away Black's active minor pieces, and
16 ...ll:id4 17 f3 Itf6 (of course 17 ... as! is
again a likely improvement) 18 ll:ib3
l:th6 19 l:ta2 ll:ixb3 20 'i¥xb3 'iff7 21 cs
 Th e M a r def Plata Va riatio n : 9 b4 tl'i h s with o u t 1 0 l:t e1
was a little better for White in
V.Grabliauskas-R.Zukauskas, Vilnius
199S.
17 ll'ib3'i¥e7 18 'ifd3 hS 19 cs
19...ll'if4+
This is a bit speculative. Black could
consider 19 ... as or 19 ... tl'igs with coun­
terplay.
20 gxf4 exf4 21 f3
Panczyk and llczuk suggest 21 'it>hl
ii.xc3 22 i.xfs i.xal 23 ii.xc8 l:taxc8 24
tl'ixal h!.ce8 2S 1'.xf4 'i¥e4+ 26 'ifxe4
h!.xe4 27 ii.g3 when White may be a
little bit better in the endgame.
21... ii.es 22 hi.a2 'i¥g7 23 tl'id1
Instead Smirin gives 23 ii.xfs ii.xfs
24 ll'ie4 l:tae8 2S �hl 'i¥g6 with com­
pensation in Info rmant 69.
23 ... g3 24 htgl gxh2+ 25 �xh2 ll'ig3
Black had some compensation for
the piece and managed to overwhelm
his opponent in J.Manion-1.Smirin, Las
Vegas 1997.
B) 10 'i¥c2
Epishin's move looks funny but it
should be taken seriously. Black pro-
tects the c3-knight in anticipation of
...ll'if4; ilxf4 exf4, after which he will try
to overrun Black in the centre. One
downside to this move is that with his
queen off the d-file, White will lack
control of dS and perhaps d6, and this
could give Black some additional possi­
bilities to undermine the centre.
A similar idea for White is 10 'ifb3
ll'if4 (Black can also consider 10.. .fs! ? or
10 ... h6) 11 ii.xf4 exf4 12 l:tadl h6 13 cs
gs 14 es! (this is basically forced, as
Black would otherwise play 14 ...tl'ig6
with a great position) 14... dxes lS d6
cxd6 (after lS ...tl'ifs both 16 dxc7 'ifxe7
17 tl'ids and 16 h 3 ! ? cxd6 17 ll'ie4 ds 18
l:txds 'ife7 19 .�c4, as in S.Skembris­
A.David, Cannes 1991, give White com­
pensation for the pawn) 16 l:txd6 and
here both 16 ...'ifC7 and 16 ...'i¥e8 have
been tried with unclear play.
10.. .fs
I like this straightforward move. In­
stead 10...tl'if4 11 1'.xf4 exf4 12 h!.adl
seems to play into White's hands: for
example, 12 ... as 13 bs h6 14 ll'id4 gs 1s
ii.h s ii.es 16 h3 tl'ig6 17 tl'if3 'i¥e7 18
95
A t ta c k i ng C h e s s : Th e King 's Indian, Vo l u m e 1
b6!? cxb6 19 ll:ia4 hta6 20 'it'b3 and
White had the initiative in A.Zhigalko­
V.Kovalev, Minsk 2010.
A decent alternative, however, is
10 ... as 11 bxas hi.xas 12 hi.el cs 13 a4
h!.a6!? with a position similar to the line
10 hi.el as 11 bxas hi.xas 12 ll:id2 ll:if4
13 ii.fl cs 14 a4 h!.a6, but in this case
Black will not have to waste as much
time playing the typical retreat ... lLif4-
hS-f6. As that line is not part of our
repertoire after 10 l:tel, I will stick with
the text move, which I think gives Black
good chances.
11 ll:igs
After 11 cs I would propose
11 ...ll:if6!? 12 ll:igs fxe4 (this is logical
because playing 'ii'c2 and cs has left
the dS-pawn needing White's atten­
tion), with the idea of ... ll:ifs when the
queen on c2 is a target..
11...ll:if4
Also good is 11 ... ll:if6 when 12 ii.f3
(possible too is 12 f3, but I cannot see
how White's queen on c2 can be an
improvement on the normal hi.el in
this position) 12 ... h6 (again 12 .. .fxe4
with the idea ... lLifS-d4 is also a reason­
able idea) 13 ll:ie6 ii.xe6 14 dxe6 was
played in A.Yermolinsky-D.Sharavdorj,
Berkley 200S, and here Panczyk and
llczuk suggest 14...'it'c8 ! lS lLids ll:ifxds.
This is rarely this a good idea, but it
works here: 16 cxds c6 and Black has
good counterplay.
12 ii.xf4
This is always played, but 12 ii.f3
must also be considered: 12 .. .fxe4 in-
96
tending ...ll:ifs looks like a good answer,
while Black could consider throwing in
12 ... as!? too.
12 ...exf4 13 h!.ael
White is loading up in the centre,
but Bl ack has a good tactical answer.
13 ...ll:ic6! 14 dxc6
After 14 ll:ie6 Black has a couple of
good responses:
a) 14...ii.xe6 lS dxe6 ii.xc3! 16 'ii' x c3
fxe4 17 1'.g4 hs 18 1'.h3 (or 18 ii.d1
'ii'f6) 18 ...gs with good play.
b) 14 ... ll:id4 lS ll:ixd4 ii.xd4 16 ii.f3
ii.es 1 7 cs as 18 'it'b3?! (18 a3 is more
solid) 18 ... axb4 19 'ii'xb4 was V.Epishin­
V.Menoni, Bratto 1999, and here
19 ... dxcs 20 'ii'xcs 'ii'd6 looks good for
Black.
14...'ii' xgs 15 ii.f3
After lS cxb7?! ii.xb7 White has
problems on the long diagonal.
1s ... bxc6 16 bs 1'.d7
A decent alternative is 16 .. .fxe4 17
ll:ixe4 (17 ii.xe4 f3!) 17...'it'fs 18 bxc6
h!.b8 with good play for Black.
17 exfs 'ii'xfs 18 1'.e4 'ifgs 19 bxc6 1'.h3
20 'ii'd 3 ii.xg2! 21 ii.xg2 ii.xc3 22 h4
 Th e M a r def Plata Varia tio n: 9 b4 tl'i h s with o u t 1 0 hi. e l
'it'f6 23 hi.e6 'ifg7 24 ii.ds
Here the game V.Epishin-F.Nijboer,
Apeldoorn (rapid) 2001, was agreed
drawn, although after the forced
24...�h8 I prefer Black.
C) 10 cs
This move is the most direct. Al­
though it is not bad, this line is not
seen very much nowadays.
10...tl'if4
Black can also play 10 .. .fs, but I pre­
fer the text move, which is considered
the main line.
11 ii.xf4 exf4
12 hi.cl
This is by far the most common
move. White logically breaks the pin
along the long diagonal, but he can
also protect the loose c3-knight with
his queen:
a) 12 'it'd2 h6 (oth erwise, 12 ... 1'.g4
13 ll'id4 is annoying, but 12 ... as!? is
worth considerin g) 13 .rf.adl gs 14 es
g4!? lS exd6 cxd6 16 ll'iel and here
Black can play 16 ... tl'ig6 or 16 ...dxcs!? 17
d6 tl'ig6 18 bxcs (not 18 d7 ii.xd7 when
the c3-knight is loose) 18 ... ii.d7 with an
unclear position.
b) 12 'tlfb3 1'.g4 (12 ... h6 13 h!.adl was
considered via the move order 10 'ii'b3
in Line B) 13 hi.adl ii.xf3 14 ii.xf3 gs 1s
ii.h s (or 1s 1'.g4 tl'ig6 16 tl'ie2 lLle s)
1 S ...tl'ig6 16 ii.xg6 hxg6 was pretty level
in A.Schneider-E.Gufeld, Helsinki 1992.
12... as
This move scores well and was
played by Kasparov, which is quite a
good endorsement. Again 12 ... 1'.g4 is
met by 13 ll'id4, but Black has a couple
of other lines worth considering:
a) 12 .. .fs is worth looking at, espe­
cially to compare it to 12 ... as 13 a3
axb4 14 axb4 fs below. After 13 es
dxes 14 d6 cxd6 1s cxd6 tl'ic6 16 'ii' ds+
�h8 17 bs ll'id4 18 tl'ixes 'it'e8! 19 'it'xd4
'ii' x es 20 'ii'xes ii.xes 21 hi.fd1 hi.d8 22
hi.ds 1'.g7 23 ii.f3 ii.e6 24 hi.d3 the posi­
tion was unclear in J.Vilela-L.Perez, Villa
Clara 1998; Black has the bishop-pair,
but White has a passed d-pawn and
some queenside pressure.
b) 12 ... h6 is the main line. White has
some choice here:
97
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
bl) 13 ll:id2 g s 14 ll:ic4 ll:ig6 (14 ... a6
l S ll:ia4! ?) lS 1'.g4 ll:ies 16 ll:ixes ii.xes
17 ii.xc8 h!.xc8 18 ll:ibs and here, rather
than 18 ...'it'f6 19 c6!, Black should play
18 ... dxcs 19 bxcs c6! with counterplay.
b2) 13 lLid4 g s 14 ii.hs as lS a3 axb4 16
axb4 c6! 17 ll:ide2 would transpose to
A.Beliavsky-V.Spasov, Manila Olympiad
1992, and here the simplest move is
17 ... ii.es with equality according to
Beliavsky.
b2) 13 h 3 ! ? g s 14 a4 ll:ig6 l S as l:te8
(both 1S ...'ife7 and 1S ... a6 are possible
too) 16 ll:id2 ll:ies 17 cxd6 cxd6 18 ll:ibs
l:te7 19 'i¥c2 a6 20 ll:id4 was M.Diesen­
L.Day, Lone Pine 1977, and here Black
should play the simple 20 ... ii.d7! when
the position is fairly level.
13 cxd6
Other moves are also possible:
a) 13 a3 is the obvious move, but
opening the a-file should favour Black
compared to the lines above. After
13 ... axb4 14 axb4 fS! ? (of course, 14 ... h6
is also possible) lS es dxes 16 d6 cxd6
17 cxd6 ll:ic6 18 'it'ds+ �h8 19 bs the
idea of 19 ... Itas!? becomes a new pos-
98
sibility: 2 o 'ifd2 (20 l:tfdl 1'.d7) 2 o...ll:id4
21 ll:ixd4 exd4 22 ll:ids f3 ! (better than
22 ... ii.e6 23 ll:ixf4) 23 ii.xf3 (after 23
gxf3 ii.e6! 24 ll:if4 'it'gs+ wins for Black
and here 24 ll:ic7 Ita2 2s ll:ixe6 Itxd2 26
ll:ixd8 h!.xd8, with the idea of ...ii.es, is
also much better for him) 23 ...l:txbs 24
'i¥e2 1'.d7 and Black has an extra pawn.
b) 13 ll:ibs could also be played im­
mediately. Then 13 ... axb4 14 cxd6 cxd6
l S 'it'b3 1'.g4 16 htc7 ll:ic8! is a very typi­
cal move. The knight heads to a7 to
challenge White's own knight and un­
dermine the Cl-rook. Now White has:
bl) 17 ll:ifd4 ii.xe2 (Black could also
consider 17 ... ii.xd4 18 ii.xg4 ii.cs, with
the idea 19 l:txb7 'it'as!) 18 lLixe2 f3 19
gxf3 ll:ia7 20 hi.xb7 ll:ixbs 21 Itxbs 'ifgs+
22 �hl hi.a3 23 'it'xb4 l:txa2 gave Black
compensation for the pawn in
F.Schirm-G.Schmid, German League
1991.
b2) 17 h!.xb7 Vas 18 1'.c4 and here:
b21) 18...ii.xf3 19 gxf3 ! ll:ib6 20 l:tbl
ll:ixc4 21 'ifxc4 ii.es (Grivas suggests
21 ... h!.fc8 22 'it'xb4 l:tcl+ 23 �g2 'i¥xb4
24 Itxb4 Itxa2 2S htb8+ ii.f8 26 ll:id4
 Th e M a r def Plata Variati o n : 9 b4 tl'i h s with o u t 1 0 hi. e 1
Itdl, but after 2 7 tl'ie6! fxe6 2 8 dxe6
�g7 - 28 ... l:ta7 29 l:ta4! is an important
point - 29 l:t4b7+ �f6 30 l:txf8+ �xe6
31 l:tff7 �es 32 l:tbe7+ �d4 33 l:tf6
White can still try to press) 22 'it'xb4
'it'xb4 23 l:txb4 l:txa2 24 J:ic4 was a little
better for White in E.Grivas-J.Cooper,
Novi Sad Olympiad 1990.
b22) 18...tl'ib6! is very strong: 19 l:ic7
tl'ixc4 20 'ii'xc4 l:tfc8! 21 l:tc6 l:txc6 22
'ii'xc6 (22 dxc6 ii.e6!) 22 ...'ii' xa2 (or
22 ...ii.xf3 23 gxf3 'ii'xa2 24 ll'ixd6 b3) 23
tl'ixd6 was E.Grivas-J.Murey, Tel Aviv
1991. Here Black shouldn't play
23 ...ii.xf3 24 gxf3 b3, transposing to the
previous note, but rath er the immedi­
ate 23 ... b3! with a clear advantage.
13 ... cxd6
14 tl'ibs
After 14 'ii'd 2 axb4 1s tl'ibs Black
could play 1S ... 1'.g4 16 h!.c7, which is
our main line below, or lS ...fS !? 16 ii.d3
fxe4 17 ii.xe4 h!.a4 with counterplay.
In stead 14 bS limits White's possi­
bilities on the queenside and 14... a4!?
1s 'it'd3 (1s tl'ixa4 'it'as) 1s...ii.d7 16
tl'id2 fs 17 hi.fe 1 tl'ic8 18 ii.f3 tl'i b6 was
unclear in N.Petre-L.Vajda, Sovata 1999.
14...1'.g4
Also possible is 14... axb4 which
leads us to note 'b' to White's 13th,
above.
15 l'.!c7 axb4 16 'ii' d2
16...ii.xf3
Th is was Kasparov's choice, but I
think that Black may be better off with
the older 16 ... tl'ic8!? and then:
a) 17 l:tfcl ii.xf3 18 ii.xf3 ii.c3 !? (I
think Black can improve with 18 ...tl'ia7
19 'it'xb4 tl'ixbs 20 'it'xbs l:txa2 21 'ii'xb7
1'.d4 when he is obviously better - f2 is
weaker than f7 and Black has much the
better bishop) 19 'it'e2 (better is 19
'ifxf4 tl'ia7 20 h!.7xc3 tl'ixbs - after
20 ...bxc3 21 tl'ixd6 White has compen­
sation - 21 l:tb3 when White is at least
equal) 19 ...ll'ia7 20 l:txb7 tl'ixbs 21 l:txbs
was B.Malich-L.Vogt, East German
Championship, Erfurt 1973, and here
21 ... l:ta3 would be better for Black.
b) 17 ll'ifd4 ii.xd4!? (17 ...ii.xe2 18
tl'ixe2 was agreed drawn in
K.Grigorian-E.Geller, USSR Champion­
ship, Leningrad 1971) 18 ii.xg4 ii.cs
99
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
(with the idea of ... ll'la7) 19 es (19 .i::!.xb7
�as is better for Black) 19 ... ll)a7 20
exd6 tllxbs 21 .i::!.x cs tllxd6 22 'ii'xf4
.i::txa2 23 'ii'xb4 'ii'f6 with equality.
17 �xf3 �es
Again 17 ...tllc8! ? looks good: 18
'ii'xb4 (18 .i::!.x b7 'ii'a s) 18 ...ll'la7 19 .i::!. x b7
tllxbs 20 'ii'xbs .i::tx a2 and if anyone is
better it's Black.
18 .i::!.xb7?!
This may be good enough to main­
tain equality, but White will have to
work for it. Instead after 18 .i::!.fcl Galla­
gher gives 18 ....i::t a s! and then:
a) 19 'ii'xb4 .i::txa2 20 .i::!.xb7 .i::!.b2 21
'ii'a3 'ii'a 8! 22 .i::!.x e7 (or 22 'ii'x a8 .i::txa8
23 tllxd6 .i::tc2! ) 22 ....i::t x bs when Black is
better.
b) 19 .i::!.xb7 �a8 20 �xb4 (not 20
.l::!.ce7 .i::txa2 winning, while 20 .i::!. x e7
.i::!.xbs also favours Black) 20...'ii'x b7 21
'ii'xas .i::!.a 8 22 'ii'b4 ( 2 2 .l::!.e7 'ii'b 8!)
22 .. 1Ixa2, again with a slight advan­
tage for Black.
White has an improvement on all of
this, however, as he can play the clever
18 .i::tb 1 ! ? .i::t a s 19 .i::!.xb7 'ii'a 8 20 'ii'xb4
1 00
'ii'xb7 21 'ii'x as .i::!.a8 22 'ii'd2 when he
has an extra pawn, although it will be
difficult to convert.
18 'ii'as 19 ll'ld4 'ii'x a2 20 'ii'xa2 .i::tx a2
•••
21 .i::txb4
21 .i::!.fa8
•••
Black has also tried 21 ... 3'.xd4 22
.i::!.xd4 g s 23 .l::!.b4 (White should really
prevent Black's next by playing 23 �hS
with equality) 23 ...tll g 6 24 .i.g4 tll e s 2s
h3 .i::!.fa8 with an edge in H.Herndl­
S.Kindermann, Austrian League 1996.
22 �g4
Gallagher suggests that White
should at least activate his bishop with
22 tllc6 tllxc6 23 dxc6 .i::tc2 24 .i.d1 .i::!.xc6
2S �b3, although he will clearly be
grovelling here as well.
22 .i.xd4 23 I!xd4 gs 24 h4?!
•••
After 24 �hS .l::!.b8 Black can try to
improve his position with ... '>t>g7 (Gal­
lagher), but White should hold here.
24 gxh4 25 'it>h2 tll g6 26 '>t>h3 .i::t b2 27
•••
�fs tll es 28 'it>xh4 h6 29 �h3 '>t>g7
Now White's king is in some trouble
and Black went on to win in G.Kamsky­
G.Kasparov, New York (rapid) 1994.
 Th e M a r de/ Plata Var i a t i o n : 9 b4 li'i h 5 with o u t 1 0 I! e 1
D ) 10 g3
This favourite of Van Wely is
White's most dan gerous alternative to
10 I!el and sometimes the play is simi­
lar. Rather than discouraging ...li'if4,
White prevents it altogether.
10 fs 11 li'igs li'if6 12 f3 f4
...
This is Black's most popular choice
and is favoured by the top players. In­
stead 12 ...c6 is also popular, but White
has done well after 13 �e3. The imme­
diate 12 ... h6 13 li'ie6 �xe6 14 dxe6 c6
lS �e3 also looks good for White, and
after 12 ...'it>h 8 both 13 �e3 and Van
Wely's 13 'it>g2 score well for White, so
we will stick with 12 .. .f4, which looks to
take advantage of White's 10th move.
13 bs
This has been Van Wely's latest
choice, but we may have seen the end
of it. White has a couple of other op­
tions:
a) 13 cs has fallen out of favour be­
cause of 13 ... dxcs!? 14 �c4 (14 bxcs h6
lS li'ie6 �xe6 16 dxe6 'ii'd4+! is Black's
idea) 14 ... cxb4 lS d6+ 'it>h8 16 li'ibs (af­
ter 16 dxe7 'ii'x e7 Black threaten s both
...bxc3 and ...'ii' c s+) 16 ... h6 17 li'if7+
(Nunn gives 17 dxe7 'ii'xe7 18 li'if7+
I!xf7 19 �xf7 'ii'xf7 20 'ii'd8+ 'ii'e8, win­
ning, and 17 li'ixc7 hxgs 18 li'ixa8 li'ic6
19 li'ic7 !i'id4 as good for Black, because
20 li'ibs? li'ixbs 21 �xbs 'ii'b6+ wins)
11 ...I!xf7 18 �xf7 cxd6 19 li'ixd6 �h3
gave Black good compensation for the
exchange in Cu.Hansen-J.Nunn, Wijk
aan Zee 1991.
b) 13 'it>g2 is actuallythe most popu­
lar move.
A good part of the theory of this line
is based on Van Wely's games, but
more recently he has preferred 13 bS.
Here Black has a wide choice:
bl) 13 ... as 14 bxas (instead 14 �a3
axb4 lS �xb4 �h6 16 li'ie6? loses to
16 ... .i.xe6 17 dxe6 cs, but 14 bS!?
should be considered) 14 ..l!xas lS 'ii'b3
li'ie8 16 �d2 I!a8 17 cs was a little bet­
ter for White in E.Lobron-B.Gelfand,
Munich 1992.
b2) 13 ...h6 14 li'ie6 �xe6 lS dxe6 c6
(not 1s ...li'ic6?! 16 li'ids) 16 bS 'ii' c 7 17
bxc6 bxc6 18 �a3 I!fd8 19 'ii'a4 'ii'c8 20
I!abl 'it>h 8 21 .i::!.fdl 'ii'xe6 (21 ... as? has
101
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
been suggested, but 22 .i::!.xd6 just wins)
22 'ii'a 6 hS 23 .l::!.b 7 gave White good
compensation for the pawn in L.Van
Wely-Ye Jiangchuan, Biel lnterzonal
1993, although Black has some coun­
terplay on the kingside.
b3) 13 ... 'it>h8 14 cs h6 lS cxd6 (after
lS ll:le6 �xe6 16 dxe6 dS 17 exds
ll:lexds 18 ll:lxds ll:lxds 19 �c4 Black
has 19 ... c6 with the idea of ....i::!.e 8)
1s ...�xd6 (now after 1s ...cxd6 16 ll:le6
�xe6 17 dxe6 Black cannot support the
dS-square with ... c6) 16 ll:lbs �b6 17 a4
ll:lfxds 18 exds hxgs 19 as 'ii'f6 20
ll:lxe7 .l::!.b8 21 g4 and now 21 ... �d7? 22
bS .i::!.fc8 23 d6 left White much better
in L.Van Wely-J.Nunn, Wijk aan Zee
1992, but both 21 ...�d6 21 ...e4 improve
and look satisfactory for Black.
b4) 13 ... ll:lhs!? and here:
b41) 14 cs fxg3 lS hxg3 ll:lf4+ 16
gxf4 exf4 17 �el (after 17 lLlxh7 Black
can play the simple 17 ... 'it>xh7 18 .l:!.hl+
'it>g8 19 �b2 gs OT 17 ...�XC3 18 ll:lxf8
�xf8 19 .l::!.b l �h3+! 20 '>t>gl �xfl 21
�xfl 'ii'f6 22 �h3 '>t>g7 with a good po­
sition) 17 ...ll:lfs 18 �xf4 (18 lLlh3 ! ?)
18 ... �XC3 19 �XC3 ll:lh4+ 20 '>t>f2 (not
20 '>t>g3 .i::!.xf4 21 'it>xf4 ll:lg2+ 22 '>t>g3
�xgs+) 20....i::txf4 21 ll:le6 �xe6 22 dxe6
'ii'e7 23 cxd6 cxd6 24 .i::ta cl .i::taf8 with a
messy position, L.Van Wely-0.Cvitan,
Moscow Olympiad 1994.
b42) 14 g4 �f6! lS ll:le6 �xe6 16
dxe6 ll:lg7 17 cs 'it>h8 (17 ...ll:lc6!?) 18
cxd6 cxd6 19 ll:lbs ll:lc8 (19 ...ll:lxe6!?) 20
�c4 .l:!.e8 21 �dS .l::!.e 7 (also after
21 ... �e7 22 �b2 ll:lb6 23 a4 a6 24 ll:lc3
1 02
ll:lxe6 2 S as ll:lc8 White has compensa­
tion) 22 'ii'b 3 'ii'e 8 23 a4 a6 24 ll:lc3
ll:lxe6 2s �xe6 .i::!.xe6 26 ll:lds �d8 27
�b2 with compensation for the pawn
in L.Van Wely-A.Zapata, Matanzas
199S.
bS) 13 ...c6 increases the tension and
has scored very well.
White has:
bSl) 14 bS cs! lS 'ii'd3 ll:le8 16 ll:le6
�xe6 17 dxe6 was B.Finegold­
G.Michelakis, Groningen 1993. Here
17 ... 'ii'c8 18 ll:lds 'ii'xe6 should be
somewhat better for Black
bS2) 14 cs looks logical enough, but
remains untried. Gallagher gives
14 ... h6 (14 ... cxds! ? 1s exds ll:lfs is an­
other idea) lS �c4 (or lS ll:le6 �xe6 16
dxe6 dS!) 1s ... hxgs 16 cxd6 'it>h7 17
dxe7 �xe7 18 bS (both 18 g4 'ii'xb4 and
18 d6 'ii'd7 should be at least okay for
Black) 18 ... g4 with counterplay.
bS3) 14 'ii'b3 h6 lS ll:le6 �xe6 16
dxe6 �c8 17 .i::!.d1 .l:!.d8 (not 17...'ii'xe6?
18 .i::!.xd6!), with a final divide:
bs31) 18 ll:lds ! ? cxds 19 cxds gs 20
�d2 .i::!.f 8 21 g4 hS 22 h3 ll:lg6 23 .i::!.dcl
 Th e M a r de/ Plata Varia t i o n : 9 b4 lLJ h 5 with o u t 1 0 .i::!. e 1
� 8 2 4 ..i.el hxg4 2S hxg4 lLJe8 2 6 a4
�f6 27 as �d8 was unclear in L.Van
Wely-A.Fedorov, Batumi 1999.
bs32) 18 cs ds 19 exds lL:ifxds 20
lLJe4 'ii'xe6 (instead 20 ...lL:ifs was sug­
gested by Mikhalevski, but after 21 e7!?
lLJfxe7 22 lLJd6 'ii'd7 23 bS White has
the initiative) 21 g4 'it>h8 (probably too
slow; Mikhalevski suggests 21 ... b6,
while 21...bS!? is also possible) 22 '>t>hl
lLJg8 (again 22 ... b6 looks better) 23 �b2
lLJgf6 24 gs lLJxe4 2s fxe4 lLJC7 26 gxh6
�f6 27 �c4 'ii'g4 28 .i::!. d 6! With an ini­
tiative in L.Van Wely-R.Pruijssers, Dutch
League 2006.
We now return to 13 bS.
At first this move seems illogical
from a positional standpoint, but
White's play is based on his light­
square control.
13 ... h6
After 13 .. .fxg3 14 hxg3 lL:ih s (it is
probably better to head back to the
main line with 14 ...h6 lS lLJe6 �xe6 16
dxe6 'ii'c 8) White has:
a) lS 'ii'e 1 h6 (better is 1s ...�f6 16
lLJe6 �xe6 17 dxe6 lLJg7 18 �h6 lLJxe6
19 �xf8 �xf8 with some compensa­
tion) 16 lLJe6 �xe6 17 dxe6 'ii'c 8 18
lL:ids 'ii'xe6 19 g4 lLJf4 20 lLJxc7 'ii'c 8 21
lLJxa8 'ii' cs+ 22 'it>h2 I!xa8 23 'ii'f2 was
better for White in G.Prakash­
K.Sasikiran, New Delhi 2001.
b) lS 'it>f2! lLJf4? (this fails, but after
1S ... ..i.f6 16 lLJe6 �xe6 17 dxe6 both
17 ... lLJg7 18 cs and 17 ... 'ii'c8 18 cs 'ii'xe6
19 cxd6 cxd6 20 �3 leave White with
the initiative) 16 gxf4 exf4 17 'ii'd3 h6
18 lLJe6 ..i.xe6 19 dxe6 lLJc6 20 .l:thl lL:ies
21 'ii'd2 'ii' gs 22 .i::!.h 3 'ii'f6 23 lL:ids �xe6
24 'it>g2 �f7 2S 'ii'xf4 'ii'd7 26 'ii'g 3 and
1-0 was L.Van Wely-S.Dyachkov, Rus­
sian Team Championship 2008.
14 lLJe6 �xe6 15 dxe6 fxg3
Black only plays this because he has
a concrete idea in mind.
16 hxg3 �cS 17 lLJds
17 ...'ii'xe6 18 lLJxc1 'ii'h 3
This is the point of Black's play.
19 l!f2
After 19 lLJxa8 'ii'xg3+ 20 '>t>h l Black
can, of course, take the perpetual
check, but he could also try 20 ... lLJhS ! ?
21 'ii' el .i::tx a8 22 .i::tf2 'ii'h4+ 23 'it>g2
1 03
A ttacking C h e s s : Th e King 's I n dian, V o l u m e 1
ll:lf4+ 24 �xf4 exf4 2S .i::!.d1. Then
2s ...�d4 26 .i::txd4 'ii'g 3+ 27 'it>h1 'ii'h4+
is still a draw, but Black could also go in
for either 2s ... �es or 2s ....i::!.c 8! ? with
the idea of ... .i::tc s.
19 ll:lxe4!
.••
Correct, whereas after 19 ... .i::tac8 20
.i::th 2 'ii'xg3+ 21 .i::!.g 2 'ii' h3 22 'ii'xd6 .l::!.f7
23 cs �f8 24 ll:le6 White was winning
in L.Van Wely-M.Golubev, Romanian
Team Championship 2000.
20 fxe4
This leads to a very delicate posi­
tion. Instead 20 .i::th 2 'ii'd7 21 ll:lxa8
ll:lxg3 22 �xh6 �xh6 23 .i::!.xh6 '>t>g7 24
.i::th 2 ll:lefS! gave Black compensation
for the exchange in L.Van Wely­
J.Degraeve, Mondariz Balneario Zonal
2000.
20 ....i::txf2 21 'it>xf2 .i::!.f8+ 22 '>t>e3 'ii'xg3+
23 'it>d2 .i::tf2 24 ll:leS
After 24 'ii'b3 'ii'g 2 2S 'ii'e 3 h S ! (Gal­
lagher) Black gets his h-pawn moving
and threatens ... �h6.
24...'ii'f3!
A tremendous improvement on
1 04
24 ... h s (24 ... ds has also been consid­
ered), although this is playable too: 2S
ll:lxg7 'it>xg7 26 'ii'b3 'ii'g 2 27 'ii'e 3 ll:lg8!
28 cs dxcs 29 � b2 and now, instead of
29 ... 'ii'h 2?! 30 .i::!.e l ll:lf6 31 '>t>dl b6 32
�c3 'it>h7 33 'ii'g s when White was
much better in L.Van Wely-T.Radjabov,
Dresden Olympiad 2008, Black should
have played 29 ... ll:lf6! 30 .i::!.el (not 30
�xes? .i::!.x e2+ 31 'ii'xe2 'ii' g s+) 30 ... ll:lg4
when he has counterplay with his ac­
tive pieces and dangerous h-pawn.
2 s ll:lxd6 .i.f6! 26 cs
After 26 '>t>el 'ii'g 3! 26 'ii'e l .i::!.h 2 27
cs Black has 27...ll:lc8! 28 '>t>dl (or 28
ll:lxc8 'ii'xe4) 28...'ii'g 2 29 'ii d2 ll:lxd6 30
cxd6 'ii'h l+ 31 'ii'e l 'ii'xe4 when White
is in trouble.
26...ll:lds! 27 exds e4! 28 'it>e1
Now 28 ...'ii'g 2 29 'it>d2 �xal was
good enough to win in L.Van Wely­
D.Stellwagen, Dutch Championship,
Amsterdam 2009, but even stronger
was 28 ... �c3+! with the idea of 29 �d2
'ii'g 3 30 �xc3 e3 31 'ii'd3 .i::tfl+ 32 'it>xfl
'ii'f2 mate!
 Chapter 6
The Mar del Plata Variation
9 b4 tbhs 10 l::.e 1
1 d4 ll:lf6 2 c4 g6 3 ll:lc3 �g7 4 e4 d6 5
ll:lf3 o-o 6 �e2 es 7 o-o ll:lc6 8 ds ll:le7 9
b4 ll:lhs 10 .i::te1
This move was introduced into
modem practice by Ivan Sokolov. Soon
hordes of top GMs, headed by Vladimir
Kramnik, started playing this way. The
game Kramnik-Kasparov, Novgorod
1997, sounded the alarm and Kasparov
quickly stopped playing the King's In­
dian. The world took notice and the
King's Indian became a 'dubious' open­
ing and virtually vanished from the
upper echelon of the chess world.
Truth be told, Kasparov never
'blamed' this loss to the Bayonet for the
nixing of the opening from his reper­
toire. Since then, however, not only has
Teimour Radjabov resurrected the
King's Indian, but he has had great suc­
cess in this particular 'death knell'
variation from the black side. Kramnik
himself has faced Radjabov twice in the
Classical King's Indian, and both times
he has avoided the Bayonet...
10 fs
...
The main alternative is 10 ... as 11
bxas .i::txas 12 ll:ld2 ll:lf4 13 �fl cs 14 a4
.i::ta6 15 .i::ta 3 with a stodgy game ahead.
This has never appealed to me and the
fact that nowadays the retreat
15 ... ll:lhs, with the idea of a further re­
treat with ...ll:lf6-e8, has become con­
sidered Black's most reliable method of
play furthers my suspicions.
11 ll:lgs ll:lf6
The alternative 11 ...ll:lf4 is not so
bad. After 12 �xf4 exf4 13 .i::!. cl Black
can choose between 13 ... �f6, 13 .. .fxe4
105
A ttacking Chess: Th e King 's I n dian, Vo l u m e 1
and 13 ... j.h6!?. The position tends to
simplify greatly, but Black is rarely in
any 'danger' of winning the game.
After 11...lL'if6, White's e4-pawn is
attacked, so his choice is limited. White
has two main lines here and they are
both quite complicated.
A: 12 .i.f3
B: 12 f3
A) 12 j.f3
White increases his piece play on
the light squares. The diagonal for the
light-squared bishop can easily be
opened if White plays lL'ig5-e6 and exf5,
1 06
and this variation tends to lead to forc­
ing play.
12 c6
...
Black shuts down the f3-bishop's
diagonal and increases the tension in
the centre. Black may also initiate the
sequence 13 ... h6 14 lL'ie6 j.xe6 15 dxe6
followed by either .. .fxe4 and ... d5 or
... lL'ie8-c7 depending on the circum­
stances.
The only other sensible continua­
tion is really 12 .. .fxe4 with the idea of
... lL'if5. After 13 lLicxe4 lL'if5 14 �b2 or
13 lLigxe4 lL'if5 14 .i.g5 lLid4 15 lL'ib5!
White is somewhat better, although
Black has a solid position. In general I
do not like this structure when the c­
pawns are still on the board, because
the advance c4-c5 gains space for
White and Black always has to consider
the possibility of both cxd6 and c5-c6.
After 12 ...c6 White has several pos-
sibilities:
Al: 13 'W'b3
A2: 13 b5
A3: 13 .i.b2
A4: 13 .i.e3
Other moves are less dangerous:
a) 13 lL'ie6 looks premature. After
13 ... j.xe6 14 dxe6 fxe4 15 lLixe4 lLixe4
16 j.xe4 d5 17 cxd5 cxd5 18 .i.c2 e4 19
.l::!.bl 'ii'b6 (but not 19 ...'ii'd 6? 20 j.xe4
j.e5 21 .i.f3 j.xh2+ 22 'it>h1) Black is
already better.
b) 13 .l::!.bl is Bareev's favourite move
in various positions, but here it looks
 Th e M a r de/ Pla ta Va ria tio n : 9 b4 li'i h 5 1 0 I! e 1
irrelevant after 13. . .h6 1 4 li'ie6 �xe6 15
dxe6 fxe4 16 li'ixe4 li'ixe4 17 �xe4 d5
18 .i.c2 'ii'd6 19 'ii'g4 (E.Bareev­
T. Radjabov, Sarajevo 2003), when both
19 ...e4 and 19 ... .i::!.f6 look good for Black.
c) 13 .i.a3 is not bad and will gener­
ally transpose to Line A2, although
there are a couple of independent pos­
sibilities:
cl) 13 ...h6 is the most forcing: 14
li'ie6 .i.xe6 15 dxe6 fxe4 16 �xe4 (16
li'ixe4 li'ixe4 17 �xe4 d5 18 cxd5 cxd5 19
b5 reaches the main line of 'A2') 16...d5
17 �c2 ! ? (again, both 17 cxd5 cxd5 18
b5 li'ixe4 19 li'ixe4 dxe4 and 17 b5 li'ixe4
18 li'ixe4 dxe4 19 'ii'xd8 .i::!.fxd8 20 .i.xe7
I!e8 21 .i.c5 cxb5 22 cxb5 bring us to
Line A2) 11 ... 'ii'd6 18 cxd5 cxd5 19 �b3
'it>h7 20 'ii'e2 li'ih5 21 I!ad1 li'if4 22 'ii'f1
and now, rather than 22 ...'ii'xe6 23 b5!
.i::!.ad8 24 �xe7 'ii'xe7 25 li'ixd5 li'ixd5 26
.i.xd5 with an edge for White in J.Chab­
anon-0.Touzane, Marseille 2010, Black
should try 22 ... .i::ta c8! with good play.
c2) 13 ... cxd5 gives White the possi­
bility to deviate, but Black should still
be fine:
c21) 14 cxd5 h6 15 li'ie6 �xe6 16
dxe6 fxe4 (also possible is 16 .. 1Ic8 17
'ii'b3 d5, as in A.Shariyazdanov­
M.Cebalo, Rabac 2003) 17 .i.xe4 d5 18
b5 li'ixe4 19 li'ixe4 dxe4 and once more
we have transposed to Line A2.
c22) 14 exd5!? e4 15 .i.e2 and then:
c221) 15 ...li'ifxd5 16 li'ixd5 .i.xal 17
li'ixe7+ 'ii'xe7 18 'ii d5+ �g7 19 .i::!.x al
'ii'xg5 20 .i.b2+ 'it>h6 21 'ii'x d6 gave
White very good compen sation for the
exchange in L.Portisch-P.Acs, Rethym­
non 2003.
C222) 15 ... !i'Jh5 ! ? 16 'ii'd 2 li'if4! 17
'ii'xf4 .i.xc3 18 'ii'h4 h5 19 .i.xh5 was
J.Lautier-T.Moriuchi, Tokyo (simul)
2002, and now 19 ... gxh 5 20 �xh5 .i::!.f6
21 'ii'h 7+ �8 22 'ii'h 8+ li'ig8 23 li'ih7+
�7 24 li'ig5+ is a draw.
c223) 15 ... h6 looks the simplest: 16
li'ie6 .i.xe6 17 dxe6 d5 18 cxd5 li'ifxd5
19 li'ixd5 �xd5 is equal according to
Portisch.
Al) 13 'ii'b3
This is a tricky move. White creates
pressure on the a2-g8 diagonal, but
107
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
now there is no pressure on the d6-
pa wn, a consideration that Black
should logically try to take advantage
of.
13 h6
...
This is the main move, but 13 ... 'it>h8
is also possible.
14 lL'ie6 �xe6 15 dxe6 'ii'c 81
Because nothing is attacking the
d6-pawn, Black can attack the e6-pawn
right away. He has a good alternative
in 1s ...fxe4 16 lLixe4 lLixe4 17 .i.xe4
�c8 and now after 18 'ii' h3 dS 19 .i.c2
lL'ifs Black has counterplay, while 18
.i::!.d 1 .i::!.d8 19 bS (after 19 'ii' h 3 dS 20
.i.c2 e4 Black looks to be doing well)
19 ...'ii'xe6 leaves White with some
compensation for the pawn, but
probably no more than that.
16 bs
White plays on the light squares.
There are several alternatives:
a) 16 cs?! looks panicky and back­
fires quickly: 16 ...fxe4 17 cxd6 exf3 18
dxe7 .i::!.e 8 19 �b2?! (better is 19 gxf3
I!xe7 20 .i::!.x es lL'ihs 21 .i::t e 3 lLJf4 22 .i.b2
'ii'f8 23 .i::t a el 'ii'f6, but Black is doing
1 08
well here too) 19 ... .i::tx e7 20 .i::!.x es 'ii'C7
21 .i::!. e el fxg2 (White's kingside is a
wreck and Black threatens ...lL'ig4) 22 h3
.i::!.f8 23 .i::t e 3 lL'ihs gave Black a strong
attack in Y.Pelletier-E.lnarkiev, Istanbul
2003.
b) 16 �a3 is an interesting move
order.
After 16 ... 'ii'xe6 17 .i::!.adl .i::!.fd8 18 bS
'it> h7 (note that compared to the main
line, 18...cs?! is not so good here be­
cause after 19 exfs gxfS 20 �xb7 .i::!.ab8
White has better piece control to sup­
port 21 ..i.dS!) 19 .i::t d2 .l::!.d7 20 bxc6 bxc6
21 .i::!.edl .i::!.ad8 22 'ii'a4 f4 White had
some compensation for the pawn in
P.Eljanov-T.Radjabov, Moscow 200S,
although Black has chances here as
well.
c) 16 .i::!. d1 .i::!.d8 17 bS (instead 17 cs
fxe4 18 cxd6 exf3 19 dxe7 .i::!.x dl+ 20
'ii'xdl 'ii'xe6 21 'ii'd8+ 'it>h7 was H.Kallio­
V.Kotronias, Batumi 2002, when White
could not take the rook because 22
'ii'x a8? 'ii'g4 wins for Black) 17...'ii'xe6
18 �a3 (after 18 bxc6 both 18 ... lLixc6
19 exfs gxfs 20 'ii'xb7 lL'id4, E.Bacrot-
 The M a r de/ Plata Va ria tio n : 9 b4 li'i h 5 1 o .i::!. e 1
T.Radjabov, F I D E World Championship,
Tripoli 2004, and 18 ...bxc6 19 �a3 'it>h7
20 'ii'a4 .i::!.ab8 21 �a6 fxe4 22 li'ixe4 li'ifs
23 li'ixf6+ 'ii'xf6 24 �xc6 .l::!.b6 25 'ii'a4
li'id4, Y.Pelletier-A.David, French League
2007, gave Black enough piece play)
and here:
cl) 18 ...'it>h8 19 bxc6 li'ixc6?! 20 exfs
gxfs (after 20 ... 'ii'xfs 21 'ii'xb7 li'id4 22
�e4 li'ixe4 23 li'ixe4 .i::!.ab8 24 'ii'ds li'ic2
25 li'ixd6 'ii'f6 White has 26 li'if7+!) 21
�ds 'ii'e7 22 li'ibs li'ie8 23 .i::tab1 and
White had the initiative in E.Bacrot­
T.Radjabov, Bled Olympiad 2002.
c2) 18 ...'>t>h7 is given by Bologan.
This is a subtle improvement that I had
to figure out and 19 bxc6 (or 19 .i::!.a bl
.l::!.d7 20 bxc6 li'ixc6 21 exfs 'ii'xfs 22
�xc6 bxc6 23 .i::!.xd6 .i::!.ad8 24 cs e4 -
Bologan) 19 ...li'ixc6!? 20 exfs 'ii'xfs 21
'ii'xb7 li'id4 22 �xd6 (it transpires that
this time after 22 �e4 li'ixe4 23 li'ixe4
.i::tab8 24 'ii'ds li'ic2 25 li'ixd6 'ii'f6 there
is no check on f7 and so Black wins ma­
terial ! ) 22 ....i::tac8 23 li'ibs li'ixf3+ 24
'ii'xf3 'ii'xf3 25 gxf3 a6 is equal accord­
ing to Bologan.
16...'ii'x e6 17 �a3
White keeps the tension, but now
Black can close the position. The im­
mediate 17 bxc6 has not been played
or mentioned, probably because after
17 ... bxc6 18 �a3 .i::!.ab8 19 'ii'a4 li'id7!
White cannot take the a7-pawn and
...li'ib6 is threatened.
11 . .cs!?
.
Black can also play 17...'>t>h7 when
18 .i::!.adl .i::!.fd8 19 bxc6 bxc6 (after
19 ... li'ixc6?! 20 exfs both 20...gxfs 21
'ii'xb7 li'id4? 22 .i::!.xd4 and 20...'ii'xfs 21
'ii'xb7 li'id4, as in V.Mikhalevski­
C.Matamoros Franco, Drammen 2005,
and then 22 �xd6! are bad for Black)
20 'ii'a4 .i::!.a c8 21 'ii'a6! gave White com­
pensation for the pawn in P.Eljanov-
1.Nataf, Mallorca Olympiad 2004. Here
Black could try 20...f4, with play similar
to Eljanov-Radjabov in note 'b' to
White's 16th move, above.
18 exfs
This wins back the pawn, but Black
will get good piece play. Instead 18
.i::!.adl f4 19 .i::t d3 .i::!.ad8 20 h3 was
S.Porat-R.Djurhuus, Drammen 2005,
1 09
A ttacking C h e s s : Th e King 's I n dia n , Vol u m e 1
and here Mikhalevski suggests 20 ... g5
21 .i::!.edl b6 with the idea of ...h 5 and
... g4.
18 ...gxf5 19 �xb7 I!ab8
20 .i.ds
After 20 �f3 e4 Black intends ... lLic6
or ... lLig6: for example, 21 �e2 lLic6! 22
�b2 lL'id4 23 'ii'dl lL'id7 with excellent
piece play.
20 ...lLifxds 21 cxds �f7 22 �b2 e4 23
.l:!.adl ll'ig6
Black stood well in A.Korobov­
D.Yevseev, Sochi 2008.
A2) 13 bS
This forcing continuation has ex-
110
perienced certain bouts of popularity.
At first this move looks very strange
positionally - usually White is aiming
for c4-c5, but in the Bayonet the ad­
vance b4-b5 is a common theme be­
cause White's initiative is largely based
on his light-square play, especially with
lL'ig5-e6 always in the air. These days it
is known that the main line basically
leads to a drawish ending. Black does
have one main deviation, but it entails
some real risks.
13 ...cxds
Instead 13 ... h6 14 lLie6 �xe6 15
dxe6 fxe4 16 lLixe4 lLixe4 17 �xe4 d5
18 cxd5 cxd5 19 �a3 leads to the same
thing. This move order would deny
Black of the option given at move 16,
however.
14 cxds
Sometimes the recapture 14 exd5
has to be taken into account when
Black plays ...cxd5 before ...h6, but in
this case White has problems on the
long diagonal after 14 ... e4.
14...h6
If Black really wants to avoid the
forcing lines that follow, the rare
14 .. .fxe4 could be played. After 15
lLigxe4 lLixe4 16 lLixe4 lL'if5 White is a
little better, but these positions are not
so bad for Black when the c-pawns
have been exchanged.
1s lL'ie6 �xe6 16 dxe6 fxe4
Thi s is the theoretically approved
line, but if Black wants to take a chance
there is 16 ....i::!.c 8!?.
This looks attractive (at least in the
 Th e M a r de/ Plata Va ri a t i o n : 9 b4 li'i h 5 1 0 .i::!. e1
sense that i t does not just lead t o a
level endgame), but I am not so sure it
is worth the trouble ... Here White has:
a) 17 �b2 fxe4 18 li'ixe4 li'ixe4 19
.l:txe4 (19 .i.xe4 dS 20 �a3 dxe4 21
'ii'xd8 .i::!.fxd8 22 �xe7 .i::!.e 8 is a familiar
ending, but Black has the extra move
_ ..i::tac8) 19 ... ds 20 .i::!. x es �xes 21 .i.xes
is very similar to Line A3. After 21 ...�b6
(21 ... .i::tc4 is also possible) 22 �b2 .l::!.c4
(not 22 ...'>t>h7 23 �e2 d4 24 h4 .i::!.f6 2S
.l:!.e1 d3 26 �es when White was win­
ning in T.Nyback-D.Stellwagen, Ger­
man League 2oos) 23 'ii'e2 'ii'a 24 �e3
�f4 2S 'ii'xa7 (after 2S �a3 'ii'd2 26
.l:!.dl l:te4! 27 '>t>fl 'ii' c 2 28 �xe7 'ii'xb2
29 'ii'd7 .i::!.xf3 30 gxf3 'ii'e2+ 31 'it>g2
.l::!.g 4+! it's a draw) 2s ....i::t c2 26 �a3 'ii' e s
27 .i::tf1 .i::tx a2 28 'ii' c s .i::!.xf3 29 gxf3 �gs+
30 'it>h1 'ii'fs 31 'it>g2 �gs+ 32 'it>h1 �fs
White could not avoid a draw in
Z.Peng-Li Shilong, Wijk aan Zee 2008.
b) 17 �b3 ds 18 li'ixds li'ifxds 19
exds e4 20 d6 and here:
(seefollowing diagram)
bl) 20 ... .i::t c3 ?! 21 dxe7 (much
stronger than the 21 �a3? .i::!.xb3 22
dxe7 .i::txa3 23 exd8'ii' .i::!.x d8 24 .i::!.adl of
Ki.Georgiev-R.Ponomariov, Istanbul
Olympiad 2000, when both 24...l:te8
and 24....i::!.c 8 are good for Black)
21 ...'ii'xe7 22 'ii'a4 exf3 23 �b2 f4 (or
23 ... �gs 24 g3 'ii'g4 2s 'ii'xg4 fxg4 26
.i::tadl .i::ta 27 e7 .i::!.e 8 28 .i::!.d8 1-0,
S.Savchenko-F.Guilleux, Le Touquet
2007) 24 'ii'e4 'ii'g s 2s g3 .i::t e 3 26 �c4
was drawn here in R.Szuhanek­
N.Grigore, Eforie Nord 2008, but White
is much better.
b2) 20 ... �xal 21 dxe7 'ii'xe7 22 �a3
'ii'f6 23 e7+ .l::!.f7 24 �e2! (Zakhartsov's
idea; instead 24 �dS .i::!.e 8 2S �dl 'ii'c 3
was good for Black in S.Savchenko­
M.Mozharov, Sochi 2008) 24 ... �es (also
bad are 24... '>t>g7 2S .i::t xal 'ii'xal+ 26
.i.fl and 24 ... '>t>h7 2S b6! a6 26 'ii'a4) 2 S
b 6 ! a 6 ( 2 s...�xb6 26 'ii'xb6 axb6 2 7
.i.bs) 26 �c4 '>t>g7 21 �xf7 'ii'xf7 28
.i::!. d l and the e-pawn will give Black big
problems.
b3) 20...'ii'xd6 21 �a3 �f4 (worse
are 21 ...�dS? ! 22 .l:!.adl 'ii'xb3 23 axb3
111
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
exf3 24 �xe7 .i::!.fe8 2 S .l::!.d7 �f6, as in
Xu Jun-Ye Jiangchuan, Taiyuan 2004,
and then 26 �b4!, and 21 ... 'ii'b6 22
�xe7 .i::!.fe8 23 �b4 �xal 24 .i::txal exf3
2 S e7+ 'l;g7 26 �a3 ! with a strong ini­
tiative) 22 .i::!.a dl (not 22 �xe7 .i::!.fe8 23
�a3 .i::tc 3 24 'ii'ds exf3 2s �d6 'ii'g s 26
�g3 f4 when Black is better) 22 ....l::!.C 3
23 �4 .i::txa3 (better than 23 ... .i::!.e8 24
�e2) 24 'ii'xa3 exf3 2S .l::!.d7 �f6 26
'ii'xf3 'ii'b4 27 'ii'e2 with an unclear po­
sition according to Zakhartsov.
11 ll:lxe4
17 �xe4 ll:lxe4 18 ll:lxe4 dS 19 �a3
makes no difference
17 ...ll:lxe4 18 �xe4 ds 19 �a3
This leads by force to an ending
where White hopes to be marginally
better. Instead he must avoid 19 �c2
e4 20 .l::!.bl 'ii'b 6 when Black is hitting
both f2 and e6.
19...dxe4 20 'ii'x d8 .i::!.fxd8 21 �xe7
21 ....i::te s
The alternative 21 ....i::td s!? has only
seen one outing, although it was suc­
cessful: 22 .i::!.adl .i::!.x bs 23 �d8 .i::!.c 8 24
.i::!.d6 .i::!.c6 2S .i::!.xc6 bxc6 26 h4 �8 27
112
.i::!. cl 'l;e8 28 e7 V2-V2 L.Fressinet­
M.Hebden, Lausanne 2001.
22 �cs .i::!.x e6 23 �e3
This is White's latest attempt to
keep a little something in the position.
He hopes to develop some pressure
against Black's queenside. In general,
taking on e4 does not lead to anything
because Black's king can become active
very easily, which makes up for the
weak es-pawn. My first experience
with this line was back in 1996 when
GM Igor Nataf asked me to play the
white side of this endgame in some
blitz games on line to see if it was ten­
able for Black. It was. A more serious
example: 23 .i::!.xe4 a6 24 b6 .i::!.c 8 2S �e3
.l:!.d6! and Black was at least equal in
M.Brodsky-A. Shimanov, St Petersburg
2008; Black's king will come to e6.
Instead 23 .i::!.a dl a6 24 bxa6 .i::!.exa6
2S a3 �f8 was also level in E.Bacrot­
T.Radjabov, Dresden Olympiad 2008.
23 ...a6
Black removes the a7-pawn from
the scope of White's bishop. The care­
less 23 ...b6?! 24 a4 a6? runs into the
breakthrough 2s as! axbs 26 axb6
when Black is in big trouble.
24 b6
The only try to eke anything out of
the position is to advance the pawn,
but it could become weak here.
24....i::!. d S
This looks like the simplest move to
me. Black should certainly not go to
sleep just yet, though, as these exam­
ples show:
 Th e M a r de/ Plata Varia t i o n : 9 b4 li'i h 5 1 o .i::!. e 1
a) 24...�f6 intending ... �gs has al­
ways been considered a simple solu­
tion, but White may still push a little
bit here: 2S g3 (2S a4 has also been
tried) 2s ... �gs 26 �xgs hxgs 27 .i::!.xe4
.i::!.xb6 28 .i::!.x es .i::t bs 29 .i::t ael (after 29
.i::!.xbs axbs 30 .i::t d 1 '>t>g7 31 .l::!. d7+ 'it>h6
32 .i::!.xb7 .i::txa2 33 .i::txbs g4 Black drew
easily in J .Lautier-V.lvanchuk, Mallorca
Olympiad 2004) 29 ... g4 30 .i::t1e4 .i::!.x es
31 .i::!.x es bS 32 .i::tg s �7 33 .i::!.xg4 and
White went on to win in S.Savchenko­
V.Talla, Legnica 2007.
b) 24....i::!.c 8?! 2s .i::tacl .i::!.ec6 (2s ....i::tcc6
26 .i::!.edl �f8 27 'it>fl also gives White
some hope) 26 .i::!.c dl! .i::!.f6 27 .l::!. d7 .l::!.f7
28 .i::!.edl .i::!.e8 29 .l::!.c 7 and White main­
tains some pressure.
c) 24 ... '>t>f7!? leads to the main line
after 2s .i::ta cl .i::!.d 8 26 .l::!. e7+ .l::!. e 7 27 �cs
.i::!.e d7.
25 .i::t a cl '>t>f7
Or 2s ... .i::!.ed6 with the idea of .. 1I8d7
and ... '>t>f7-e6.
26 .l::!. c 7+ .l::!. e7 27 �cs .i::!.e d7
Black is totally fine and after 28 g3
�f8 29 .i::!.xe4 .l::!. xe7 30 bxe7 .l::!. c 8 31 �b6
�d6 32 'it>g2 'it>e6 33 .l::!.g4 gs 34 h4 the
game was agreed drawn immediately
in N .Rashkovsky-F.Florian, Biel 2001,
although Black could have tried
34 ... �xe7 with the idea of 3S �xc7
(better is 3S .l::!. c 4) 3S .. 1Ixc7 36 hxgs
'it>f S! when he is suddenly much better.
Another game went 28 g4 'it>e6 29
'it>g2 .l::!.xe7 30 bxe7 .i::!.c8 31 �b6 �f8 32
.i::!.xe4 �d6 and Black even went on to
win in A.Van Beek-G.Kodentsov, Gron­
ingen 2007.
A3) 13 �b2
This move became popular after
Shirov used it to defeat Radjabov. The
bishop does not look very effective on
b2, but White plans a dangerous ex­
change sacrifice to justify it.
13 h6
•••
Black heads down the rabbit hole.
The alternative is the immediate
13 ... cxds 14 cxds (Black must also
worry about 14 exds) 14...fxe4 (instead
14 ... h6 lS li'ie6 �xe6 16 dxe6 fxe4 17
li'ixe4 li'ixe4 18 I!xe4 dS brings us back
to the main line) lS li'igxe4 li'ifs 16
113
A ttacking C h e s s : Th e King 's I n dia n, Vo l u m e 1
ll:lxf6+ and now both 16 ... �xf6 17 lLle4
�g7 18 .i::!.cl �d7 (V.Laznicka-G.Jones,
European Team Championship, Herak­
lion 2007), and 16 ... .i::txf6 17 lLle4 .l::!.f7 18
g3 (D.Vigorito-J.Rihel, Somerville 2009)
18...�d7 are only a little bit better for
White.
14 ll:le6 �xe6 15 dxe6 fxe4
Black can try to avoid the coming
complications with 1S ...'ii'C7 16 'ii'b 3 (16
cs ds 17 exds ll:lfxds 18 �xds .i::ta d8
looks okay for Black), and here
16 .. .fxe4!? looks worth a try after either
17 i.. xe4 .i::!.a e8 or 17 ll:lxe4 ll:lxe4 18
�xe4 .i::!.f6. In practice Black has pre­
ferred 16 ... .i::tad8, but 17 .i::!.adl fxe4 18
ll:lxe4 ll:lxe4 19 �xe4 �c8 20 'ii' h3 I!f6
21 ..tc2! .i::!.xe6 22 f4 I!f8 23 fxes (worse
is 23 cs dS 24 fxes, as in W.Paschall­
D.Karatorossian, Budapest 2004, be­
cause of 24...h s with the idea of ...ll:lfs)
23 ... dxes 24 cs ll:lds 2s �b3 .l::!.e 7 26
�xc8 .i::!.xc8 27 �xds+ cxds 28 I!xds e4
29 ..txg7 'it>xg7 30 'it>f2 was clearly bet­
ter for White in L.Vrbica-T.Pupak, corre­
spondence 2004.
16 ll:lxe4
Instead 16 �xe4 gives Black a rather
pleasant choice:
a) 16 ...'ii'b6 17 �b3 dS! 18 cxds cxds
19 �xds ll:lg4 gave Black counterplay
in S.lvanov-B.Avrukh, Beersheba 1998.
b) 16 ...ll:lxe4 17 ll:lxe4 ll:lfs 18 'ii'g4
'it>h7 (18 ... 'ii'e8 19 e7!) 19 .i::!.adl 'ii'e 7 20
Ild3 .i::tae8! (not 20...'ii'xe6 21 .i::!.xd6) was
satisfactory for Black in O.Averkin­
V.Zakharstov, Krasnodar 1998.
c) 16 ... ds 17 cxds (Black was better
114
after 1 7 �d3 e4 1 8 �fl 'ii'b6 19 .l::!.bl
ll:lhs in E.Bareev-T.Radjabov, Enghien­
les-Bains 2003) 17 ...cxdS 18 �c2 and:
cl) 18...'ii'd6 19 �b3 .i::!.ad8!? (this
looks better than 19 ... 'ii'xe6 or
19 ...'ii'xb4) 20 ll:la4 ll:le4 21 f3 ll:lgs 22
ll:lcs ll:lxe6 looks okay for Black.
c2) 18 ... e4 19 ll:la4 (it may be better
to just play 19 �b3 'ii'b6 20 .l:!.cl 'it>h7 21
bS 'ii'xe6 22 �a3 .i::!.ad8 with an unclear
position) 19 ... b6 20 f3 exf3 21 'ii'xf3
�d6! 22 .i::!.adl .i::tac8 23 �bl .l::!.c4 and
the strong threat of ... ll:lg4 gave Black
the initiative in A.Mista-R.Antoniewski,
Trzebinia 1998.
16...ll:lxe4 17 .i::!.x e4
Worse is 17 �xe4 dS 18 cxds cxds
19 �c2 �b6 20 'ii'd2 e4! 21 �xg7 'it>xg7
22 �b3 .i::tfs and Black was better in
M.Peek-A.David, Amsterdam 2000.
11 ...ds
The only real alternative is Galla­
gher's old suggestion of 17 ...ll:lfs. After
18 bS! .i::!.c8 19 .i::te 2 .i::!.e8! ? (worse is
19 ...�e8 20 bxc6 bxc6 21 'ii'a4, V.Golod­
T.Nedev, European Championship,
Plovdiv 2008) White has:
 Th e M a r de/ Plata Varia tio n : 9 b4 li'i h 5 1 o .i::!. e 1
a ) 20 bxc6 bxc6 21 cs (21 �a4 'ilia
22 .l::!.bl .i::!.xe6 23 �c3 .l::!.e 7 24 �e4 'it>h7
2S g4?! li'ih4 26 �el .l::!.f7 27 .i::!.b3 .l::!.f4!
gave Black the upper hand in A.Wirig­
N.Pokazanjev, Winterthur 2008) 21 ... ds
22 �xes �xes 23 .i::!. x es �f6 24 .i::t e 1
was agreed drawn here in R.Dautov­
S.Kindermann, Nussloch 1996. After
24 ... .i::txe6 the position is pretty level.
b) 20 �a4 'fib6 21 .i::!.d1 (or 21 bxc6
bxc6 22 �e4 I!xe6 23 .l:tb1 �c7 24 �c3
li'id4 2S .l:teb2, which gives White com­
pensation for the pawn) 21 ... .i::txe6 22
�b3 �d8 23 cs ds 24 .l:txds! cxds 2 s
�xds .i::!. x cs 2 6 �xe6+ '>t>h7 27 g3 li'id4
28 �xd4 �xd4 29 'it>g2 and White kept
an edge in R.Buhmann-M.Klenburg,
Neustadt an der Weinstrasse 2009.
18 cxds
Retreating with 18 .i::!. el gives Black
no problems after 18 ... 'fid6.
18...cxds
19 Ilxes!
The point of White's play.
19...�xes 20 �xes 'fib6 21 �b2!
This was Shirov's innovation. White
opens the e-file to protect the e6-pawn
and introduces the threat of a battery
on the a1-h8 diagonal. other moves are
less dangerous:
a) 21 �d4 �xe6 22 �cs was played
in J.Vilela-R.Vazquez, Cuba 1996. Here
Markos suggests 22 ... .i::tfd8 23 �d2 li'ifs
when White probably does not have
quite enough compensation for the
exchange.
b) 21 'fid2 'illx e6 (21 ... '>t>h7 22 .i::!.e l as
was played by a young Radjabov, but is
not as trustworthy) 22 .i::!.e l (or 22 �d4
I!xf3) 22 ....i::txf3 ! is a thematic idea.
Black gives back the exchange to ruin
White's pawn structure and give his
knight a strong outpost. After 23 gxf3
li'ifs the knight is well placed to both
attack and defend. One example, which
was a serious upset, T.Nyback­
M.Matthiesen, Copenhagen 2003: 24
�g3 �f7 2s .i::t c1 d4 26 .i::t a 'fids 27
'fid3 .i::!.e8 28 a4 hS 29 'it>g2 �e6 30 .i::!.xb7
h4 31 .l:txa7 h3+ 32 '>t>gl 'flel+ 33 'fifl
li'ixg 3 34 fxg3 'fle3+ 0-1.
21....i::ta d8
This is considered best by Markos.
Black creates the possibility of shutting
115
A ttacking C h e s s : The King 's I n dian, Vo l u m e 1
the long diagonal with ...d4. others:
a) 21 ...'>t>h7 was Black's original try.
Then 22 'ii'e2 d4 23 h4 .i::!.f6 24 .l:tel
'ii'xb4 2S a3 'ii'd6 26 h S .i::!.af8 27 'ii'e4
ll:lc6 28 hxg6+ '>t>g7 29 �cl! 'ii'e 7 30
�xh6+ 'it>xh6 31 'tlfli4+ 'it>xg6 32 �xc6
bxc6 33 .l:te S leads to a further theoreti­
cal divide:
al) 33 ... 'ii'xe6? 34 .i::!.xe6 .i::!.xe6 3 S
'ii'g4+ '>t>f7 3 6 'ii'xd4 was winning for
White in A. Shirov-T.Radjabov, Linares
2004.
a2) 33 ... .i::txe6 is a better try: 34
'tlfli s+ '>t>g7 3 S .i:tgs+ 'ii'x gs 36 'ii'x gs+
.i::!. g 6 3 7 'ii'e s+ 'it>g8 38 'ii'xd4 .l::!.f7 39
'ii'd8+ (after 39 g3 .i::!.f g7! White will
have trouble advancing his pawns)
39 ....l:tf8? (39 ...'>t>h7 has to be a better
try, with the idea of 40 g3 .l:tfg7!, al­
though White will still certainly try to
probe with 40 'ii'd3 !) 40 'ii'd7 .l::!.f7 41
'ii'e 8+ '>t>g7 42 g3 and White won the
ending in M.Casella-1.Zenyuk, U S
Championship, S an Diego 2004. I can
still remember preparing this line with
Casella before the game ...
b) 21 ...'ii'xb4 is Bologan's recom­
mendation. However, after 22 .l:tbl
.i::!.ac8 (instead 22 ...'ii'c s 23 �d4 'ii'c 7 24
�b2 'ii'c s 2s �d4 'ii'c 7 26 �b2 was
drawn by repetition in W.Fademrecht­
F.Gerhardt, correspondence 2004, but
White could also consider 24 'ii'd 2 'it>h7
2 S h4 with compensation) 23 �g7 (23
�es 'ii'a 3 eyes cl and Bologan gives 24
'ii'd2 'it>h7 2S .i::!.xb7 'ii cl+ 26 �dl 'ii'xd2
27 .l:txe7+ 'it>g8 28 .l::!. g 7+ with perpetual
check) 23 ...'ii'xbl 24 'ii'xbl 'it>xg7 2 S
116
'ii'b 2+ (Bologan only gives 2 S h4 b6 as
unclear) 2s ... .i::!.f6 26 h4 b6 27 g4 g s 28
�g2 .l::!.c4 29 'ii'e s White kept some
pressure in M.Boccia-S .Giuliani, corre­
spondence 2004.
22 'ii'e 1
This is an interesting try. White pro­
tects the b4-pawn and may still con­
tinue with 'ii'e s at some point.
a) 22 'ii'e 2 'ii'xb4 23 'ii'e S allows
Black to execute his idea with 23 ... d4
when he stands better.
b) 22 'ii'd3 'ii'x e6 23 'ii'd4 transposes
to variation 'd2', below.
c) 22 bS 'ii'x e6 23 'ii'd4 ll:lfs 24 'tlfli8+
'>t>f7 2 S 'ii'h 7+ 'it>e8 26 �g4 .l::!.f 7?! gave
White excellent compensation after 27
'tlfli8+ 'it'd 7 28 'ii'c 3 .i::!. c 8 29 'ii'd2 .l::!.c4 30
�f3 in B.Lalith-A.Lahiri, Bhubaneswar
2010, but Black can improve here with
26 ... 'ii'f 7! 27 .l:tel+ ll:le7 28 'ii'xf7+ (28
'ii'xh6 'ii'xf2+) 28 ....i::txf7 when he is bet­
ter.
d) 22 'ii'd2 has been White's most
common choice. Then following
22 ... 'ii'x e6! (after 22 ... d4 23 .i::!. el .i::!.f6 24
..lli.g4 'it>h7 2S h4 h S 26 �h3 Black could
 Th e M a r de/ Plata Va ria tio n : 9 b4 li'i h 5 1 o .i::!. e 1
hardly move in R.Buhmann-T.Nedev,
Hersonissos 2007) we have:
dl) 23 'ii'x h6 d4 24 'ii'd 2 .i::!.xf3 ! 2 5
gxf3 li'if5 is given by Markos. This is a
typical set-up for Black. White's king
position is also loose and the strong d­
pawn and knight give Black compensa­
tion for the pawn.
d2) 23 'ii'd4 looks scary, but the
checks do not lead to anything: 23 ...li'if5
24 'tlfli8+ 'l;f7 25 �7+ 'l;e8 26 'ii'xb7
.l::!.f7 27 'ii'h 5+ �8 when Black is able to
get 'castled' again and his pieces coor­
dinate well. Still, White retains com­
pensation for the exchange.
d3) 23 I!el could lead to perpetual
check after 23 ... 'ii'f7 (forced) 24 'ii'x h6 (if
24 'ii'd4 'l;h7 25 'ii'xa7 li'ic6 Black can
fight for the initiative after 26 'ii'b6
.i::!.fe8 or 26 'ii'c 5 d4 27 b5 J:tfe8! ) 24 ... d4
and then:
d32) 25 'ii'd2 li'if5 is unclear, but 26
�dl? (26 �e4) 26 ...'ii'xa2 27 h4 .i::!. de8
28 .i::!.xe8 .i::!.xe8 29 g4 'ii'e6 30 'l;h2 'ii'd6+
31 rt;g2 'ii' d5+ 32 �h2 li'ixh4 33 rt;g3
'ii'g 2+ and 0-1 was a significant upset
in A.Fier-K.Sai, Bhubaneswar 2010.
d32) 25 .i::!.x e7 (forcing the draw)
2 5 ... 'ii'xe7 26 'ii'xg6+ 'ii'g7 27 'ii'e6+ 'ii'f7
(27... .l::!.f7 28 �h5 does not help Black,
while 27 ... 'l;h8 28 �h3+ �7 29 'ii'g4
'ii'b 1+ 30 �dl 'ii'xb2 31 'ii'h 5+ is also a
draw) 28 'ii'g4+ 'ii'g 7 29 'ii'e6+ with per­
petual.
22 .i::!.d 6
•.•
Black could also try 22 ...d4 23 .i::!.d1
(after 23 �g4 h5 24 �h3 .i::!.de8! Black
intends ...li'if 5) 23 ... h 5! ?.
23 I!dl
Putting pressure on the d5-pawn
makes it more difficult for Black to play
... li'if5. Instead 23 'ii'c 3 d4 24 'ii'b3 .i::!.f6
brings White's initiative to a standstill,
while 23 'ii' d2 .l:!.xe6 24 �d4 (after 24
'ii'x h6 d4 25 'ii'd2 Black plays the usual
2 5 ... .i::txf3 ! 26 gxf3 - not 26 �xd4? .i::!.d6
winning - 26 ...li'if5 with thematic play)
24 ... 'ii'd6 25 'ii'xh6 is again not as scary
as it looks after 25 ...li'if5 26 �8+ 'l;f7
27 �7+ 'l;e8 28 �c3 J:te7 29 'ii'h4 d4!?
with unclear play.
23 .i::txe6 24 'ii'c3 .i::!.ef6 2 5 'ii'd 2
.•.
2 s....i::txt3
This is certainly safer than 2 5 ... l!f4!?
117
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
26 �es .i::!.xb4 27 'ii'xh6 when White has
the initiative. After 27 ...'l;f7 28 'ii'd2
'l;e8 29 �d4 .i::!.xd4?! (29 ... 'ii'd6 is better,
but White has compensation for the
exchange) 30 'ii'xd4 'ii'xd4 3 1 .i::!.xd4
White had the better ending in
V.Golod-N.Huschenbeth, Hamburg
2008.
26 gxf3 .i::!.xf3 27 .i::te1 .i::tf7 28 �d4 �d6
29 'ii'xh6
Or 29 ..i.es 'ii'b6 30 �d4 'ii'd6 with a
repetition.
29 ... .l:!.h7 30 �cs .i::!.xh6 31 �xd6 lllfs
The endgame is equal (Golod).
A4) 13 �e3
This move used to be the main line
and it has been played more than any
other move. Nowadays it is not consid­
ered so dangerous for Black.
13... h6
Instead 13 ... 'l;h8 14 exfs! was
pointed out by Markos and this looks
quite good for White.
Bologan states that Black should
play 13 ... cxds immediately, although
he gives no explanation why this is so. I
118
do not like this move order because it
allows the extra option of 14 exdS! ?
which scores rather well for White. In­
stead after the more common 14 cxds
h6 lS tlle6 �xe6 16 dxe6 fxe4 17 tllxe4
tllxe4 18 �xe4 dS we just transpose
back into the main line with 13 ...h6.
A more serious option is 13 .. .f4!?.
After 14 �cl it seems as though White
has just wasted time, but the centre
has been stabilized and now the bishop
will head to a3. Following 14... h6 lS
tll e6 �xe6 16 dxe6 tllc 8 17 bS 'ii' e 7
(worse is 17 ...'ii'e 8?! 18 bxc6 bxc6 19 cs
'ii'xe6 20 �a3 dxcs 21 ll'la4 tll b 6 22
tllxcs 'ii'f7 23 'ii' c l �c4 24 �dl! with
the better chances for White in
V.Kramnik-B.Gelfand, Belgrade 1997)
18 bxc6 bxc6 19 cs dxcs 20 �a3 tllb6
21 ll'la4 .i::tfd8 22 'ii'b 3 tllxa4 23 �xa4
�xe6 24 �xcs ll'ld7 2s .i::!.e cl tllxcs 26
I!xcs .i::!.d6 the position is equal.
14 tlle6 �xe6 15 dxe6
1s ...fxe4
Instead 1S .. .f4 16 �cl takes us back
to 13 .. .f4 in the previous note.
Black could also consider l S ... gs.
 Th e M a r de/ Plata Va ria tio n : 9 b4 li'i h 5 1 o .i::!. e 1
This has not had much success, but it
does not look too bad. After 16 exfs
li'ixfs 17 �d3 li'ixe3 18 .i::!.x e3 "ille7 19
.i::!.d1 .i::!.ad8 (19 ... flixe6 20 "illxd6 �xc4 21
bS gives White the initiative) 20 li'ie4
Black has:
a) 20... g4?! 21 li'ixf6+ flixf6 22 �xg4
�xf2+ 23 '>t>hl gave White a big advan­
tage in V.Kramnik-F.Nijboer, Wijk aan
Zee 1998. After this game nobody
wanted to play 1s ... gs anymore.
b) 20 ... ds 21 li'ixf6+ flixf6 22 cxds
cxds 23 e7 (23 �xds? .i::!.xds) 23 ... �xe7
24 �xds+ 'it>h8 is given as slightly bet­
ter for White by Kramnik. This certainly
looks pretty miserable for Black.
c) 20 ... li'ixe4 21 �xe4 fllxe6 22 bS
gives White good compensation for the
pawn.
d) 20 ... �xe6 and now Kramnik
claimed that White has the upper hand
after 21 li'ixd6, but 21 ... li'ie8 22 cs li'ixd6
23 cxd6 I!f6 looks tenable, since Black
will win the d-pawn and have an extra
pawn. White certainly has some com­
pensation, but Black does not really
appear to stand worse.
16 li'ixe4 li'ixe4 17 �xe4 dS 18 cxdS
White can also refrain from this ex­
change with 18 �c2, but this does not
look too dangerous: 18 ... b6 (Black plays
as if the exchange had already oc­
curred on dS, but 18 ... d4 19 �d2 and
now 19 ... .i::!.f6 or 19 ... "illd6 could seek to
punish White's omission as the a2-g8
diagonal is not yet open) 19 "illg 4 (19
cxds would transpose to the main line)
19 ....i::!.f6 20 bS!? (White tries to break
apart Black's centre) 20 ... fllc8 (taking
immediate aim at the e6-pawn; 20... d4
21 .i::!.adl "illd6 is an alternative) 21 bxc6
(2 1 �b3 d4 22 �d2 "ilia 23 c s ! ?, main­
taining the e6-pawn, is also possible)
21 ... d4 22 �d2 fllxe6 23 fllxe6+ .i::!.xe6
(Black is likely to pick off the c6-pawn,
but White's bishop-pair will be fair
compensation) 24 �e4 was Huang
Qian-Ju Wenjun, Hefei (rapid) 2010.
Now Black could play 24....i::!.c8 because
2s .i.ds? fails to 2s ...li'ixds 26 cxds .i::!.d6
when the pawns begin to fall.
18...cxds
19 �C2
This leads to a complicated middle­
game. White can also head for an end­
ing with 19 �cs dxe4 20 flixd8 .i::!.fxd8
21 �xe7 .i::!.e 8 22 �cs. The position is
identical to the main line of Line A2,
but here White's pawn is still on b4. If
anything, this should help Black be­
cause his queenside will not be under
any pressure. Both 22 ... b6 23 �e3 I!xe6
24 a4 �f8 2S bS (N.lbrayev­
S.Mamedyarov, Khanty Mansiysk 200S)
2s ...a6! and 22 ...I!xe6 23 I!xe4 .i::!.d8 24
119
A ttacking C h e s s : Th e K i n g 's I n dian, Vo l u m e 1
h4 a6 2 s a4 .l:!.ds 26 ii.e3 .l::!.c6 27 .l:tb1 bS
(Z.Gyimesi-V.Baklan, Baile Tusnad
2oos) are equal.
19 b6
...
This is a very important move. Black
must prevent ii.e3-cS or his centre will
fall apart. He has a strong centre and
White's e6-pawn is likely to fall, but
Black must be careful here because
White has the bishop-pair and the
safer king. Indeed, Black must try to
keep his pieces well coordinated and
should not necessarily be in a hurry to
win the e6-pawn. Any misstep could
result in a serious tactical accident. As
long as Black is not too careless, his
chances are good, though, and in prac­
tice he has done pretty well from this
position.
20 'lli'g4
White must play actively or the e6-
pawn will fall. Black is likely to win it
anyway, but he must not allow White
too much activity in return. Some other
tries:
a) 20 ii.b3 immediately is inaccu­
rate because White will not be attack-
120
ing the g6-pawn after 'lli'g4. After
20 ... 'lli'd6 21 'lli'g4 h S 22 'lli'h 3 (E.Ovod­
V.Rajlich, Budapest 2002) 22 ... e4, in­
tending ....:tfs, the b4- and e6-pawns
are weak.
b) 20 ii.a4 'lli'd6 21 ii.d7 is mis­
guided. The e6-pawn is secure, but
White's bishop is totally out of play.
After 21 ...'l!Vxb4 22 .l::!.b1 'l!Vh4 23 f3 J:tfs
24 ii.f2 'lli'f6 2 S ii.g3 h S 26 h3 .l::!.f8 27
�hl? 'l!Vg s 28 ii.h2 e4! Black broke
through to White's king in
V.Malakhatko-F.Jenni, Istanbul Olym­
piad 2000.
c) 20 bS 1'6d6 21 ii.cl is not an un­
common idea, but here it is painfully
slow. Now 21 ... .l:tad8 22 ii.b3 �h7 23 a4
'l!Vxe6 24 ii.a3 was V.Malakhatko­
S.Sivokho, Polanica Zdroj 1999, and
here 24 ... e4! 2S .:tel ii.es 26 ii.xe7 (26
.l::!.x e4 ii.xh2+) 26 ...'l!Vxe7 is good for
Black because 27 ii.xds? loses to
27 ...'l!Vg s, intending ....l::!.xd6 or ...ii.g7.
20 e4
...
The alternative 20....l:tf6 is not so bad
theoretically, but Black must work for a
draw after 21 .l:tadl 'lli'd6 22 ii.b3 .l:td8
 Th e M a r de/ Plata Varia tio n : 9 b4 l'i::i h 5 1 o .l:. e 1
(22 ... .l:r.xe6 can be met by 23 J:txds l'i::ixds
24 .l:tdl when White will recoup his ma­
terial with a slight edge) 23 bS (other­
wise, 23 .l:txds oversimplifies: 23 ...l'i::ix ds
24 .l:r.dl 'l!Vxe6 2s 'l!Vxe6+ J:txe6 26 .l:txds
.l:txds 27 ii.xds �f7 and the ending is
drawn) 23 ... 'l!Vxe6 24 'l!Vxe6+ .l:txe6 2S
ii.cl! �7 26 ii.a3 e4.
Here White can try to improve his
position or he can win back the pawn
immediately:
a) 27 ii.xe7 seems to be the most ac­
curate: 27 ... �xe7 28 .l::!.x ds .l:txds 29
ii.xds .l::!.e s (instead 29 ... ii.c3 30 .l:te2 .l::!.e s
31 ii.xe4 .l::!.xbs 32 ii.d3+ .l:tes 33 .l:r.xes+
ii.xes 34 ii.xg6 is Kramnik-Shirov, be­
low) 30 ii.xe4 (White must play this,
because 30 .l:txe4 �d6 31 .l:txes �xes 32
ii.f7 gs 33 a4 was already agreed
drawn in V.Mikhalevski-M.Klinova, Gi­
braltar 2008). Now 30 ... .l:txbs 31 ii.d3+
.l::!. e s 32 J:txes+ ii.xes 33 ii.xg6 is Kram­
nik-Shirov again, so Black should play
30...�6 31 a4 which is Mikhalevski­
Melekhina below, but with a couple of
tempi less for each side.
b) 27 g3 and here:
bl) 27 ... .l:td7 28 ii.xe7 �xe7 29 J:txds
.l:txds 30 ii.xds ii.c3 (also better for
White is 30 ....l::!.e s 31 ii.xe4 .l:txbs 32
ii.xg6+ �6 33 ii.d3 .l:tb2 34 a4) 31 J:te2
J:tes 32 ii.xe4 and White was much bet­
ter in V.Kramnik-A.Shirov, Linares 1998.
b2) 27....:tes ! 28 .l::!.d4 with a final di­
vide:
b21) 28 ....l:te6 29 ii.xe7 �xe7 30
.l:txds .!:txds 31 ii.xds J:tes 32 ii.xe4 �f6
33 a4 was V.Mikhalevski-A.Melekhina,
Philadelphia 2008. White can certainly
play on for a fair while, although Black
did manage to draw this game.
b22) 28 ... t'i::ifs ! is best. Then Mik­
halevski gives 29 .l:r.dxe4 l'i::id4 30 .l::!.x es
ii.xes 31 �g2 ii.f6 with equality.
21 .l:r.ad1
White threatens ii.xe4, so Black
must move his queen off the d-file.
21 'l!Vc7 22 ii.b3
•••
Instead 22 ii.a4 still looks like the
wrong idea. After 22 ....l::!.fs (Black could
look into 22 ... 'l!Vc3 or 22 ...'l!Vc4) 23 ii.d7
.l:taf8 24 bS, thematic is 24 ... hS when
White must commit his queen: 2S 'lli'h4
(2S 'l!Ve2 could be met by 2s ... ii.es or
2s ... ii.c3 26 .l:r.c1 'l!Ves) 2s ...ii.f6 26 'lli'g 3
(26 'l!Vh3 is possible, but the queen does
not seem well placed here) 26 ... ii.es 27
'lli'h4 ii.f6 repeats. Instead 24 ..'lli'c3 2 S
'l!Ve2 'l!Ves 26 'lli'd2 'lli'd6 (now 26 ...h s can
be met with 27 'l!Vb4! ii.f6 28 'lli'a4 'lli'b8
29 .l::!.e 2!, preparing to double rooks on
the d-file) 27 h3?! (not 27 ii.xh6 ii.xh6
28 'lli'xh6 .:ths, but 27 a4 gs 28 'l!Vc1 is
possible) 27...g S ! 28 a4 was B.Lalith­
V.A.Rajesh, Delhi 2010, and here Black
121
A tt a c k i n g C h e s s : Th e King 's I n d i a n, Vo l u m e 1
should play 28 ...d4! 29 ii.xd4 (not 29
'l!Vc2 dxe3 30 .l::!.xd6 exf2+) 29 ....l::!.ds 30
.l:!.xe4 .l:tf4! when White has some prob­
lems.
22 .l::!.fs!
...
This move protects dS, cuts off the
e6-pawn and Black may also double
rooks on the f-file or play ..J::t e s. White
has tried many things here, but has
been unable to prove much.
23 .l:td2
White prepares to double rooks on
the d-file and leaves the dl-hS diagonal
open for his queen. There are many
other moves:
a) 23 'l!Ve2 .l::!.af8 24 'lli'a6 (24 .l::!.fl and
1/2 -Yi was T.Radjabov-J.Moreno Carnero,
Pamplona 2002) 24 ... ii.c3 2S .!:tel 'l!Ves
26 .l:tedl ii.xb4 27 ii.d4 (27 'l!Vxa7!?)
27 ... 'lli'b8 28 'lli'a4 ii.d6 29 g3 hS gave
Black a good position in A.Janusonis­
D.Chocenka, correspondence 2oos.
b) 23 ii.d4 ii.xd4 (23 ... .l:taf8!?) 24
l:i'.xd4 'l!Vc3 2S 'lli'dl .!::taf8,, 26 .!::t e 2 'l!Vc7 27
g3 'l!Vc6 28 'l!Vc2 1/2-Yi was S.l skusnyh­
A.Motylev, Moscow 1999.
c) 23 .:tel 'Ii' d6 24 .l::!. e dl .l::!.af8?! (bet-
122
ter is 24...'l!Vxb4 2 S .l::!.c7 h S , leading to
variation 'e2' below with Black having
avoided the note to White's 24th move
there) was L.Gerzhoy-V.Rajlich, Buda­
pest 2002, and White could play 2S
'l!Vxe4! here.
d) 23 bS is logical. The pawn is not
hanging anymore, c6 is controlled and
ii.c1-a3 becomes a possibility.
Here Black has:
dl) 23 ...'l!Ves?! 24 ii.cl! .l::!.af8 2s ii.a3
hS 26 'lli'g 3 'l!Vxe6 27 ii.xe7 'l!Vxe7 28
ii.xds+ �h8 29 ii.xe4 .l:tg s 30 'lli'd3 and
White was up a pawn with better posi­
tion in L.Lenic-A.Jankovic, Sibenik 2006.
d2) 23 ....l:taf8 24 .l:te2 'l!Vc8 2S ii.cl
(White should consider 2S .l:tc2 'l!Vxe6 26
.!::tc7) 2s ... hs 26 'i!Vh4 ii.f6 27 'l!Vh3 �h7
28 ii.a3 'l!Vxe6 29 g3 .l::td8 30 f3 'lli'c8 gave
Black a good position in M.Bosiocic-
1.Saric, Split 20os.
d3) 23 ...�h7 24 .l:te2 (after 24 ii.cl
.!::taf8 2S ii.a3 .l:txf2 26 ii.xe7 'l!Vxe7 Black
is much better following 27 .l:txds e3 or
27 ii.xds 'l!Vcs) 24 ... .l:td8 2s ii.cl ii.es 26
g3 hS 27 'l!Vh3 'l!Vcs is comfortable for
Black, who is well coordinated.
 Th e M a r de/ Plata Varia t i o n : 9 b4 l'i::i h s 1 0 .l:. e 1
e ) 23 .l:te2 is similar t o the main line,
but here White may bring the rook to
c2 as well.
Black has:
el) 23 ....l:td8 24 bS hS 2S 'lli'h 3 �h7
26 ii.cl 'l!Vcs 27 .l:tc2 'l!Vxbs? (better is
21...'l!Vd6) 28 .l:te7 'l!Vb4 29 J:txa7 'l!Vcs 30
'l!Ve3 gave White a winning position in
N.Alfred-S.Nagle, Budapest 2004.
e2) 23 ...'lli'c 3 24 .l:tc2 (it is probably
better to play 24 bS .l:taf8 2S .l:tc2 with
some initiative for White, V.Mikha­
levski-A.David, Vlissingen 2000)
24 ... 'l!Vxb4 2S l:te7 h S 26 'l!Vh3 (and here
better is 26 'l!Ve2 'lli'd6 27 l:txe7 'l!Vxe7 28
.l:txds .l:txds 29 ii.xds .l:tc8 30 g3 ii.f6,
which is pretty even) 26 ... as! 27 g4!
(forced, because 27 a4 is good for Black
and after 27 .l:tb7? a4 28 ii.c2, as in
J.Sales-Lim Chuing Hoong, Kuala Lum­
pur 20os, Black has 28 ... .l::!.c 8! 29 ii.bl d4
winning, as pointed out by Mik­
halevski) 27 ...hxg4 28 'lli'h4 .l:te8 29 ii.gs
e3! and Black has good counterplay
according to Bologan.
e3) 23 ...h S ! takes immediate advan­
tage of the fact that the queen cannot
retreat along the dl-hS diagonal. After
24 'l!Vh4 ii.f6 2 s 'lli'g3 ii.es (2s ...'lli'c 6!?
could be tried) 26 'lli'h4 .�f6 27 'lli'g 3
ii.es 28 'lli'h4 ii.f6 the game was drawn
by repetition in Xu Jun-R.Ponomariov,
Shenyang 2000.
Returning to 23 .l:td2:
23 'l!Vc3!?
.•.
This is the most enterprising move.
Others:
a) 23 ... ii.c3? 24 .:tel 'l!Ves loses to 2S
.l:txds ! l'i::ixds 26 'l!Vxg6+ �h8 2 7 'lli'x h6+
�g8 28 'lli'g 6+ �h8 29 'l!Vh6+ �g8 30
e7!, R.Sherbakov-S.lskusnyh, Tula 1999.
b) 23 ...J:taf8 24 .l:tedl 'lli'c6 was
S.Savchen ko-R.Ponomariov, Alushta
2000. Now 2S .l:tc2 'lli'd6 (worse is
2s. . .'l!Vxe6 26 l:te7 - Ponomariov) 26
'l!Vxe4 ii.es ! 27 .l:txds ii.xh2+ 28 �hl
.l:txds 29 ii.xds ii.f4 was level in
A.Koz lowicz-F.Schwarz, correspondence
2003.
c) 23 ...'l!Vc6 24 bS!? 'l!Vxe6 2S f3 exf3
26 .�f2 'lli'f6 27 .l:txe7 'l!Vxe7 28 ii.xds+
.l::!.x ds 29 .l:txds 'l!Ve2 30 gxf3 .l:tf8 31
'l!Vxg6 'lli'xf3 led t o a draw in Z.Straka­
D.Fridman, correspondence 2004.
123
A tt a c k i n g C h e s s : Th e King 's I n d i a n, Vo l u m e 1
d) 23 ... �h7 24 .l:tedl 'Ii'c6 25 .l:tc2 (25
'l!Vh3 .l:td8) 2s ...'l!Vxe6 (2s ...'lli'd6 26
'l!Vxe4!) 26 .l:tc7 was B.Lalith-A.Lahiri,
New Delhi 2009. Now 26 ... hs with the
idea of ...'lli'd6 would be unclear.
e) 23 ... .l:td8!? is untried but looks
sensible.
24 'l!Vd1
Instead after 24 .l:tedl 'l!Vxb4 25
ii.d4?! (better is Panczyk and llczuk's 25
'l!Ve2 when 2s ....l:td8 26 'l!Va6 would give
White counterplay) 2s .. Jk8 26 'l!Ve2
ii.xd4 27 .l:txd4 'l!Vcs 28 g4 .:tes 29 'lli'd2
.l:tf8! 30 'lli'x h6 .l:txe6 Black stood well in
J.Werle-A.David, Amsterdam 2000.
24 ....:tds
Black has some other options:
a) 24... .l:taf8 25 bS (25 g4 .l:tes)
2s ...�h 7 (2s ...'lli'b 4!? has been proposed
by Panczyk and llczuk) 26 J:tc2 'l!Ves 27
ii.cl d4 28 ii.a3 d3 29 ii.xe7 (29 J:td2
ii.f6 30 .l::!.xd3 .l:txf2 31 .l:td7 ii.h4! is un­
clear) 29 ... dxc2 30 ii.xc2 .l:te8 31 .l:txe4
'l!Vxbs 32 ii.a3 'lli'ds 33 'lli'xds 1/2-1/2,
A.Eriksson-G.Glatt, correspondence
2002. After 33 ... .l:txds 34 e7 ii.f6 35 ii.b3
.l:td7 36 ii.a4 .l::!.e xe7 the position is to-
124
tally equal.
b) 24 ...'l!Vxb4 25 g4 .l:tes (instead
2s ... .l:tf6 26 .l:td4 'l!Vd6 27 .l:txds lllxds 28
'lli'xds 'lli'xds 29 ii.xds .l:te8 30 .l::!. c 1 gives
White compensation for the exchange)
26 .l:td4 'l!Vc3 ! (not 26 ... 'l!Vas 27 ii.d2 'lli'bs
28 a4 'lli'c6 29 ii.b4 with a big advan­
tage for White - Tsesarsky) 27 ii.f4 .l:tf8
28 ii.xes ii.xes 29 .l::!.xds (forced)
29 ... ii.xh2+ 30 �g2 .l:txf2+ 31 �xf2
lllxds 32 ii.xds 'lli'g 3+ 33 �1 (not 33
�e2? 'IJVf3+ 34 �d2 'IJVd3+ 35 �Cl Ji.f4+
36 �b2 ii.es+ 37 �cl 'lli'c3+ 38 'l!Vc2
'l!Vxel+) 33 ...'lli'h 3+ with a draw.
2s bs 'lli'b4
Black can also play 2s ... 'lli'c8 when
26 ii.d4 ii.xd4 27 .l:txd4 'l!Vcs transposes
to the main text, and in fact this was
the actual move order of Xu Jun-Ye Ji­
angchuan, below. However, White has
an extra option here in 26 g4!? .l:tff8 27
ii.xds ii.c3 28 ii.xe4 ii.xd2 29 ii.xd2
'l!Vxe6 30 'l!Ve2 with good compensation.
26 ii.d4
But not 26 a4? ii.c3.
26....:tdfB 27 a4 ii.xd4 28 .l:txd4 'l!Vcs 29
.l:te2
 Th e M a r de/ Pla ta Va r i a t i o n : 9 b4 tb h s 1 0 .l:. e 1
White covers the vulnerable f2-
pawn.
29 .l::!. es 30 h3
...
Instead 30 .!:tdd2 'lli'd6 31 .l:tc2 .l::!.xe6
32 'lli'c1 .l:tf7 33 .l:te7 (not 33 'lli'xh6? .l:th7)
33 ...tt:ifs 34 .l:txf7 �xf7 3S 'lli'd2 �f6!? 36
ii.xds .:tes 37 ii.b3 'lli'xd2 38 .l:txd2 .:tcs
gave Black a slightly better ending due
to his active pieces in J.Skeels-S.Boyd,
correspondence 2006.
After 30 h3, the game Xu Jun-Ye Ji­
angchuan, Shanghai 2001, was agreed
drawn, but Black could play on after
30 ...J:tffS OT 30...�g7.
B) 12 f3
This move is less forcing than 12
ii.f3 and it is for that very reason that it
has become very popular. Indeed, a lot
of tension remains in the position, al­
though Black has a broader choice
here:
B1: '12 c6
•..
82: 12 ¢>hl
•••
Black's other main move, 12 ... tl:ihs,
is not looking very attractive these days
because of Bareev's idea 13 cs tl:if4 14
ii.c4 �h8 lS .l:tbl!.
Bl) 12 c6
...
Black increases the tension as in
Line A while also threatening ... 'lli'b6+.
This choice is somewhat more risky
than Line B2, but it is quite playable
and leads to a double-edged game.
White has two fundamental lines:
811: 13 'ili>h1
812: 13 .i.eJ
Other moves do not mind the g1-a7
diagonal and are rare: for example, 13
bS h6 14 tl:ie6 ii.xe6 lS dxe6 fxe4 16
fxe4 'lli'b6+ 17 �hl ex bS 18 ex bS .l:tac8
when White must be careful, because
19 tl:ids loses to 19 ...tl:iexds 20 exds
tl:ie4 and 19 ii.b2 fails to 19 ....l::!.xc3 20
ii.xc3 tl:ixe4.
Instead 13 'l!Vb3 is reasonable, al­
though it takes pressure off the d-file
and 13 ...h6 14 tl:ie6 ii.xe6 lS dxe6 'lli'c 8
16 .l:tdl .l:td8 looks fine for Black.
125
A ttacking C h e s s : T h e King 's I n d i a n , Vo l u m e 1
811) 13 �h1
White safeguards his king by re­
moving it from the g1-a7 diagonal.
13 ... h6
Black forces the play.
14 l2Je6 ii.xe6 15 dxe6 lbe8
Threatening to play ... lbe7xe6-d4.
Bologan suggest the rarer 1s ... 'l!Vc7
when 16 bS (instead 16 ii.b2 .l:tad8 17
'l!Vb3 .l:tfe8 18 .l:tadl �h7 19 ii.d3 fxe4 20
lbxe4 lbxe4 21 J:txe4 l2Jfs 22 f4 J:txe6 23
fxes was V.Epishin-1.Nemet, Biel 1996,
and here Epishin suggests 23 ...ii.xes)
16 ....l:tfd8 17 bxc6 bxc6 18 .l:tbl 'lli'c 8 19
'lli'a4 'l!Vxe6 20 .l:tb7 as 21 ii.e3 .l:td7 22
.l:tb6 .l:te7 23 .l:tebl lbd7 is fairly level in
his view.
16 'l!Vb3
After 16 bS Black should play
16 ... 'lli'c8, rather than 16 ... cs?! 17 lbds
lbc7 18 l2Jxc7 'lli'xe7 19 exf s gxfs
(19 ... lbxfs may be better, but even if
Black wins the e6-pawn White will al­
ways have light-square compensation)
20 g4 when White had the initiative in
D.Lima-A.Romero Holmes, Leon 1996.
16...l2Jc7
126
17 c5
This is almost always played, but
Black should also be ready for 17 .l:td1! ?
which looks natural. Here he can con­
sider:
a) 17 ... l2Jxe6 18 cs dS 19 exds l2Jd4
20 .l:Ixd4 exd4 21 dxc6+ (not 21 d6+
�h7 22 dxe7 'l!Vxe7 and Black wins ma­
terial) 21 ... �h7 22 cxb7 .l:tb8 23 ii.f4 is
very messy, but looks roughly balanced.
b) 17 ...�h7 18 cs dS 19 exds cxds 20
lbxds lbexds (worse is 20...lbcxds 21
ii.c4 e4 22 .l:tb1) 21 ii.c4 e4 22 ii.xds
(after 22 .l:tb1 'lli'h 4! 23 g3 'l!Vh3 24 ii.xds
lbxds 2s 'lli'xds .l:tad8 26 'l!Vb3 J:td3 ! 27
'l!Vc2 exf3 Black has good play)
22 ...lbxds (not 22 ... ii.xal 23 ii.xb7 with
a big advantage) 23 .l::!.x ds 'lli'f6 24 e7
'l!Vxe7 2S ii.b2 ii.xb2 26 'l!Vxb2 .l:tad8 and
the position is level.
11... ds 18 exds cxds
After 18...lbexds 19 ii.c4 White
maintains pressure.
19 ii.b2
Black has a tremendous centre, but
it is not easy to hold it together if
White plays aggressively.
 Th e M a r de/ Plata Varia tio n : 9 b4 l'i::i h s 1 0 .l:. e 1
19 b6!?
...
Black seeks to break up White's
queenside pawn chain. Other moves
are risky:
a) 19 ... 'lli'c8 20 t'i::ib 5! l'i::ixb5 (this is
probably a mistake; instead 20 ... a6 21
l'i::id6 'l!Vxe6 22 l'i::i xb7 was E.Bareev­
R.Polzin, Rethymnon 2003, and here
22 ... e4! ? would be unclear, while
20 ... a5! ? is another option) 21 ii.xb5
'l!Vxe6 22 ii.xe5! ii.xe5 23 f4 gives White
a comfortable advantage.
b) 19 ...'l!Ve8 is the most popular
move. Black removes his queen from
the d-file and keeps an eye on the b5-
square as well. Markos stops here,
pointing out that Black has scored well,
but the position is actually not so easy
for him. White has:
bl) 20 .l:tadl .l::!. d 8 21 a4 a6 22 b5
'it>h7 23 ii.a3 l'i::i xe6 24 bxa6 l'i::id4
(probably better is 24 ... bxa6 2 5 ii.xa6
l'i::id4 26 'l!Va2 'lli'd7 with an unclear posi­
tion) 25 'l!Vxb7 .l::!.b8 26 'lli'e7 l'i::ic2 27 l'i::ib5
was E.Bareev-V.Topalov, Dortmund
2002, and now 27 ... .l:tc8 is probably
okay for Black. However, 27 ii.b5 looks
stronger, since after 27...'lli'f7 28 c6!
l'i::i x a3 29 l'i::i x d5 l'i::i xd5 30 'lli'xf7 .l::!.xf7 31
.l:txd5 White is much better.
b2) 20 a4 a6 21 .l:tadl (the sacrifice
21 l'i::i xd5 does not work as well here:
21 ...l'i::ic xd5 22 l:!.ad1 'lli'c6 23 ilc4 l:!.fd8
24 ii.xe5 ii.xe5 2 5 .l::!.x e5 t'i::if6 26 .l:teel b5
27 'l!Vc3 l'i::ifd5 28 ii.xd5 .l:txd5 29 axb5
axb5 30 'lli'f6 'lli'e8 gave Black good
chances in Van Wely-Kotronias, Euro­
pean Championship, Warsaw 2005)
21 ....l:td8 22 l'i::i x d5 .l:txd5 (worse is
22 ...l'i::ic xd5 23 ii.c4 'lli'c 6 24 b5 axb5 2 5
axb5 'l!Vxe6 26 ii.xe5 ii.xe5 27 f4) 23
ii.c4 .l:txdl 24 .l:txdl 'l!Vb8 25 l:!.d7 and
even though White has only a pawn, he
has good compensation for the piece,
L.Van Wely-D.Stellwagen, Dutch Cham­
pionship, Leeuwarden 2005.
b3) 2o l'i::ix d5!
This immediate sacrifice is more
dangerous than in Van Wely-Kotronias
above. Leaving the respective a-pawns
on their home squares helps White.
After 20 ...l'i::i c xd5 21 .l:tadl there is:
b31) 21 ...'l!Vc6 22 ii.c4 .l:tfd8 23 b5!
'l!Vxc5 24 ii.a3 ! is the difference because
127
A tt a c k i n g C h e s s : The King 's I n d i a n, Vo l u m e 1
the a-file is not open for Black's rook
and 24 ... 'l!Vc7 2 5 ii.xd5 seems better for
White, although I suspect Black can
hold after 25 ... .l::!.xd5 26 .l:txd5 l2Jxd5 27
'lli'xd5 'l!Vc3! 28 e7+ �h7 29 .l::!.g l 'l!Vxa3
30 'lli'd8 .l::!.xd8 31 exd8'1li' 'l!Vxa2.
b32) 21 ... �h7 22 ii.c4 l2Jf6 23 ii.xe5
b5 24 ii.fl gave White good compensa­
tion for the piece in L.Van Wely­
V .Kotronias, European Team Champi­
onship, Gothenburg 2005. After 24... a6,
Bologan suggests 25 �gl (instead of
the game's 25 g3) 2S ....l::!.a7 26 a3 with
some advantage.
20 cxb6
Worse is 20 c6 �h7 21 .l::!.adl l2Jxc6
22 l2Jxd5 lbxe6 23 lbxb6 'l!Vxb6 24 'l!Vxe6
.l::!.fe8, which was drawn here in L.Van
Wely-F.Nijboer, Dutch Championship,
Hilversum 2006. However, instead of
24....l:tfe8 Black could play 24 ...'l!Vxb4 25
.l:t bl (25 'l!Vxc6 'l!Vxb2 leaves Black a
pawn up) 25 ... .l:tf6 26 'Ii'c4 'l!Vxc4 27
ii.xc4 .l:td6 and despite White's bishop­
pair, the extra pawn gives Black reason
to play on.
20 ...axb6 21 .:tad1 'lli'b8
128
Worse is 21 ...'lli'd6 because of Mik­
halevski's suggestion 22 lbb 5! lbxb5 23
ii.xb5 d4 (23 ...'l!Vxe6 allows the familiar
24 ii.xe5 ii.xe5 25 f4) 24 a4 .l::!.fc8 25
ii.d7 .l:tc7 26 b5 with the idea of ii.a3
when White has pressure.
22 lbxds
After 22 lbb5 lbxe6 23 ii.c4!? dxc4
(Black should avoid 23 ...l2Jf4 24 ii.xd5+
l2Jfxd5 25 l2Jc3) 24 'l!Vxc4 'lli'c8 Mik­
halevski gives 25 lbc7 'l!Vxc7 26 'l!Vxe6+
.l::!.f 7 27 .l::!.d 7 'lli'c6 28 .l:txe7 'l!Vxe6 29 .l:txe6
.l:txa2 30 ii.xe5 ii.xe5 31 .l:t6xe5 �g7 32
.l:te6 l:!.f6 33 .l:te7+ .l:tf7 with a draw.
22 lbcxds 23 .l:txds lbxds 24 'lli'xds
•••
.:tds 2 s 'lli'b3 �h7 26 ii.bs
White's advanced d-pawn gives him
enough compensation for the ex­
change, but no more. Black soon sacri­
fices back the exchange to equalise.
26 'lli' d 6 27 ii.d7 .l:!.a7 28 .l:td1 'lli' e 7 29
.••
a4 .l:taxd7 30 exd7
Here a draw was agreed in E.L'Ami­
F.Nijboer, Dutch Championship, Hilver­
sum 2008.
812) 13 ii.e3
 The M a r de/ Plata Va r i a t i o n : 9 b4 lb h s 1 0 .l:. e 1
White covers the diagonal towards
his king by developing.
13 ...ii.h6
This is the main line. Black tries to
exploit the pin along the c1-h6 diago­
nal and threatens .. .f4. Instead 13 ... h6 is
possible as well. After 14 lbe6 ii.xe6 lS
dxe6 -.a (Markos does not mention
this possibility, giving only 1s ...lbeS 16
-.b3 lbc7 17 .l::!.adl when White has the
upper hand) 16 .l:!.b1 (16 �3 .l:tfdS 17
bS -.cs is equal according to Bologan)
16 ... .l:!.fdS 17 bS -.cs 1S bxc6 bxc6 19
-.a4 �h7 (19 ...-.xe6 and 19 ...f4 are also
possible) 20 .l:tedl -.xe6 21 .l:tb7 .l:td7 22
-.a6, as in D.Komljenovic-A.Kuzmin,
Benasque 1999, then 22 ... hs 23 .l:txd7
-.xd7 24 ii.cs lbcS 2s ii.xd6 lbxd6 26 cs
..tfS is equal (Bologan).
14 h4
Too tame is 14 ii.d2. After 14.. .f4 lS
lbh3 (1s lbe6 ii.xe6 16 dxe6 -.cs wins
the e6-pawn right away) 1s ...ii.xh3 (not
1s ... gS?l 16 lbf2 when White was better
in E.Lobron-R.Barcenilla, Yerevan
Olympiad 1996) 16 gxh3 ii.gs Black has
a good game.
Instead 14 cs seems to ignore
Black's threat, but after 14 ...f4
(14 ...cxdSl? l S cxd6 cxd6 could be an
improvement) l S cxd6 fxe3 16 dxe7
-.xe7 17 d6 -.ds 1S ii.c4+ �g7 19 lbe6+
ii.xe6 20 ii.xe6 �6 (or 20 ... lbh s l ? with
the idea of ...lbf4 and ....l:tf6) 21 lbe2
.:tads it was unclear in E.Bareev­
J.Balcerak, German League 2001.
14...cxds 15 cxds
White has also tried lS exds. This
gives him access to the e4-square, but
it hands fs to Black. After 1s _.f4 16 ii.f2
ii.xgs 17 hxgs lb hs 1s cs lLifs 19 lbe4
(following 19 ii.bs lbhg3 20 lbe4 lbxe4
21 .l:txe4 -.xg s 22 -.d2 the position is
unclear according to Mikhalevski)
19 ... lbfg3 20 ii.xg3 (after the 20 ii.c4
lbxe4 21 .l:txe4 of M.Bosboom-F.Nijboer,
Wijk aan Zee 1997, Black should simply
play 21 ...-.xgs) 20 .. .fxg3 21 ii.c4 ii.fs 22
lbxd6 -.xgs 23 lbxfs .l:txfs 24 d6+ �g7
Black had a strong attack in
T.Kotanjian-V.Kotronias, Thessaloniki
2007.
1s ...ii.d1
Black will have to play this move at
some point, both to cover the bS­
square and to allow the e7-knight to
move off the queen's path to gs. Black
has also tried the immediate 1 S .. .f4 16
ii.f2 ii.xgs 17 hxgs lbhs 1s .l:tc1 and
then :
a) 1S ... lbg3 19 -.d2 ii.d7 (no better
is 19 ...lbxe2+ 20 -.xe2 ii.d7 21 lbbs) 20
ii.bs and White was better in
T.Radjabov-F.Nijboer, Wijk aan Zee
2001.
129
A ttacking C h e s s : Th e King 's I n dian, V o l u m e 1
b) 18 ... ii.d7 19 ii.bs ! gave White
some advantage in J.Berkvens-P.Smir­
nov, Patras 2001, but 19 lbbs ii.xbs 20
ii.xbs l2Jc8 gave Black counterplay in
A.Huzman-1.Caspi, Israeli League 2008.
16 'lli' d2
This move was recommended by
Markos and it is White's latest try.
White connects his rooks and there is
also a subtle point in lining up White's
queen against the gs-square, as we
shall see. The alternative is 16 .l:!.cl a6
17 bS and then:
a) 11 ...f4 18 ii.f2 ii.xgs 19 hxgs lbhs
20 'l!Vb3 as (not 20 ... l2Jc8 21 bxa6 bxa6
22 'l!Vb7) 21 b6 l2Jc8 22 ii.bs .l:tf7 23
ii.xd7 .l:txd7 24 lbbs was S.lvanov­
H. Rudolf, Panormo 2001, and here
Black should take the pawn with
24 ... lbxb6, although White has a strong
initiative after the simple 2S .l:tc2, in­
tending to double rooks on the c-file.
b) 11... ii.xbs! ? 18 ii.xbs axbs 19
lbxbs fxe4 20 fxe4 l2Jxe4?! 21 'lli'g 4?!
(White could play 21 lbxe4 ii.xe3+ 22
.l:txe3 'lli'b6 23 'l!Vb3! lbfs 24 l:tcel with
an obvious advantage) 21 ...ii.xgs 22
130
hxg s .l:ta4! was suddenly much better
for Black in S.Brynell-V.Kotronias,
Stockholm 2006. Correct, though, is
20...ii.xgs 21 hxgs l2Jxe4 22 l2Jc7 l2Jfs 23
l2Jxa8 'l!Vxa8 with some compensation
for the exchange in an unclear posi­
tion.
16 f4
...
Instead 16 ... a6 is a sneaky move be­
cause Black will have to play this at
some point anyway. After 17 a4?! (a
waste of a tempo; it would be much
more testing to just bring a rook to cl)
11...f4 18 ii.f2 ii.xgs 19 hxgs lbh s 20 bs
l2Jc8 21 b6 'l!Vxgs 22 .l:!.ecl 'lli' d 8 23 'l!Ve1
lbg3 ! 24 ii.xg3 fxg3 2S 'l!Vxg3 'l!Vxb6+ 26
'lli'f2 'l!Vas left Black a pawn up in
L.Murzin-V.Kotronias, Linares 2003.
11 ii.f2 ii.xgs 18 hxgs lbhs
This is the typical structure for the
13 ii.e3 line. Black will win the g s-pawn
and have some kingside pressure, but
White will obtain a strong initiative on
the other wing.
19 .l:tec1
Instead 19 ii.bs looks premature.
After 19 ... ii.xbs 20 lbxbs a6 21 l2Jc3
 Th e M a r de/ Plata Varia tio n : 9 b4 lb h s 1 0 .l:. e 1
�h8 ( 21 . .lbc8 looks better) 22 .l::!. ecl
lbg8 23 lbe2, as in l.lkonnikov­
V.Kotronias, Gausdal 2008, Black
should play 23 ....l:tf7 to cover C7, with
the idea of ...'l!Vxgs with an unclear po­
sition.
19 ...a6
Now this is necessary because after
19 ... lbc8 both 20 lbbs and 20 ii.bs look
strong.
2o bs
White consistently plays to open the
queenside. Markos also mentions the
strange 20 'l!Ve1?!, but this is playing on
the wrong flank and Black gets good
counterplay after his 20 ...lbc8 21 ii.h4
h6! 22 gxh6 gs 23 ii.f2 .l:tf6.
20 ... lbcB 21 bxa6 bxa6 22 .l:tab1 'l!Vxgs
23 .l:tb7 .l:tf7
Now Black threatens ... ii.h3. Note
that Black has trouble creating direct
threats with his queen and knight be­
cause ...lbg3 will be met by ii.xg3 when
Black cannot recapture with the pawn
- this is one point of White's queen
being on the d2-square.
24 ii.fl
Instead 24 .l:tcbl lbg3 (24 ...'lli'd8 2 S
'l!Ve1 g s 26 l:i'.b8 .l:txb8 27 .l:txb8 was bet­
ter for White in Xu Jun-S.Safin, Bled
Olympiad 2002, because the a6-pawn
is so weak) 2 S ii.xg3 'l!Vxg3 26 ii.d3 is
unclear according to Markos.
24 ...'lli'dB
Black wants to play ... lbg3 and able
to meet ii.xg3 with .. .fxg3. Still, the
immediate 24 ... lbg3 2S ii.xg3 'l!Vxg3
looks playable as well.
25 'l!Ve1 gs 26 lbd1 lbf6 27 J:tcc7 .l:tb8!?
Black gives back the pawn to relieve
some of White's queenside pressure.
27 ... ii.e8 was an alternative.
28 ii.xa6 .l::!. x b7 29 .l:txb7 g4
Here Black had counterplay in L.Van
Wely-F.Nijboer, Dutch Championship,
Leeuwarden 2001.
82) 12 ...�hB
This is considered the most reliable
move nowadays, mostly due to the ef­
forts of Radjabov. Because 12 f3 is a bit
slow, Black makes a useful waiting
move without committing his pawn
structure. White has tried many ideas
131
A ttacking Chess: Th e King 's I n dian, Vo l u m e 1
here but Black's defences continue to
hold.
B21: 13 .i.e3
B22: 13 lUe6
823: 13 cs
824: 13 .:.b1
White has also tried 13 bS and here:
a) 13 ... lbhs 14 ii.fl (14 l2Je6 ii.xe6 1 S
dxe6 l2Jf4 16 hf4 exf4 17 .:tel c 6 18
ii.d3 fxe4 19 ii.xe4 ii.es was about
equal in P.Ricardi-D.Lemos, Berazategui
2007) 14 ...ii.f6 (14 ... l2Jf4!?) lS lbe6
ii.xe6 16 dxe6 lbg7 (16 ... l2Jf4!?) 17 ii.h6
.l:te8 18 l2Jds l2Jg8 19 ii.xg7+ ii.xg7 20 cs
(20 exfs looks like a better try) 20.. .fxe4
21 fxe4 dxcs 22 'l!Vb3 c6 23 l2Je3 'lli'e7 24
bxc6 bxc6 2S ii.c4 l2Jf6 and Black took
over the initiative in F.Abbasov­
N.Mamedov, Baku 2009.
b) 13 ... l2Je8 14 ii.e3 ii.f6 (after 14.. .f4
lS ii.f2 lbxds 16 lbxds 'l!Vxgs 17 cs
White has some initiative for the
pawn) lS l2Je6 ii.xe6 16 dxe6 lbg7
(16 .. .f4 17 ii.f2 lbg7 18 cs is given as
unclear by Carlsen, although after
18 ... l2Jxe6 19 cxd6 'lli'xd6 20 'lli'xd6 cxd6
2 1 ii.c4 lbd4 22 .l::!. a cl .l:tfc8 Black can
hardly complain with the result of the
opening) 17 ii.h6 l2Jxe6! 18 ii.xf8 'lli'xf8
19 cs!? lbxcs (19... dxcs!? 20 'lli'd7 'lli'c8)
20 ii.c4 ii.gs 21 'l!Ve2 'l!Vh6 22 .l:tad1 .l:tf8
gave Black sufficient compensation for
the exchange in M.Carlsen-V.lvanchuk,
Foros 2008.
821) 13 ii.e 3
132
This has often been played, but is
not very popular nowadays.
13 ...ii.h6
Black plays along the lines of 'B12',
but 13 ...lbhs ! ? is also interesting. After
14 cs (perhaps White will look for im­
provements here) 14...l2Jf4 practice has
seen:
a) lS ii.c4 fxe4 (Wells gives the line
lS ... lbexds! 16 lbxds 'l!Vxgs 17 hf4
exf4 18 lbxe7 fxe4 19 l2Jxa8 ii.h3 20
ii.fl exf3 21 'lli'x f3 ii.g4 22 'l!Vxb7 f3
when Black indeed has a strong attack)
16 lbgxe4 lbfs 17 ii.f2 l2Jd4 has scored
well for Black; one example being 18
.:tel ii.fs 19 �hl 'lli'd 7 when Black had a
reasonable position in A.Rychagov­
N.Pokazanjev, Novokuznetsk 2008.
b) lS .:tel fxe4 (Wells's idea works
here too, although not quite as well:
lS ... lbexds!? 16 lbxds 'l!Vxgs 17 hf4
exf4 18 lbxe7 fxe4 19 l2Jxa8 ii.h3 20
ii.fl exf3 21 'lli'xf3 ii.g4 22 'l!Vxb7 f3 and
here, with White's rook on cl instead of
al, he can hold: for example, 23 .l:tc2
f2+ 24 .l:txf2 ii.d4 2S �hl ii.xf2 26 .l::!.e 7
'l!Vh6 27 .l:tf7 dxcs 28 'l!Ve7 .l:txf7 29 'lli'xf7
 Th e M a r de/ Plata Va riat io n : 9 b4 lb h s 1 o .l:.e 1
id4 30 'lli'e 8+ �g7 31 'l!Ve7+ with a
draw) 16 lbgxe4 lLifs 17 ii.f2 lbd4 18
ifl was slightly better for White in
E.Bareev-1.Cheparinov, Amsterdam
2008.
14 lbf7+
The slow 14 ii.d2 should not give
Black any trouble after 14.. .f4 lS lbe6
(1s h4 ii.xg s 16 hxgs lbh s) 1s ... ii.xe6
16 dxe6 c6 with ideas like ...'lli'b 6+,
... 'lli'c8 and maybe ... lbe8-c7.
Instead 14 h4 f4 lS ii.f2 ii.xg S 16
hxg s lbh s 17 cs lbg8 looks like a better
version of Line B12 for Black because
... �h8 is a useful move as it clears g8
for his knight. After 18 'lli'd2
(M.Thejkumar-K.Ramu, Hyderabad
2006), Black can just play 18 ... 'l!Vxgs
(18 ... ii.d7 and 18 ... a6 are reasonable
too) 19 lbbs 'lli'd8 with the idea of
... lbg3 and ... 'lli'h4.
14....:txf7 15 ii.xh6 f4
White has obtained the bishop-pair,
but it has taken time and Black has a
clear plan on the kingside.
16 ii.gs
White wants to keep his dark-
squared bishop. Instead 16 cs makes a
lot of sense because 16 ... lbeg8 is rather
forced. Then 17 ii.g s h6 18 ii.h4 gs 19
ii.f2 .l:tg7 20 g4 was R. Pogorelov-1.Nataf,
Reykjavik 2004, and here Black should
probably just continue with 20 ... hs.
White could also play 16 g4 imme­
diately. After 16 ...lbeg8 17 ii.gs hS
(Black can also play 17 ...h6 18 ii.h4 gs
19 ii.f2 hS, heading into the main line)
18 gxhs (18 h3 hxg4 19 hxg4 .l:th7 20
�g2 fbh6 21 :J:!.h l fbf7 22 J:txh 7+ �Xh7
23 ii.h4 �g7 should give Black enough
play) 18 ... .l:th 7 we have:
a) 19 hxg6 .:th S 20 h4 was given as
much better for White by Gofshtein,
but Black has 20 ....i::!.x gs+ 21 hxgs lbxds
22 �f2 lbxc3 23 .l:th1+ �g7 24 l:th7+
�xg 6 2s 'l!Vh1 'l!Vxgs 26 .l:tg1 'l!Vxgl+ 27
�xgl lbxe2+ 28 �2 lbg 3. This looks
quite risky, but there is no mate: 29
'lli'h4 ii.e6 30 .l:txc7 lbf6 just looks very
unclear.
b) 19 h4 .!:txhs 20 �2 was
Z.Gyimesi-P.Acs, Budapest 2004. Now
Black has 20 ....l:txh4! 21 ii.xh4 lbg4+ 22
fxg4 'lli'xh4+ 23 �g2 'lli'g 3+ 24 �1 wh en
he can take the draw or play 24...'l!Vxc3
with good compensation for the ex­
change.
16...lbegB
16 ...h6!? looks even stronger be­
cause 17 ii.xh6?! lbh7 18 h4 lbg8 19
ii.g s lbxgs 20 hxgs 'l!Vxg s is nice for
Black. Instead 17 ii.h4 gs 18 ii.f2 .l:tg7 is
an improved version of the main line
because Black has not spent a tempo
on ... lbeg8. Indeed, after 19 cs g4! 20
133
Attackin g C h e s s : Th e King 's I n dian, Vo l u m e 1
ii.h4 l'Llg6 21 ii.xf6 'lli'xf6 22 fxg4 h S !
Black had good play i n A.Vlasov­
V.Tukhvatullin, Sterlitamak 2008.
11 g4 h6 1s ii.h4 gs 19 ii.f2 hs 20 h3
20...hxg4!
This looks better than 20 ....:th7 21 cs
l'Lle7. In this position Black has lost two
tempi compared to 16 ...h6 and after 22
'it>g2 l'Llg6 23 .l:thl White was somewhat
better in B.Gelfand-T.Radjabov, Sochi
2008, although Black managed to win
anyway.
21 hxg4 l'Llxg4 22 fxg4 f3! 2 3 ii.xf3 'lli'f6
24 .l:tf1!
After 24 'it>g2? ii.xg4! 2S J:te3 .l:taf8
White's position falls apart - Mik­
halevski.
24...'l!Vxf3 25 'lli'xf3 .l:txf3 26 l'Llbs! ii.xg4
21 l'Llxc1 .l:tc8 28 l'Llbs .l:txc4
Black has equal chances in this
double-edged endgame.
822) 13 t'Lie6
The most direct move. White does
not even wait for ...h6. Van Wely liked
this approach for a while, playing it in
four of his games against Radjabov.
134
Van Wely won the first one, but then
lost the next three, at which point he
headed towards other variations ...
13...ii.xe6 14 dxe6
14 ...l'Llhs
This has been played the most, but
it may not be best. Others:
a) 14.. .fxe4 lS fxe4 l'Llc6 is a typical
idea, but it does not fare well here: 16
l'LldS! l'Llxe4 (compared to the position
where .l:tbl and ... h6 are thrown in,
16 ...l'Llg8 is not so good because after
17 bS l'Lld4 18 e7 l'Llxe7 White has 19
ii.gs! .l:te8 20 ii.g4 h6 21 ii.xe7 .l:txe7 22
l'Llxe7 'l!Vxe7 23 'lli'd 3 and Black does not
have enough) 17 ii.f3 l'Llf6 18 bS! was
very good for White in L.Van Wely­
T.Radjabov, Khanty-Mansiysk 200S.
b) Van Wely suggests that 14 ....l::!.e 8!?
is Black's best in New in Chess.
(seefollowing diagram)
The game A.Grischuk-T.Radjabov,
Elista 2008, continued lS l'Llds fxe4 16
l'Llxf6 (after 16 fxe4 Black could try
16 ...l'Lleg8!?; the alternative is 16 ...l'Llxe4
 The M a r de/ Plata Varia tio n : 9 b4 lb h s 1 0 .l:.e1
1 7 ii.f3 lbf6 when both 1 8 ii.gs lbeg8
and 18 lbxf6 ii.xf6 19 ii.xb7 .l:tb8 seem
satisfactory for Black)
16 ... ii.xf6 17 fxe4 lbc6 18 ii.g4 lbd4
19 1i.e3 c6 (Black should also consider
both 19 ... lbxe6 20 ii.xe6 .l:txe6 21 'lli'ds
ii"c8 and 19 ...'l!Ve7 20 1i.xd4 exd4 21 .l:tfl
:ts) 20 1i.xd4 exd4 21 J:tf1 .l:tf8 22 'lli'd3
-W-e7 when Black was better, but the
position was drawish. Indeed, after 23
:n 1i.g7 24 .l::!.af1 .l:txf 3 2s .l:txf3 .l:tf8 26
:xf8+ 'lli'xf8 27 'lli'f3 hS 28 'lli'xf8+ hf8
29 1i.e2 'itig 7 30 bS �6 31 bxc6 bxc6 32
.i.d3 'it>xe6 the game was drawn. This
method of play looks like a simple solu­
tion to 13 lbe6, which will probably not
be seen so much anymore. Neverthe­
less, we will still examine the sharp
14 ...lbh s.
15 g3
White has a serious alternative in
lS cs. After 1s ...lbf4 16 ii.c4 fxe4 17
.l:txe4 (17 fxe4 lbc6 with the idea of
... lbd4 is at least equal for Black)
11 ... lbfs and here:
a) 18 ii.xf4 exf4 19 lbds .l::!.e8 20
lbxf4 c6 (not 20 ... ii.xal? 21 'l!Vxal+ lbg 7
2 2 e 7 .l:txe7 23 lbxg6+ hxg6 2 4 .l:th4
mate) 21 e7 .l::!.xe7 22 lbe6 'l!Vb8 (after
22 ...'lli'd7 23 lbxg7 .l:txe4 24 fxe4 lbe3 2S
'l!Ve2 'l!Vxg7 26 .:tel lbxc4 27 'l!Vxc4 dxcs
28 bxcs White is a little better accord­
ing to Mikhalevski) 23 lbxg7 .l:txg7 24
cxd6 (Mikhalevski suggests 24 1i.e6 !?
dS 2s .l:te2 'lli'f4 26 ii.xfs gxfs 27 'l!Ve1
.l:tag8 28 'l!Vc3) 24 ...lbxd6 2s .l::!. d 4 lLifs
(playing for a win; 2s ... lbxc4 26 .l:txc4 is
just equal) 26 J:td3 .l:te7 27 'lli'd2 'l!Ve S 28
.l:tdl .l::!. a e8 was level in L.Van Wely­
T.Radjabov, Monaco (blindfold) 2007,
although Black did indeed manage to
win.
b) 18 g3 lLih 3+ 19 'it>g2 lLigs 20 .:tg4
(Radjabov gives both 20 ii.xg s 'l!Vxgs 21
lbds c6 22 cxd6 lbxd6 23 .l:tg4 'l!Vh6 24
e1 cxds 2s exf8'1li'+ .l:txf8 26 'lli' x ds lLifs
27 .l:te4 lbe3+ 28 J:txe3 'l!Vxe3 29 .l:tdl b6
and 20 e1 'l!Vxe1 21 lbds 'lli'd8 22 ii.xgs
'l!Vxg s 23 lbxc7 lbe3+ 24 .l:txe3 'l!Vxe3 2S
lbxa8 e4! 26 f4 .l:txa8 27 cxd6 .l:td8 as
unclear) 20...lbxf3 ! 21 'it>xf3 e4+ 22
lbxe4 1i.xa1 23 ii.gs lbe1+ 24 'it>g2 ii.es
was rather unclear in L.Van Wely­
T.Radjabov, Biel 2007.
135
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1

1s ...ii.f6 L'Ami) 30....l:txe4! 31 'l!Vxe4 gxfs 32 'lli'd3

A typical idea. Black wants to play a6 33 a4 'lli'e6 34 ii.xg7 �xg7 3S gxfs
...lbg7xe6. 'l!Ves is level according to L'Ami.
16 c5 16 ...f4
This does not lead to much for
White. More interesting is 16 exfs lbxfs
17 g4 (else ... l2Jd4) 17 ... e4 18 l2Jxe4
ii.xa1 19 ii.gs and here:
a) 19 ...ii.f6 20 l2Jxf6 l2Jxf6 21 gxfs
(Mikhalevski prefers 21 ii.d3, but after
21 ... l2Je7 22 'l!Va1 �g7 23 ii.xf6+ .l:txf6 24
gs 'lli'f8 the position is unclear) 21...gxfs
22 f4 .l:tg 8 23 'lli'd4 cs 24 'l!Va1 'lli'e7 2S
�2 .l:taf8 26 bxcs dxcs 27 ii.f3 .l:tg6 28
'l!Ves when White has excellent com­
pensation and went on to win in 17 �g2
V.Mamonovas-C.Gatto, correspondence Worse is 17 g4?! lbg7 18 ii.c4 lbc6
2007. 19 cxd6 cxd6 20 lbe2 (after 20 lbbs
b) 19 ..id4+ 20 �hl 'lli'e8 21 e1 lbxe1 lbxe6 21 'lli'xd6, as in T.Bakre-A.Kuzmin,
(21 ..Af7 22 gxfs ii.es 23 f4 l2Jxf4 24 f6 Dubai 2002, Black has 21 ...lbg s! 22
lbe6 2S ii.h4 lbd4 is messy, but looks 'lli'xd8 .:taxd8 23 l:tfl a6 24 l2Jc3 l2Jd4
better for White) 22 'lli'xd4+ lbg7 23 bS with the threats of ... .l:tc8 and ...l2Jgxf3)
l2Jg8 (L'Ami relates how at this point 20....l:!.c8 21 ii.ds lbxb4 22 .l::!.b1 l2Jc2 23
both players were playing very quickly!) .l:tfl b6 was much better for Black in
24 ii.d2 'lli'f7 (after 24 ... 'l!Ves White has L.Van Wely-T.Radjabov, Wijk aan Zee
the simple 2s 'l!Vxes dxes 26 ii.c3 or can 2007.
play 2S 'l!Ve3 .l:tae8 26 'l!Vxa7! .l::!.a8 27 17 lbc6 18 cxd6 cxd6 19 lbds
•••

'l!Vxb7 .l::!.xa2 28 'lli'xe7 which looks very

strong) 2s ii.c3 h6 (2s ... .l:tae8 26 'l!Vxa7
b6 27 'l!Va3 h6 28 ii.b2 .l:ta8 29 'l!Vb3 �h7
30 ii.d1 .l:tae8 31 'l!Vc3 also looked good
for White in V.Mamonovas-G.Staf, cor­
respondence 2007) 26 f4 .l:!.ae8 27 .:tfl
l2Jf6 28 ii.f3 was 1.Cheparinov-T.Radja­
bov, Sofia 2008. Here 28...l2Jxe4 29 ii.xe4
�h7! 30 fs (not 30 ii.xb7?! lbe6 31 'l!Vxa7
lbxf4 32 'lli'd4 cs 33 bxc6 'lli'e7 34 �xf4
.l:Ixf4 3S 'lli'xf4 .l:tf8 with a strong attack -

136
 Th e M a r de/ Plata Vari a t i o n : 9 b4 lb h s 1 0 .l::!. e 1
19 ...lbd4
19 ... lbg7 also looks good.
20 ii.b2 lbxe6 21 g4 lbhg7
Chances were fairly balanced in this
murky position in A.Shirov-T.Radjabov,
Wijk aan Zee 2001.
823) 13 cs
This obvious and direct move was
considered harmless for a long time,
but lately White has been trying to
squeeze something out of the end­
game that arises.
13...h6
Black has also tried 13 ...dxc5 14
bxc5 h6 1 5 lbe6?! ii.xe6 16 dxe6 'lli'd4+!
17 'lli'xd4 exd4 18 lbb5 fxe4 with good
counterplay in Z.Gyimesi-T.Radjabov,
Moscow 2005, but Avrukh's suggestion
15 lbh3 ! fxe4 16 fxe4 ii.xh3 17 gxh3
indeed loo ks good for White.
Black could check out 13 ... a5!?.
14 lbe6 ii.xe6 15 dxe6 ds
Black cannot sneak back into Gy­
imesi-Radjabov with 15 ... dxc5, because
of 16 'l!Vxd8! (not 16 bxc5 'lli'd4+)
16 ....l:tfxd8 17 bxc5 with a big plus.
16 exds lbfxds 11 lbxds 'lli'x ds!
Worse is 17...lbxd5 18 'l!Vb3 (both 18
ii.b2 and 18 ii.c4 lbxb4 19 .l:tbl look
dangerous too) 18 ... c6 19 ii.c4 with
pressure.
18 1'6b3
White's hopes for an edge have be­
come associated with this finesse. By
threatening ii.c4 he induces Black to
exchange queens, after which the a-file
will be opened for the al-rook.
The immediate 18 'l!Vxd5 should also
not be taken too lightly. After 18 ... lbxd5
19 ii.c4 lbxb4 20 .l:tb1 lbc6! White has:
a) 21 J:txb7 lba5 22 .l:tb4 lbxc4 23
J:txc4 .!:tfe8 is no problem for Black.
b) 21 ii.d5 .!:tfd8 22 ii.xc6 bxc6 23
.l:tb7 .l:te8 24 .l:txe7 .l:txe6 25 .l:td1 ii.f8 was
fine for Black in G.Rechlis-B.Avrukh,
Israeli League 2003. White has com­
pensation for the pawn, but no more
than that.
c) 21 ii.e3 .l:tab8 22 .l:tbdl .!:tfd8 23 .l:td7
.l:txd7 24 exd7 ii.f8 25 f4 .l:td8! (worse is
25 ... e4 26 .l:tbl!, A.Korobov-N.Huschen­
beth, Pardubice 2008) 26 fxe5 .l:txd7 27 e6
.l:td8 and if anything Black is better.
137
A t ta c k i ng C h e s s : Th e K i n g 's Indian, Vo l u m e 1
d) 21 e7! ? looks like White's best try,
intending 21 ....l:tfe8 (not 21 ...l2Jxe7 22
.l:txb7) 22 i..f7 .l::!. x e7 23 i..x g6.
Black has:
dl) 23 ... e4 is considered Black's best
by Avrukh, but Black will suffer here: 24
i..xfs!? (24 fxe4 fxe4 25 .l:txe4 i.. d4+ 26
.l:txd4 l2Jxd4 27 i..b 2 .l:td8 28 .l:tdl �g7
29 i..c2 .l:te2 30 .l:txd4 .l:txc2 is given as
equal by Avrukh) 24 ...exf3 25 .l:txe7
lbxe7 26 i..e4 i.. d4+ 27 �fl fxg2+ 28
�xg2 .l::!. g 8+ 29 �h3 b6 30 cxb6 cxb6 31
i..x h6 should be a draw, but White can
certainly play on for a while with his
bishop-pair.
d2) 23 ... .l:tf8 24 .l:txb7 is given as
slightly better for White by Avrukh , but
24... e4 gives Black counterplay.
d3) 23 ... l2Jd4 also looks okay after 24
.l:txb7 .l:te6 or 24 i..b 2 e4!.
18 'lli'x b3 19 axb3 lbc6 20 .l:ta4
...
White protects the b4-pawn and
may double on the a-file. Instead 20
i..bs .l::!.fe8 21 i..xc6 bxc6 22 bS .l:txe6
(22 ...cxbs 23 .l:ta6 gives White compen­
sation) 23 b6 cxb6 24 cxb6 cs 25 bxa7
.l:te7 is equal according to Giri.
1 38
20 .l:tfeB!
...
Radjabov's move looks like Black's
best road to equality. Black has also
pushed his a-pawn:
a) 20 ... as?! is loosening. Instead of
21 bS?! l2Jd4 22 i..c4 .l:tfe8 23 i..e 3?! l2Jc2
24 J:te2 l2Jxe3 25 .l:txe3 e4 26 i..d s .l::!.ad8
27 fxe4 fxe4 28 i..xe4 .l:!.xe6 when Black
was better in Z.Gyimesi-A.Volokitin,
German League 2007, the simple 21
bxas! just looks much better for White.
b) 20... a6 is a little bit better: 21
i..b2 (21 bS axbs 22 i..xbs .l::!.fe8 23
J:txa8 .l:txa8 24 i..b2 .l:te8 25 i..xe 5 i..x es
26 .l:txes �g7 27 J:tds .l:te7 should hold)
21 ....l:tae8 22 i..c4 .l:te7 23 i..d s .l:td8? 24
i..xc6 bxc6 25 i..xe s i..x es 26 .l:!.xes with
a healthy extra pawn for White in
Z.Gyimesi-E.Hagara, Jenbach 2009. A
better try is 23 ...l2Jd8 24 .:ta2 (24 i..x es
i..x es 25 .l:txes c6 26 i..c4 �g7) 24 ... c6
25 i..c4 lbxe6 26 i..xe 5 l2Jc7 27 i..x g7+
.l:txg7, as given by Golubev, but White is
somewhat better here as well.
21 i..c4
This does not lead to much, but
White has trouble getting anywhere:
 Th e M a r de/ Pla ta Va ria tio n : 9 b4 lb h s 1 0 .l:. e 1
a) 21 ii.b2 .l:txe6 22 bS lbd4 23 ii.c4
J:tee8 24 ii.f7 .l:tf8 25 ii.xg6 lbxb3 26
ii.xes lbxcs is level.
b) 21 ii.e3 has been used success­
fully by Arun Prasad. Now Black has:
bl) 21 ...lbd8 22 .:teal lbxe6 23 .l::!.x a7
.l:tab8 24 ii.c4 (or 24 .l:tdl lbd4 25 ii.c4)
24.. .f4 2 5 ii.d2 e4 26 fxe4 ii.xal 27 J:txal
.l:tbd8 28 ii.c3+ �h7 29 h4 and White
had good compensation in 5.Arun
Prasad-A.Lahiri, Gurgaon 2010.
b2) 21 ... a6 22 bS axbs 23 .l:txa8 .l:txa8
24 ii.xbs .l:te8 25 .l:td1 .l::!.xe6 26 .l:td7 was
clearly better for White in 5.Arun
Prasad-V. Rajesh, Chennai 2010, as
26 ....l:te7? loses to 27 ii.xc6.
b3) 21 ... .l:txe6 is the simplest. Now
22 ii.c4 J:tee8 23 bS lbd4 transposes to
variation 'c'.
c) 21 bS lbd4 22 ii.c4 lbxe6 23 ii.e3
lbd4 and now:
cl) 24 ii.xd4 exd4 25 J:txe8+ .l:txe8 26
.l::!.xa7 d3 27 ii.xd3 ii.d4+ 28 �fl ii.xcs
29 l:txb7 ii.b6 and with White's rook
trapped on b7, Black cannot be worse.
c2) 24 ii.f7 lbxb S ! ? (Black could also
play 24.. 1H8 25 ii.xd4 exd4 26 ii.xg6
d3) 25 ii.xe8 (25 ii.xg6 J:tf8 is fine for
Black) 2s ....l:txe8 26 ii.d4 .l:te7 27 ii.b2 c6
28 .l:taal �g8 29 .l::!.adl �f7 and Black
had two pawns for the exchange in
M.Kozakov-R.Vidonyak, Lvov 2009.
21... a6! 22 bs axbs 23 .l:txa8 .l::!. xa8 24
ii.xbs .l::!.e8!
Black finally rounds up the e6-
pawn.
2 s .l::!. d 1
25 ii.c4 lbas ! does not help White,
so he activates his rook.
2s ....l:txe6 26 .l:td7 ii.f8! 27 .l:txc7 ii.xcs+
28 �fl .l:te7 29 .l:txe7
Perhaps simpler was 29 .l:tc8+ �g7
30 ii.xc6 bxc6 31 J:txc6 with equality.
29 ... ii.xe7 30 ii.b2
After 30 ii.xh6 gS! White will have
to give up a pawn to save his bishop.
30...lbd4 31 ii.xd4 exd4
Despite Black's extra pawn, the
endgame is a draw, A.Grischuk­
T.Radjabov, Linares 2009.
824) 13 l:tb1
This flexible move has become
White's latest attempt to get some­
thing from the position. The rook gets
off the long diagonal and takes a far­
sighted look at the b-file which, as we
have seen, can open up in some varia­
tions. This line was also recommended
by Markos in Beat the KID. Moreover,
after a few reversals with 12 lbe6, Van
Wely turned to this variation to beat
Radjabov. He was not going to aban­
don the Bayonet and he praised Rad­
jabov for sticking to his guns as well,
139
A ttacking C h e s s : Th e King 's I n dian, V o l u m e 1
expressing that he would rather 'lose a
game than lose face'. It will certainly be
interesting to see what their future
battles bring.
13 ... h6
Forcing the pace. After 13 ... tllh s 14
cs tllf4 lS il.c4 we have transposed to
Bareev's 12 f3 tllh s 13 �bl! which was
discussed at the beginning of Line B.
The main alternative is 13 ... c6 14
il.e3 and now:
a) 14 ... il.h6 lS h4 cxds 16 cxds f4 17
il.f2 tllh s 18 'lli'd2 il.xgs 19 hxgs is very
similar to Line B21. As Markos points
out, the additional moves �bl and
... 'iii>h 8 help both sides to some degree,
because White will often open the b­
file with b4-bS (Black usually has to
play ... a6 to prevent a white piece from
hopping to bS), while ...'iii>h 8 could be
useful too because Black can play a
quick ... tll g 8 to open his queen's path
towards capturing the gs-pawn.
b) Black can also play 14 ... h6 lS tlle6
il.xe6 16 dxe6, but this should compare
unfavourably for him to the line 12 ... c6
13 il.e3 h6, because of the extra moves
�bl (which White plays anyway) and
...'iii>h8 (which Black may or may not
play). However, maybe it is not so bad
after 16 ...tll e 8 17 bS tll c 7 18 bxc6 bxc6
19 �b7 tllx e6 20 'lli' a4 (not 20 il.xa7??
�xa7 0-1, S.Kojima-G.Timoshenko,
Cappelle la Grande 2009) 20 ...tll d4 with
an unclear position in R.Markus­
Y.Dembo, Budapest 2002.
14 tlle6 il.xe6 15 dxe6 fxe4
Instead lS ... tllhs 16 exfs! opens the
140
position favourably for White after
both 16 ... tllxfs 17 il.d3 and 16 ... gxfs 17
g3, intending f4 when Black is weak on
the light-squares.
16 fxe4
Exchanging pieces helps Black. After
16 tllxe4 tllxe4 17 fxe4 tllc6 Black is fine
because White has no knight coming to
ds.
16...tllc6 17 tlld s
17 tllg S!
...
Black cannot play 17 ...tlld4? 18 e7
and 17...tllx e4? 18 'lli'd3 tllf6 19 'l!Vxg6
must also be avoided.
18 il.d3
Opening a path for the white queen.
Other moves do not lead to much:
a) 18 l:tfl �xfl+ 19 il.xfl tlld 4 20
'lli'g 4 c6 (with the idea of ... 'lli'f6) virtu­
ally forces White to sacrifice a piece
with 21 'l!Vxg6 cxds 22 exds, but this
looks insufficient after 22 ...'l!Vh4!.
b) 18 cs tlld4 19 il.c4 tllxe6 20 'lli'g4
'lli'e 8 21 cxd6 cxd6 (with the idea ... tlld4)
22 il.bs ! ? 'l!Vxbs 23 'l!Vxe6 'l!Vd3 24 il.e3
tllf6 2S �bdl 'lli'c 4! is also satisfactory
for Black - Markos.
 The M a r de/ Plata Va ria tio n : 9 b4 ll'i h s 1 0 .U. e 1
c ) 18 'lli'd3 ll'id4 19 'l!Vh3 .U.e8! 20
�g4!? (an interesting pawn sacrifice;
instead 20 e7 ll'ixe2+ 21 l:txe2 'lli'c8 22
il.e3 ll'ixe7 23 'l!Vxc8 .U.axc8 24 ll'ixe7
:.Xe7 2S il.xa7 is equal according to
Markos) 20 ... c6 21 ll'ic3 'lli'e7 22 bS ll'ixe6
(similar is 22 ... 'iii>h7 23 bxc6 bxc6 24 ll'ie2
!Lixe6 2S il.xe6 'l!Vxe6 26 'l!Vxe6 .U.xe6,
which has led to draws in some corre­
spondence games) 23 il.xe6 'l!Vxe6 24
'l!Vxe6 .U.xe6 2S bxc6 bxc6 26 il.e3 l:te7 27
.U.edl �d7 gives White compensation for
the pawn, although probably no advan­
tage. Then 28 a4 ll'if6 29 as �c8 30 a6
dS!? (30 ... �ce7 is solid, but obviously
rather passive) 31 exds cxds 32 ll'ixds
was A.Korobov-K.Maslak, Pardubice
2010, and now 32 ... ll'ig4! would keep the
game within the bounds of equality.
18 ... ll'id4 19 'lli'g4 gs
This move is forced. 19 ... 'iii>h 7?! 20
�e3 ll'if6 21 'l!Vh3 with the idea of e6-e7
is a problem, and 19 ... c6?! 20 'l!Vxg6! is
now very strong: for example, 20...'lli'e 8
(20 ...cxds 21 exds ll'if6 22 il.xh6 �g8 23
�gs is crushing) 21 'l!Vxe8 �axe8 22
!Lie3 �xe6 23 bS gives a good endgame.
20 h4
This move was Van Wely's choice in
his defeat of Radjabov. There are some
othertries, though:
a) 20 il.e3 c6 21 .U.fl! ? is speculative,
but White was . also successful after
21 ... cxds 22 exds ll'if6 23 'l!Vh3 e4? 24
il.xd4 exd3 2S 'lli'xd3 'lli'e7 26 cs in
S.Arun Prasad-J.Ashwin, Kavala 2009.
The calm 23 ...'iii> g 8 would have been
more testing, but the idea is interest­
ing.
b) 20 bS hopes to open the b-file in
the event of ... c6 by Black, but Markos
points out that Black can play 20 ... 'lli'c 8!
21 h4 'l!Vxe6 22 'lli'hs 'lli'f7 23 'lli'xf7 .U.xf7
24 hxgs hxgs 2 s il.xgs il.h6 with good
prospects.
c) 20 'l!Vh3 was Ponomariov's choice
when this position first occurred and
this move is considered best by Markos.
After 20 ... c6 White can try:
cl) 21 e7 ll'ixe7 22 il.xgs cxds 23
il.xh6 is a sharp piece sacrifice, but
Markos illustrates how Black can neu­
tralize it: 23 ... 'lli'c 8! (not 23 ... 'iii>g8 24
il.xg7 'iii>x g7 2s 'lli'g4+ ll'ig6 26 exds �f6
27 l:tfl with a strong initiative) 24 'lli'h4
dxc4! (again holding on to the piece is
dangerous after 24 ... 'lli'd7 2 S il.e3+ 'iii>g 8
26 il.xd4 exd4 27 exds) 2 s il.xg7+ 'iii>xg7
26 'l!Vxe7+ �f7 27 'l!Vgs+ 'iii>h8 28 'l!Vh6+
�h7 29 'lli'xd6 ifa 30 'lli'xe7 �xe7 and
Black has excellent compensation for
the pawn in the endgame.
c2) 21 ll'ie3 'lli'f6 and now:
c21) 22 ll'ifs ll'ixe6 23 �fl ll'if4 is al­
ready better for Black.
141
A ttacking C h e s s : Th e King 's I n d i a n, Vo l u m e 1
c22) 22 tllg4 'l!Ve7! 23 il.e3 (better is
23 tll e 3 'l!Vxe6 24 'l!Vxe6 tll xe6 2s tllfs
.U.ad8 26 il.e3 with some compensation
for the pawn) 23 ... tllxe6 24 g3 'lli'd7! ?
(perhaps even better is 24...tll d 4!? 2 S
�fl 'lli' e 6) 2S �fl was R.Ponomariov­
T.Radjabov, Wijk aan Zee 2003. Here
the simplest is 2s ... tll d4 with the idea
of 26 il.xg s? tll f 3+.
c23) 22 �fl 'l!Vxe6 23 �xf8 (Markos
suggests 23 tllf s tllx fs 24 exfs 'lli'f6 2s
il.b2 dS 26 bS with compensation for
the pawn) 23 ...�xf8 24 'l!Vxe6 tllxe6 2S
tllfs .U.d8 26 il.e3 il.f8 27 g3 tll g 7 28
tllx g7 'iii>xg7 29 bS cs 30 �b3 il.e7 31
.U.a3 .U.a8 32 'iii>f2 and White had enough
compensation for the pawn in
A.Klimov-A.Shomoev, Kemerovo 2007.
20...tllf6 21 'lli'g3
After 21 'l!Vh 3 g4 Black keeps the
kingside closed.
21...tllxe6!
This is much better than 21 ... gxh4?!
22 'lli'xh4 tllxe6 23 il.xh6 'iii>g 8? (this
loses, although after 23 ...tll h7 White
has a choice between an endgame
edge, with 24 'l!Vxd8 �axd8 2S il.e3, or a
142
continuing initiative after 24 il.xg7+
'iii>x g7 2S 'lli'g4+ tllh gS 26 l:tfl OT 26 tll e 3)
24 'l!Vh3 il.xh6 2S 'lli'xh6 c6 26 J:te3 ! with
a winning attack in L.Van Wely­
T.Radjabov, Wijk aan Zee 2009.
22 hxgs tllx ds
Worse is 22 ...tllxg S 23 il.xg S hxgs 24
'l!Vxgs when White has all the play.
23 gxh6
After 23 exds tllxg s 24 'lli'g4 'iii>g8
(24... �f6 is also possible) 2S il.e3 'lli'c8
26 'lli'h4 'lli'd8 27 'lli'g4 'lli'c8 28 'lli'h4 'lli'd8
29 'lli'g4 the game was drawn in
K.Landa-A.Shomoev, Ulan Ude 2009.
23 ... il.f6
This is fine, but 23 ...tlldf4 24 hxg7+
tll x g7 2S 'l!Vh2+ 'iii>g 8 26 hf4 �xf4 27 cs
.U.h4 28 'lli'g 3 'lli'f6 also looks okay for
Black.
24 exds il.h4
An important resource. The position
remains very complicated.
25 'lli' h 3
After 2S 'lli' g 6? il.f2+ 26 Wfl il.e3+ 27
il.fs l:txfs+! 28 'lli'xfs tll d 4 29 'lli'f7 'lli'f8
Black wins, while 2S 'lli'g4 il.xel 26 dxe6
(26 il.e3 'lli'h4 27 'lli'x h4 il.xh4 28 dxe6
 Th e M a r de/ Plata Va ria t i o n : 9 b4 ll'i h s 1 0 .U. e 1

:ae8 is much better for Black) 28 il.d2 l:tae8 29 .U.f1 -.xe6 30 -.xe6
26 ...il.f2+ leads to: �xe6 31 .il.e4
a) 27 �1 il.h4+ 28 'iii>e 2 (28 'iii> g l
..if2+ is a draw) 28.. ltf2+ 29 'iii> d l -.e7
intending ....U.g8 is unclear.
b) 27 'iii>h 2 .U.g8 28 -.e4 -.e7!
(28 ... �4+ 29 -.xh4 il.xh4 was given as
much better for Black by Velickovic and
Sasa in the Inform ant, but White has
two dangerous passed pawns and
looks to be doing well after 30 'iii>h 3 ! ) 29
i.d2 .U.af8 with counterplay.
25 ... il.xel 26 dxe6 il.f2+ 27 'iii>h 1
White looks to have sufficient com­
pensation for the exchange, especially
if the kingside pawns get moving, but
Black finds a clever resource to take
over the initiative.
31 ...1If4!?
Instead 3 1 .. ltef6 32 g4 il.g3 33 �xf6
.U.xf6 34 il.gs is not so easy for Black.
32 il.xf4 exf4 33 il.f5
Better was 33 J:txf2 l:txf2 34 g4!.
33 ...1Ixh6+ 34 iLh3 il.g3
21 ...-.f6I Black had a big advantage in Shen
This is more ambitious than the sim­ Yang-Ju Wenjun, Xinghua Jiangsu
plifying 27 ...°ii'h4. 2009.

143
 Chapter 7
The Mar del Plata Variation
White's other Ninth Moves
1 d4 tllf6 2 c4 g6 3 tllc3 il.g7 4 e4 d6 5
tllf3 o-o 6 il.e2 es 7 o-o tllc6 8 ds tlle7
In this chapter we look at moves
other than 9 tlle l, 9 tlld2 and 9 b4. The
first of these, 9 il.gs, seems illogical,
but it has had its bouts of popularity. 9
il.d2 was once a main line, but it seem s
too passive for modern tastes. Finally,
both 9 a4 and 9 'iii>h l look a bit indul­
gent, but they should not be underes­
timated. Although these various lines
are not considered to be theoretically
dangerous, they all have their ideas
and Black should still be well prepared.
144
A: 9 �gS
B: 9 .il.d2
C: 9 34
D: 9 �h1
There are a few other moves to con­
sider, although they are rather experi­
mental:
a) 9 'l!Vc2 looks a bit pointless, but it
cannot be too bad: 9 ... tllh s (9 ... tlld7 and
9 ... tlle8 are possible as well) 10 g3 (10
b4 is Epishin's line considered at the
beginning of Line B in Chapter 5) 10 ...fs
11 tllgs was P.Eljanov-M.Al Sayed,
Khanty-Mansiysk 2009. Here 11 ...tllf6
looks sensible and best.
b) 9 il.e3 rather invites 9 ...tllg4 (oth­
erwise White will play 10 tlld2), but it is
not so bad: 10 il.gs (after 10 il.d2 fS 11
tllg s tllf6 the position is the same as the
main line of the Bayonet, but White has
played il.d2 rather than b4 and .U.e1, and
after 12 exfs tllxfs! Black is fine) 10 .. .f6
(worse is 10 ... h6 11 il.h4 fS 12 tlld2 tllf6
 Th e M a r de/ Pla ta Va ria t i o n : White 's O t h e r N i n t h M o ves
13 f3 gs 14 il.f2 f4 lS cs with the initia­
tive) 11 il.h4 (11 il.d2 fS is the same as
10 il.d2) and now after 11 ... gs 12 il.g3
�h6 we reach a line of the Gligoric
where White has castled rather prema­
turely - see note 'c' to White's 12th
move in Line Bl of Chapter 9.
c) 9 �bl is undoubtedly playable,
but it should not be too scary for Black
theoretically.
When faced with a move like 9 �bl
Black must mainly rely on his under­
standing of the Mar del Plata in gen­
eral, as he did in E.Lobron-L.Brunner,
German League 1990: 9 ... tllh s (this is
very logical; instead 9 ... tlld7 10 b4 as 11
a3 'iii>h 8 12 il.e3 ! tllg 8 13 cs! axb4 14
axb4 fS lS tllg s favoured White in
l.Nikolaidis-A.Jankovic, Leros 2010) 10
tt:Je1 tllf4 11 tlld 3 tllx e2+ 12 'l!Vxe2 fS
and there was a more or less normal­
looking position similar to some found
in Line A, below.
d) 9 tllh4 looks quite extravagant,
but it is playable too: 9 ... tlle 8 (after
9 ... tlld7 10 g3 fs 11 exfs tllxfs 12 tllxfs
we can see some point to White's idea
following 12 .. ..i:!xfs 13 il.g4 or 12 ...gxfs
13 f4) 10 g3 (another point to Black's
9th move is that it covers the Cl-pawn,
so 10 il.e3 can be met by 10 ... tllx dS!)
and now, rather than 10...ilh3 11 �el
fs 12 exfs tllxfs 13 tllf3 ! ? h6 14 tlle4
when the h3-bishop was a bit mis­
placed in G.Hertneck-W.Watson, Ger­
man League 1994, I would prefer 10.. .fs
11 exfs tllxfs 12 tllf3 tllf6 13 il.d3 h6
with an equal position in C.Bauer­
F.Libiszewski, Montpellier 2004.
A) 9 iLgS
This move was brought to promi­
nence by the creative Ukrainian Grand­
master Vereslav Eingom and was
favoured by the Swiss Grandmaster
Yannick Pelletier for a while. It does not
seem logical for White to trade off his
dark-squared bishop, but if Black forces
the exchange then White hopes to gain
time for a queen side assault.
9 tll h s
...
Black has a decent alternative in the
popular 9 ... h6. After 10 il.xf6 il.xf6 11
b4 Black has:
145
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
a) 11 ...c6!? was recommended by
Bologan. We have seen this pre­
emptive strike before and after the 12
.U.cl (12 a4 as! 13 bxas cs is comfort­
able for Black) 12 ... as 13 a3 il.g7 14
'l!Vb3 cxds l S exds!? b6 16 �fdl axb4 17
axb4 il.d7 18 bS of Y.Pelletier­
V.Bologan, Cap d'Agde 2002, Bologan
suggests 18 ... g S ! ?.
b) 11 ...il.g7 has been the most
common choice: 12 cs (12 tlld2 fS
transposes) 12 .. .fs 13 tlld2 fxe4 (this is
usually played, but it scores badly; still,
13 .. .f4? is bad because of 14 il.g4 and
other moves like 13 ... 'iii>h 7 look silly) 14
tlldxe4 tllfs looks like a better version
of the 9 tlld 2 variation for White after
lS a4, 1 S �Cl OT 1 S il.g4.
c) 11 ...'iii> g 7 has always looked stur­
diest to me: 12 cs tllg 8 13 a4 (13 cxd6?!
cxd6 releases the tension prematurely
and 13 tlld 2 il.e7 14 c6 bxc6 lS dxc6 fS
gives Black good play) 13 ... il.e7 14 c6
(after 14 cxd6 both 14 ... il.xd6! ? and
14...cxd6 lS as fS 16 tlld2 tllf6 give
Black counterplay), and now Black can
play 14 ... bxc6 lS dxc6 fS, 14 .. .fS l S cxb7
146
il.xb7 or the untried 14... as!? with
counterplay.
10 tlle1
White attacks the hS-knight and
prepares to bring his own knight to d3.
Other moves are less dangerous:
a) 10 .U.el h6 11 il.d2 tllf4 12 il.xf4
exf4 gives Black a position similar to
those in the Bayonet with 10 cs, but
here Black is better placed. After 13
�d2 gs 14 h3 tllg 6 lS tllh 2 tlle s 16
.U.acl a6 Black had no problems in
J.Gonzalez Zamora-V.Bologan, Turin
Olympiad 2006.
b) 10 g3 h6! (this move will prove
useful because it controls gS; instead
10 .. .f6 11 il.d2 fS transposes to a line
which can arise from 9 il.d2 too, but
Black should avoid it as White has 12
tllg S!) 11 il.d2 (there is not much point
now in exchanging the bishop with 11
il.xe7?! 'l!Vxe7) and here:
bl) 11 ... il.h3 12 .U.el fS 13 tllh4 tllf6
14 exfs g s ! l S tllg 6 tllx g6 16 fxg6 il.fs
17 il.e3 'lli'e8 18 cs a6 19 .U.cl 'l!Vxg6 20
cxd6 cxd6 was fine for Black in l.Bem­
W.Watson, Gausdal 1991.
 Th e M a r de/ Plata Varia tio n : White 's O t h e r N i n t h M o v e s
b2) 11 .. .fs 12 tllh 4!? (after 12 exfs
Black should play 12 ... tllxfs, because 13
g4? fails to 13 ...tlld4 14 gxhs tll x e2+ lS
'l!Vxe2 il.g4) 12...tllf6 13 exfs g s 14 tll g2
(14 tll g 6 tllx g6 lS fxg6 il.fs is similar to
variation 'bl', but White's rook is not
on el) 14 ... tllxfs l S il.d3 tlld4 was level
in M.Krasenkow-W.Watson, Cappelle la
Grande 1990.
10 tllf4 11 tll d3 tllxe2+
...
Black secures the bishop-pair. Prac­
tice has show that White keeps some
edge after 11 ... tllx d3 with both 12
'l!Vxd3 and 12 il.xd3.
12 'l!Vxe2
12 h6
•••
Again this move will prove useful.
Instead 12 ...f6 13 il.d2 (less incisive is
13 il.e3 fS 14 f3 f4) 13 .. .fs 14 f4! com­
pares favourably (for White!) with the
note to Black's 13th move, below, be­
cause Black does not have control of gs.
13 il.d2
Instead 13 il.xe7 'l!Vxe7 is harmless,
while 13 il.e3 fS 14 f3 (after 14 f4 exf4
1s tllxf4 g s 16 tllh s il.es 17 exfs tllxfs
Black hits the e3-bishop - compare
with variation 'c' in the note to Black's
next move) 14... g S ! (14.. .f4 lS il.f2 g s
16 cs tll g6 i s also possible) lS c s tll g 6
16 .U.fcl �f7 17 .U.C2 (better is 17 exfs
il.xfs according to Pelletier) 17 ... tllf4 18
'lli'd2 tllxd3 19 'lli'xd3 g4 20 fxg4 f4 21
il.f2 il.xg4 22 cxd6 cxd6 23 .U.acl a6 24
tl:Ja4 'l!Vg s 2s tllb6 .U.af8 and Black had
an excellent attacking position along
with the bishop-pair in Y.Pelletier­
A.Fedorov, Plovdiv 2003.
13 gs
...
Black clamps down on the kingside.
The natural 13 ...fs leads to sharp play,
but is risky after 14 f4! (instead 14 f3 is
well met by 14.. .f4 or 14...c6!?) 14... exf4
1s tllxf4 gs.
Here White has:
a) 16 tlld3 f4 is fine for Black, as 17
es dxes 18 tllx es fails to 18....U.e8!.
b ) 16 tlle6 il.xe6 17 dxe6 f4 (intend­
ing ... tllg6) 18 g3 tll g6 (but not 18.. .fxg3
19 hxg3 tll g 6 20 tll d s! c6 21 'l!Vh s with
the initiative in A.Yermolinsky­
V.Topalov, Wijk aan Zee 1999) 19 gxf4
il.xc3! 20 bxc3 gxf4 21 'iii>h l 'lli'f6 gave
Black promising counterplay in A.Blees-
147
A ttacking Chess: Th e Kin g 's I n d i a n, Vo l u m e 1
H.Klarenbeek, Heraklio 1993.
c) 16 tllh s ! il.d4+ (worse is 16...il.es
17 exfs tllxfs when, instead of 18 g4
tllg 7 19 tllx g7 'iii>x g7 20 �xf8 'lli'xf8, as
seen in D.Berczes-L.Valdes, Budapest
2006, White can play 18 tlle 4! il.xb2 19
il.c3 il.xc3 20 tllx c3 with a strong initia­
tive for the pawn) 17 'iii>h l f4 sees Black
threatening to consolidate with ...tllg 6.
Thus White must act quickly with
18 g3! and here:
cl) 18 ... il.h 3 19 gxf4! looks promis­
ing: 19 ...il.xfl 20 �xfl tllg 6 21 es! (after
21 'lli'd3 Black should avoid 21 ...il.h8 22
es 'lli'e8 23 e6 as in M.Galyas-S.Husari,
Balatonlelle 2003, and play 21...il.xc3!
22 il.xc3 tllxf4 23 tllxf4 �xf4 24 �xf4
gxf4 2s 'lli'd4 'l!Vgs 26 'lli'h8+ 'iii>f7 when
White does not have more than a per­
petual check with 27 'lli'h 7+ because 27
'l!Vxa8? gets mated after 27.. .f3)
21...dxes 22 'lli' d 3! tllxf4 (White wins
after 22 ...'iii>h 7 23 fxg s hxgs 24 �fS!
!Ixfs 2s 'lli'xfs with a decisive attack) 23
l:.xf4 (not 23 il.xf4 'lli'e 8!, turning the
tables) 23 ...exf4 24 'lli'xd4 'l!Ve7 2S tlle4
and White has a huge advantage.
148
c2) 1 9. ..fxg3 1 9 �xf8+ 'lli'xf8 20 tllb s!
(after 20 �fl iLh3 ! 21 �xf8+ !Ixf8 Black
will win back the queen with ...�f2, as
in the game G.Prakash-P.Konguvel,
Nagpur 1999) 20 ...'lli'f2 21 'lli'xf2 il.xf2
22 tllx c7! (worse is 22 hxg3 il.b6 23 il.c3
a6 24 tlld4 'iii>f7 2S 'iii> g 2 'iii> g 6 26 �hl
il.d7 when Black was at least equal in
G.Prakash-P.Konguvel, Chennai 2000)
22 ... il.h3 23 hxg3 �c8 was P. Lukacs­
V.Rajlich, Budapest 2001, and now 24
tllb s �xc4 2S 'iii>h 2! would give White a
comfortable plus.
14 g4
White blocks the kingside. Instead
14 f3 fS gives Black counterplay, while
14 h4?! g4 lS f4 gxf3 16 'lli'xf3 fS gave
him the initiative in L.Oll-A.Shirov, Til­
burg 1992.
14...tll g6
The knight heads for f4.
15 f3 tllf4 16 tllxf4
White could also try 16 il.xf4!? exf4
17 .U.acl with the idea of opening the c­
file, but Black's strong bishops should
certainly hold the balance.
16...exf4
 The M a r de/ Plata Va ria t i o n : White 's O t h e r N i n th M o ves
Black is already quite comfortable
with the bishop-pair and th e possibility
of opening the h-file with ... hs.
17 �fdl
More solid is 17 tlldl with the idea
of exch anging Black's dark-squared
bishop immediately. However, after
17 ...c6 18 il.c3 cxds 19 il.xg7 'iii> x g7 20
c:xds il.d7 21 tllf2 �c8 22 'Ii' d2 'lli'b6
Black was still very comfortable in L.Oll­
S.Dolmatov, Rostov 1993.
11 hs! 18 h3 il.es 19 il.e1
...
Safer was 19 'iii>g2 with the idea of
defending with �hl at some point.
19. 'iii>g7
..
Strengthening the long diagonal
with 19 ...'lli'f6 was also possible.
20 J:td3?! �h8 21 tlld 1 hxg4 22 fxg4
After 22 hxg4 both 22 ...'lli'f6 and
22 ... �h3 with the idea of ... 'lli'h 8 are very
good for Black.
22 iff6 23 il.c3 il.d7 24 tllf2 �ae8
...
Black had a clear advantage in
V.Pelletier-T.Radjabov, Biel 2006.
B) 9 il.d2
This old move is rarely seen nowa-
days, but it is not completely harmless.
White prepares to bring his rook to cl
and remains very flexible.
9 ...tlle8
Black can also play 9 ...tlld7 when 10
�cl fS 11 tll g s tllf6 leads us back to the
main line. However, White has an extra
option in 10 b4 fS 11 tllgs tllf6 12 f3, so
9 ... tll e8 seems more accurate.
Instead 9 ... tllh s was for a long time
considered the main line, but after 10
g3 fS 11 tllg s ! (more common is 11 exfs
tllxfs 12 tlle4, but Black is fine after
12 ...tllf6) 11 ...tllf6 12 f3 White has
scored well. This position is similar to
the Bayonet with 12 f3, but here White
has played il.d2 instead of b4. This may
not favour him, but it could well intro­
duce a new world of complications that
is not really worth delving into consid­
ering how rare 9 il.d2 is nowadays.
1o �c1 fs
After 10 tll el fS 11 tlld3 tllf6 we
have transposed back to Chapter 1. In­
stead 10 b4 fS 11 'lli'b 3 (11 tll g s makes
less sense here because the e6-square
is defended) 11 ...tllf6 12 exfs should be
149
A ttacking Chess: Th e King 's I n dian, Vo l u m e 1
met by 12 ... tllx fs! when 13 tllg s? runs
into 13 ...tlld4.
11 tll gs
Others:
a) 11 exfs gxfs (11 ...tllxfs! ?) 12 tll g s
h6 13 tlle6 il.xe6 14 dxe6 'lli'c8 lS 'l!Vb3
c6 16 il.hs 'l!Vxe6 17 'l!Vxb7 tllf6 18 il.e2
�fb8 19 'lli'a6 .U.xb2 gave Black good
play in M.Taimanov-R.Fischer, 1st
matchgame, Vancouver 1971.
b) 11 'l!Vb3 b6!? 12 exfs gxfs 13 tll g s
tl\f6 (after 13 ...h6 14 tll e6 Si.xe6 lS dxe6
'lli'c8 16 tllds 'l!Vxe6 17 tllxe7+ 'l!Vxe7 18
cs+ 'iii>h8 19 cxd6 tllxd6 20 .U.c6 White
had decent compensation for the pawn
in M.Taimanov-Ma.Tseitlin, USSR 1973)
14 f4 h6 1s fxes dxes 16 cs tllfxds 17
tllxds tll x ds 18 cxb6 axb6 19 .U.c6 'iii>h8
20 tl\f3? (better is 20 'lli' h3 when Black
can play either 20.. J:tf6! ? or 20 ... tllf6 21
il.c3 f4! ? 22 'lli'h4 il.b7) 20 ... il.b7 21 �g6
tl\f4 22 Si.xf4 exf4 23 .U.dl 'lli'e7 gave
Black a big advantage in M.Taimanov­
R.Fischer, 3rd matchgame, Vancouver
1971.
11...tllf6
This looks better than 11 ... h6 12
150
tlle6 Si.xe6 13 dxe6 'Ii'c8 14 'l!Vb3 c6, as
in L.Ftacnik-Kr.Georgiev, Groningen
1976. Here lS cs!? would give White
the initiative.
After the text, again we have a posi­
tion resembling the Bayonet. Here
White has played the rather slow il.d2
and l:tcl, though, instead of b4 and
l:tel.
12 f3
12 exfs should be met by 12 ... tllxfs!,
heading for the d4-square.
12 ...c6 13 'l!Vb3 h6 14 tll e6 il.xe6 15
dxe6 'lli'c8 16 .il.e3 '1li'xe6
This is fine, but Black has decent al­
ternatives in 16 ... .U.d8 and 16 .. .f4.
17 l:tfdl
Instead 17 'l!Vxb7 is met by 17 ... .U.fb8
18 'lli'a6 �xb2 with an unclear position.
11 ... 'iii>h7 18 'l!Vxb7 .U.fb8 19 'lli'a6
19 .. .fxe4!
An important move. Here 19 .. ..i:!xb2
is bad because of 20 'l!Va3! which also
hits the d6-pawn. Likewise, 19 ... hS?! 20
cs ds 21 exds tllfxds 22 tllx ds tllxds 23
il.c4 gave White the upper hand in
J.Plachetka-V.Babula, Stare Mesto 1992.
 Th e M a r de/ Pla ta Varia tio n : Wh i t e's O t h e r N i n t h M o ves
2o fxe4
If 20 tllxe4 tllxe4 21 fxe4 then
21 ... �xb2 hits the bishop on e2.
20...�xb2 21 'l!Va3 �xe2 22 tllxe2
Not 22 �xd6 .U.xg2+! 23 'iii>xg2 'lli'g4+
24 'iii>f2 �f8 when Black has a strong
attack.
22 ...tllxe4
With two pawn for the exchange
and squares for his pieces, Black has a
good position.
C) 9 a4
This move is rare but is hardly bad.
Moreover, it has certain similarities
with and may even transpose to 9 tll e l
tt:ld7 10 il.e3 fs 11 f3 f4 12 il.f2 gs 13 a4
(Line E of Chapter 3).
9 .as
..
This is a sound reply. Black can also
play 9 ...tllh s as he would against 9 b4.
White chooses between 10 as and 10
g3 with rather unexplored play.
10 tlle1
Instead 10 b3 looks slow, but White
can still play this way: 10 ...tlld 7
(10...tllh s ! ?) 11 il.a3 il.h6 (worse is
11 ... fs 12 tll g s tllf6 13 cs!, but 11 ...b6
and 11 ...tllcs are sensible alternatives)
12 b4 axb4 13 il.xb4 fS was M.Carlsen­
A.Morozevich, Biel 2006, and Black was
a tempo up on a line of the Bayonet (9
b4 as 10 il.a3 axb4 11 il.xb4 b6).
10...tll d 7 11 tll d 3
This is more dangerous than 11 il.e3
fS 12 f3 when Black can keep the ten­
sion with 12 ...tllc s!? (instead 12 .. .f4 13
il.f2 gs leads to the main lines of Chap­
ter 3) 13 tlld3 b6 and now:
a) 14 b4 tllxd3 lS 'lli'xd3 axb4 16
tllb s 'iii>h8 17 'l!Vb3 (after 17 il.d2 Black
can consider 17 .. .fxe4 18 fxe4 .U.xfl+ 19
il.xfl tl:Jg8 20 il.xb4 tl:Jf6 21 aS il.h6 OT
17... cs!? 18 dxc6 tll x c6) 17 ... tll g 8 18
'l!Vxb4 and here 18 ... tllf6 19 exfs gxfs 20
il.gs h6 21 il.h4 il.d7 22 .U.a3 'l!Vb8 23
il.f2 .U.g8 24 �bl tllh s 2s 'iii>h l 'l!Vd8 gave
Black counterplay in V.Korchnoi­
G.Kasparov, Barcelona 1989, but later
Kasparov preferred 18 .. .fxe4! 19 fxe4
�xfl+ 20 �xfl il.h6 21 il.xh6 tllxh6 22
'l!Vd2 'iii> g 7 when Black is certainly bet­
ter.
b) 14 �a3 ! ? 'iii>h 8 lS 'lli'd2 f4 16 il.f2
h S (after 16 ... g s Black must consider 17
g4! ?) 17 tllb s gs 18 b4 axb4 19 'l!Vxb4 g4
20 �fal tll g 6 21 'lli'el (or 21 as �xas 22
�xas bxas 23 �xas? gxf3 24 gxf3 tllx d3
2S il.xd3 iLh3 when White cannot de­
fend against 26 ... 'l!Vgs+), and now
21 ...tllb7! ? was unusual but held up
White long enough on the queenside
for Black to develop good kingside play
in A.Evdokimov-D.Jakovenko, Russian
Team Championship 2010.
151
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
11...fs 12 f3
After 12 exfs tllxfs 13 tlle4 tllf6
Black is fine.
12 ...f4
This is rather committal, but it is
thematic and logical. Instead 12 ...b6 13
il.d2 tllf6 14 b4 axb4 lS tllxb4 il.d7 16
tlld3 cs!? 17 �a3 �f7 was unclear in
A.Evdokimov-L.Nisipeanu, European
Championship, Budva 2009.
Another possibility is 12 ...'iii> h8 13
il.e3 b6 14 b4 (if 14 'l!Vb1 il.a6! ? then lS
b3 tllg 8 is equal according to Bologan
and lS tll b s can be met by the reply
1 s ... tllf6, pressuring e4) 14 ... axb4 l S
tllxb4 tllf6 16 tlld3 tlleg8!? 17 tllf2 tllh s
18 �el il.h6 19 il.xh6 tllx h6 20 iLfl f4
21 tllb s g s gave Black good counterplay
in S.Khmelevsky-G.Beckhuis, Vienna
2006.
13 b3!
This is the point of White's play - he
will be able to open the queenside. This
may look slow, but it is the right plan
and if we compare it to Line E of Chap­
ter 3, White is actually getting his
bishop to b4 more quickly by playing
152
b3, il.a3, b4 and il.xb4, rather than
il.e3-f2-e1, b4 and il.xb4.
13 ...gs 14 il.a3 b6
Instead 14...tllf 6?! allows the imme­
diate lS cs, as in S.Atalik-V.Bologan,
Manila Olympiad 1992.
15 b4 axb4 16 il.xb4 tllc s 17 as
This is a critical position. Black is too
far behind to just play for mate:
17 ... il.d7?! 18 tllb s gave White the ini­
tiative in V.Popov-F.Amonatov, Sochi
2007, and 17 ... il.a6 18 axb6 (better
than 18 tllbs c6!) 18 ... cxb6 19 tllx cs
bxcs 20 il.as 'lli'd7 21 'lli'a4 looks better
for White as well.
Black's best looks like 17 ...c6!? 18
axb6 (instead 18 tll a4 tllxd3 19 'lli'xd3
bxas 20 il.c3 cxds 21 cxds il.a6!? 22
'lli'd2 il.xe2 23 'l!Vxe2 'lli'e7 24 �fbl �fb8
was fine for Black in A.Klimov­
E.Bondarenko, Tomsk 2008) 18 ... tll x d3!
(but not 18 ....U.xal? 19 'l!Vxal tllxd3 20
il.as!) 19 �xa8 (19 il.xd3? loses to
19 ...�xal 20 'l!Vxal 'l!Vxb6+) 19 ... 'l!Vxb6+
20 cs! tllxcs 21 il.as 'lli'b7 22 �xc8 .U.xc8
23 dxc6 tllxc6, which looks okay for
him.
 Th e M a r de/ Pla ta Va ria t i o n : Wh ite 's O t h e r N i n t h M o ves

D ) 9 �h1

A few possibilities:
Tony Miles's old favourite looks bl) 13 a3 ! ? looks promising. White
pointless, but it is not completely with­ will play b4 next unless Black sacrifices
out merit. a pawn with 13 ...a4 14 il.xcs dxcs lS
9 tlle 8!
... tllxa4.
One point of White's play is that b2) 13 l:tcl is a bit slow as the rook
9 ... tllh s can be met by 10 tllg 1 tllf4 11 does not have much purpose on this
l.f3, although i t is not clear that this is square:
so dangerous. Instead 9...�h8 can be
met by 10 a4 or 10 tll e l, when it is
hardly clear which player the king
moves will favour.
As I found out for myself, 9 ... tlld7 is
less accurate than 9 ...tlle 8, though, be­
cause it allows 10 g4!?. With this move
White discourages Black from opening
the kingside, after which he may tum
his attention back to the queen side.
a) 10...fs 11 gxfs gxfs 12 exfs
(White can use both the e4-square and 13 ... il.d7 14 b3 tllg 8 lS tlld2 f6
the g-file) 12 ...tllf6 13 �gl tllxfs 14 il.d3 (Black intends ... il.h6) 16 h4 fS! (now
�h8 lS tllg s left White with an edge in that White has weakened his kingside,
R.Kaufman-D.Vigorito, Chicago 2008. Black tries to open things up) 17 gs
b) 10...�h8 11 �gl! as 12 il.e3 tll cs tllxe4 18 tlldxe4 fxe4 19 tllxe4 tl\e7 20
gives White an active version of some il.g4 tllfs gave Black good counterplay
similar plans we have seen with the g4- in M.Carlsen-T.Radjabov, Biel 2006.
advance. b3) 13 'lli'd2 (with the idea of �g3

153
A ttacking Chess: Th e King 's I n d i a n, Vo l u m e 1
and .U.agl) 13 ...tll g 8 14 l:tg3 b6 lS l:tag l
tllf6 16 'l!Vc2 il.d7 17 tlld2 c6 18 f3 gave
White an edge in M.Tosic-A.Kolev,
Vrnjacka Banja 1990.
b4) 13 tlle l fS 14 f3 fxe4 lS fxe4 c6
16 tlld3 cxds 17 cxds b6 18 tllxcs bxcs
19 �d2 �b8 20 !Iafl and again White
could claim some advantage in A.Miles­
F.Gheorghiu, Cran s Montana 2001.
10 tlle1
Instead White can also play 10 a4 as
11 tlle l fS 12 tll d3 tllf6 13 f3 b6, which
is similar to Line C, although here
White has played 'iii>h l rather prema­
turely.
It looks a bit late for 10 l:tgl because
Black can play 10...fs 11 exfs when
11 ...gxfs 12 tll g s tllf6 has been played
with some success, but I prefer
11...tllxfS ! when White's kingside shuf­
fling looks rather silly.
10 ... fs 11 exfs tllxfs!
Again I prefer activating the knight
rather than allow 11 ...gxfs 12 f4!?.
12 tllf3
Covering the d4-square, but this is
obviously time-consuming. White's
1 54
king move looks out of place here.
White can also play 12 il.d3 tllf6 13
tllf3 which transposes to the main line.
Instead 12 tllc2 tllf6 13 g4? just weak­
ens the kingside too much. After
13 ...tll d4 14 h3 tllx g4! ? lS hxg4 'lli'h4+
16 'iii>g l tllx e2+ 17 'l!Vxe2 il.xg4 18 f3 a
draw was agreed in A.Miles­
D.Gormally, British Championship,
Hove 1997, but Black should really have
played on with 18 ... il.xf3! 19 .U.xf3
'lli'g4+ 20 'l!Vg2 'lli'xf3 21 'lli'xf3 .U.xf3 22
'iii>g 2 l:taf8 when he is certainly better.
Here a clearer refutation is available
even sooner with 14 ... tllx c2! lS 'l!Vxc2
il.xg4 16 il.xg4 (White gets mated after
16 hxg4 tllxg4 17 il.xg4 'lli'h4+ 18 'iii>g2
'l!Vxg4+ 19 'iii>h2 .U.f3) 16 ... tllx g4 with a
winning position.
12 ...tllf6 13 il.d3 tlld4 14 tllxd4 exd4 15
tlle4 tll xe4 16 il.xe4
16...'lli' h 41
This is more aggressive than
16... il.d7 17 'l!Vd3 cs 18 dxc6 bxc6 19
il.d2 .U.b8, which was also satisfactory
for Black in A.Miles-A.Beliavsky, Nova
Gorica 1999.
 Th e M a r de/ Plata Varia tio n : White 's O t h e r N i n t h M o ves
17 'l!Vc2 il.fs!
Black develops very quickly.
18 il.xfs �xfs 19 'iii>g1
This is a bit humbling. White 'takes
back' his 9th move. However, he has
trouble developing his queenside and a
good alternative is hard to find, be­
cause 19 il.d2 loses to 19 ...�xf2.
19 �es
••.
Black has a fantastic position. All of
his major pieces are active, he has a
passed d-pawn and White has no ac­
tive possibilities.
20 a4 il.es
Black could also play 20...c5!?.
21 h3 d3!?
This forces the pace. Again 21...c5
was possible too.
22 'l!Vxd3 .U.xf2 23 �xf2 il.d4 24 g3
This is forced, because 24 'lli'fl loses
to 24 ... il.xf2+ 25 'lli'xf2 .U.el+.
Now after 24 ... 'lli'xh3 ? White man­
aged to hold on with 25 il.f4 il.xf2+ 26
'iii>xf2 'lli'h2+ 27 'iii>t3 'l!Vh5+ 28 'iii>g 2 .U.e2+
29 'iii>fl .U.xb2 30 'lli'e4! (this covers the
hl-square and prepares to check the
black king) 30 ... 'lli'h 3+ 31 'iii> el �g2 32
'lli'e8+ 'iii>g 7 3 3 'lli'e7+ 'iii> g 8 34 'lli'e8+ and
V2-V2 in A.Miles-H.Tirard, Cappelle la
Grande 2000.
However, Black could have crowned
his excellent play with 24 ...'lli'f6! 25 'lli'fl
(or 25 il.f4 il.xf2+ 26 'iii>xf2 'l!Vxb2+)
25 ... 'lli'f3 !, paralyzing White and threat­
ening ...�e2: for example, 26 il.d2 �e2
27 il.el 'l!Vxg3+ 28 'l!Vg2 (28 'iii>h l il.xf2
29 'l!Vxe2 'l!Vg1 mate) 28 ... .U.xel+ 29 J:txel
il.xf2+.
155
 Chapter 8
The Mar del Plata Va riation
White's Eighth Move Deviations
1 d4 tl:Jf6 2 C4 g6 3 tl:Jc3 J.g7 4 e4 d6 5
tl:Jf3 0-0 6 J.e2 es 7 0-0 tllc6
Usually when White plays 7 o-o he
is willing to enter into a sharp fight.
However, that is not always the case
and in this chapter we look at moves
other than 8 dS. It seems that with the
moves 8 dxes and 8 J.e3 White is
changing his mind about what kind of
game he wants and these variations
are certainly less popular than 7 dxes
and 7 J.e3. While the lines considered
here are not terribly dangerous, they
should not be ignored, especially be-
156
cause it i s not easy to create counter­
play unless one is familiar with certain
ideas.
A: 8 dxes
B: 8 J.e3 .
Note th at 8 J.g 5 is not very good
because of 8... tllxd4 9 tllxd4 exd4 10
'lli'xd4 tllxe4! 11 J.xd8 J.xd4 12 J.xc7
J.xc3 ! (this is better than 12 ... tllxc3 13
bxc3 J.xc3 14 �abl which is just equal)
13 bxc3 J.e6 14 J.f3 l:tac8! when Black
is much better.
A) S dxes
White plays a kind of delayed ex­
change variation. This should not trou­
ble Black, but it is not as innocent as
one might think and Black should not
go to sleep just yet.
8 dxes 9 J.gs
...
Instead 9 'lli'xd8 tllxd8 is solid, but
Black can also just play 9...�xd8 10 J.gs
l:tf8! (White may keep a pull after
 Th e M a r de/ Plata Varia tio n : W h i t e's Eig h t h M o ve D e viatio n s
1 1...l:te8 1 2 tlld s tllx ds 1 3 cxds tlld4 14
..'!Jxd4 exd4 lS il.d3 or 11 ...�d7 12
.i..d l!) 11 �fdl il.g4 which actually
transposes back into the main line.
9 'lli'xdl
...
Black has an interesting alternative
in 9 ... il.g4!? which immediately fights
for the d4-square. Then 10 'lli' x d8 and:
a) 10...�fxd8 11 tlld s tll x ds 12 exds
f6 13 dxc6 fxgs 14 cxb7 l:tab8 lS h3
.i..fs was l.Rausis-1.Nataf, Enghien les
Bains 1997, and here 16 cs! would give
White some initiative.
b) 10 ... �axd8 11 il.xf6 (better is 11
..'!::l d s tllxe4 12 il.xd8 �xd8 13 �fdl
when Black probably lacks sufficient
compensation for the exchange)
11 ... il.xf6 12 tlld s il.xf3 13 il.xf3 il.g s
was fine for Black in R.Ponomariov­
V.Topalov, Vitoria Gasteiz 2007; 14
..'!::lxe7 l:td2 gives Black enough activity
for the pawn.
c) 10...tllxd8 11 il.xf6 il.xf6 12 tlld s
.i..xf3 13 gxf3 ! (after 13 tllxf6+?! 'iii> g 7 14
.i..xf3 'iii>xf6, with the idea of ...tllc 6-d4,
Black is already better, while 13 il.xf3
�g s 14 tllx c7 l:tc8 is certainly not worse
for the second player) 13 ... il.g s 14 tllxe7
.:.c8 lS tll d s tlle6 gives Black some
compensation, but it is a bit specula­
tive - White's extra pawn is 'real' on
the queenside.
10 .U.fxd1
White can also play 10 �axdl, but
then the fl-rook would be a bit stuck.
10 il.g4
...
Instead 10 ... h6 11 il.e3 il.g4 looks
fine too.
11 h3
Black is certainly not troubled by 11
�acl h6 (11 ... tlld4!?) 12 il.e3 .U.fd8, but
11 �d3 is interesting. Black could try
11 ... hf3 12 il.xf3 tlld4 13 tlld s tllxds
14 cxds fS or even the immediate
11 ... tllxe4!? 12 tllxe4 fS, but practice
has focused on 11 ... h6 12 il.e3 tllxe4!
(White is a little better after 12 ...�fd8
13 l:tadl �xd3 14 �xd3) 13 tllxe4 fS 14
tll c s e4 lS �b3 and now:
a) 1S ...f4 16 il.d4 (Black is certainly
okay after 16 tllxe4 fxe3 17 l:txe3 il.xf3
18 il.xf3 il.xb2 19 l:tbl il.d4) 16 ... tllxd4
17 tllxd4 il.xd4 18 il.xg4 il.xcs 19 �xb7
il.d6! (but not 19 ... il.b6 20 b4!) is equal
but unbalanced.
b) 1S ... exf3 16 gxf3 f4 17 fxg4 fxe3
18 fxe3 tlld4!? 19 exd4 il.xd4+ 20 'iii>hl
il.xcs 21 �fl?! �xfl+ 22 il.xfl b6 was
drawish but still a little bit better for
Black in B.ltkis-M.Golubev, Sovata 2000.
However, 21 �xb7 il.b6 22 cs! is a bet­
ter try, so I would prefer to deviate ear­
lier with 18 ... tlla s 19 �bs b6 or just
18 ... �ae8. In both cases Black has good
compensation for a pawn.
157
A tt a c k i ng Chess: Th e King 's Indian, Vo l u m e 1

11...il.xf3 12 il.xf3 tlld4 13 tlld s tllx ds the lines of the Gligoric Variation with
14 cxds the sharper 8 ... tllg4.

Even though the d4-knight looks

nice, White has the bishop-pair and 81: 8...�e8
potential pressure on the c-file, so 82: 8 tllg4
•..

Black must act quickly.

14 ...fs 81) 8 ...l:te8
Worse is 14...c6 1 5 dxc6 bxc6 16
.U.acl �fb8 17 b3 as 18 l:tc4 when White
has some pressure.
15 .U.acl .U.f7 16 il.e3
After 16 'iii>fl Black should avoid
16 ...f4?! 17 il.g4! and play 16 ... il.f8 or
16 ...l:taf8.
16 ...fxe4 17 il.g4
Wisely avoiding 17 il.xe4? tlle 2+ and
17 il.xd4?! exf3 18 il.e3 fxg 2. After 17
il.g4 tllfs 18 l:tc3 hs 19 il.e2 il.f8 20 l:tc4
tlld6 21 l:tc3 tllfs 22 l:tc4 tlld6 23 l:tc3 This move has long been known to
tllfs the game was drawn in l. Krush­ equalize.
A.Melekhina, US Women's Champion­ 9 dxes
ship, Saint Louis 2009. The point of Black's rook move is
seen after 9 dS tlld 4! 10 tllxd4 exd4 11
8) 8 .il.e3 il.xd4 tllxe4 12 il.xg 7 'iii>x g7 13 tll xe4 (or
White tries to keep the tension in 13 'lli'd4+ 'lli'f6) 13 ...�xe4 when the posi­
the centre. Black has the traditional tion is equal, although Black does have
equalizer 8 ... �e8 or he can play along slightly the better long-term chances

158
 The M a r de/ Plata Varia tio n : W h i te's Eig h t h M o ve D e viations
because of his better bishop.
9 dxes
...
1o 'lli'xd8
This is not very ambitious, of course,
but other moves do not give White
much either:
a) 10 cs it.g4! (fighting for the d4-
square) 11 .tbs 'l!Vc8 (Smirin's idea;
11 ... tlld7 12 it.xc6 bxc6 13 h3 it.xf3 14
·tvxf3 'l!Vb8 is another option) and now:
al) 12 h3 �d8 (it is probably better
to just play 12 ...it.hS! 13 it.xc6 bxc6,
transposing to variation 'a2') 13 tlld s
!Dxds 14 exds it.e6 (instead 14...it.xf3
lS 'lli'xf3 tlld4 16 it.xd4 exd4 17 �fel
·tvfs 18 'lli'xfs gxfs 19 it.c4 it.f8 20 b4 as
21 a3 d3 is possible, but Black is really
just playing to draw here) lS dxe6!
:xdl 16 exf7+ 'iii>xf7 17 it.c4+ �8 18
:taxdl gave White good compensation
for the queen in P.Lukacs-A.Stummer,
Budapest 1992.
a2) 12 it.xc6 bxc6 13 h3 it.hs 14 it.gs
(otherwise, there is 14 g4? tll x g4 and 14
'l!Ve2 h6 1s 'iii>h2 gs 16 'lli'c4 �b8 17 b3
as 18 tlld2 g4 19 f3 gxh3 20 gxh3 .U.b4
was unclear in P.Van der Sterren-
l.Smirin, Tilburg 1992) 14 ... .U.b8 lS .U.bl
it.xf3 16 'lli'xf3 tlld7 (with the idea of
...tlld 7-e6) 17 b4 as 18 bS (18 a3 axb4
19 axb4 tllf8) 18 ...tllx cs 19 bxc6 tlle6
gave Black counterplay in L.Oll-1.Smirin,
Rostov on Don 1993.
b) 10 h3 prevents ...it.g4, but costs
White time. After 10 ... it.e6 11 cs Black
has:
bl) 11 ...tllh s 12 tllg s (12 .tbs ! ?)
12 ...tllf4 13 tllxe6 tllxe6 14 .tbs .U.f8!?
(14...'lli'xdl lS �fxdl .U.ed8 16 it.xc6
bxc6 is level) lS it.xc6 bxc6 16 'lli'a4 'lli'e 8
17 .U.adl fS 18 exfs (after 18 f3 f4 19
it.f2 gs Black has counterplay on the
kingside as well) 18 ...gxfs 19 f3 'iii>h 8 20
'iii>h l .U.g8 21 .U.d2 it.f6 and Black had an
active position in P.Van der Sterren­
B.Gelfand, Biel lnterzonal 1993.
b2) 11...a6 kind of mimics White's
10th move by preventing any .tbs
ideas: 12 'lli'a4 'lli'e7 13 .U.fdl .U.ad8 14
it.c4 tlld4! lS tllxd4 (Black has no prob­
lems after lS it.xe6 tllxf3+ 16 gxf3
'l!Vxe6 17 'iii>g 2 c6) 1s ... exd4 16 �xd4 was
1.Jelen-M.Tratar, Slovenian Team
Championship 1993. Here Black should
159
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
play 16 ... ll'ld7! 17 ll'lds ii.xds 18 .l:f.xds
c6 19 l:tdd1 ll'lxcs 20 'ti'b4 ii.f8 with a
good position.
b3) 11 ... 'ti'e7!? 12 ii.bs (this seems
logical as Black has omitted ... a6, but
Black still obtains satisfactory play; in­
stead 12 'ti'c2 .l:f.ad8 13 .l:f.ad1 was
J.Speelman-J.Nunn, Hastings 1987/88,
and now Black should play 13 ... a6 ac­
cording to Speelman) 12 ....l:f.ed8 13 'ti'a4
ll'ld4! 14 ll'lxd4 (14 ll'lxes a6! 1s ii.xd4
axbs 16 't\Vb4 ll'lxe4! with the idea of 17
ll'lxe4 .l:f.a4! is good for Black) 14... exd4
1S ii.xd4 c6 16 ii.e2 ll'lxe4! 17 ii.xg7
ll'lxcs 18 't\Va3 ..t>xg7 (Black has won a
pawn, but he will have to untangle his
pieces) 19 ll'le4 b6 20 .l:f.fc1 ii.xh 3 ! ?
(tempting, but 20....l:f.dS is simpler and
preparing for the opening of the b-file
with 20....l:f.ab8!? is also possible) 21
'ti'c3+ f6 22 ll'lxcs ii.xg2! 23 .l:f.e1! ii.ds
and Black had good compensation for
the piece in D.Berczes-A.Jankovic, Sara­
jevo 2010.
10...ll'lxdS
Instead 10 ... .l:f.xd8 11 .l:f.fd1 (after 11
ii.gs .l:f.f8! 12 .l:f.fd1 ii.g4 we're back in
160
Line A) 11 ... ii.g4 12 ll'lds ll'lxe4 13 ll'lxe7
.l:f.xd1+ 14 l:txd1 .l:f.c8 1s ll'lds ll'lf6 16
ll'lxf6+ ii.xf6 17 h3 ii.e6 18 b3 was
M.Dlugy-A.Fishbein, New York 1991,
and here 18 ...l:td8 19 .l:f.xd8+ ii.xd8 looks
fairly level.
11 ll'lbs
After 11 ll'lds ll'le6 12 ll'lgs ll'lf4! ?
(12 ... ll'lxds 13 cxds ll'ld4 is also okay,
but not 13 ... ll'lxgs?! 14 ii.xgs with an
edge for White) 13 ii.xf4?! ll'lxds 14
ii.d2 ll'lf6 1s f3 ll'lhs 16 .l:f.fd1 ll'lf4 11
ii.fl ii.f6 18 ll'lh3 ii.xh3 19 gxh3 ll'le6
Black was winning in Z.Mijailovic­
V.Kotronias, Vmjacka Banja 2006.
11 ... ll'le6 12 ll'lgs l:te7 13 .l:f.fd1
Alternatively:
a) 13 ll'lxe6 ii.xe6 14 f3 c6 (14 ...b6!?)
1S ll'lc3 (1S ll'lxa7 looks risky but could
be more testing) 1s ... .l:f.d7 16 .l:f.fd1 ii.f8
was fine for Black in S.Reshevsky­
R.Fischer, Santa Monica 1966.
b) 13 ll'lxa7 ll'lf4 (or 13 ...ll'ld4 14
ii.xd4 exd4 1S ll'lxc8 .l:f.xc8 16 f3 ll'ld7
with compensation) 14 ii.xf4 (14 ii.f3
ii.d7 1S ll'lbs h6 is pleasant for Black)
14 ...exf4 1S ll'lxc8 .l:f.xc8 16 f3 ll'ld7 gives
Black reasonable compensation for the
pawn.
13 ... b6
13 ... c6 is also possible.
14 a4
Instead 14 ll'lxe6 ii.xe6 1S f3 c6 16
ll'lc3 ii.f8 is fine for Black, while 14 cs
ll'lxcs 1 S .l:f.d8+ ii.f8 16 ll'lxa7 nxa7 17
.l:f.xc8 is probably best met by 17 ....l:f.e8
(17 ... ..t>g7 18 f3 ll'le8 19 a3 ll'ld6 20 .l:f.d8
h6 2 1 ll'lh3 ll'le6 was also okay for Black
 The M a r de/ Pla ta Varia t i o n : White 's E ig h t h M o ve Deviatio ns
in B.Larsen-R. Fischer, Monaco 1967, but
White could have tried 18 ii.xcs!?) 18
.l:f.xe8 ll'lxe8 19 ll'lf3 f6 20 ii.c4+ was
agreed drawn in this even position in
J.Piket-J.Nunn, Wijk aan Zee 1991.
14. ..c6 1s ll'ld6
After 1S ll'lxe6 ii.xe6 16 ll'lc3 a clever
idea is 16 ... :tb7!. Black frees the f8-
bishop and discourages White from
playing a4-aS because then Black
would get play along the b-file. After 17
b4 ii.f8 18 .l:f.ab1 ll'ld7 (with the idea of
... as!) 19 bs .l:f.c8 20 ll'lds ll'lcs 21 bxc6
:txc6 22 as bxas 23 ll'lf6+ ..t>h8 24 J:txb7
lbxb7 2S ll'ld7 ii.d6 White faced an up­
hill battle to draw in L.Portisch-J.Nunn,
Am sterdam 1990.
1s ...ll'ld4
This works tactically and secures
Black equal chances. Also possible is
1s ... ll'lf4!? 16 ii.fl h6 17 ll'lgxf7 l:txf7 18
ll'lxf7 ..t>xf7 19 as bxas with unclear
play in L.Gonda-V.Kotronias, Hungarian
League 2007.
16 ii.xd4 exd4 17 as
Or 17 .l:f.xd4 ll'le8 18 ll'lxc8 .l:f.xc8 19
.U.d2 h6 20 ll'lf3 .l:f.xe4 21 l:tad1 ..t>f8 V2-V2,
D.Polajzer-M.Tratar, Slovenian Team
Championship 199S.
11 .. l:tbS 18 axb6 axb6 19 l:txd4 ll'le8 20
.l:f.ddl
20 ll'lxc8 is more prudent.
20 ... ll'lxd6 21 l:txd6 ii.b7 22 .l:f.a2 h6 23
ll'lf3 l:txe4
Here Black's bishops gave him an
edge in L.Gonda-J.Gallagher, Olbia
2008.
82) 8 ...ll'lg4
This is a more enterprising move.
9 ii.gs f6 10 ii.cl
In stead 10 ii.h4 is sometimes
played, but this combines the negative
aspects of the Gligoric Variation - the
bishop could be stranded on the king­
side - while giving up the positive as­
pect - White's flexibility with his king.
Play will transpose to less critical lines
in Ch apters 9 and 10 after 10... gs
(10...'it'e8 and 10...ll'lh6 are also possible
and could transpose too to the Gligoric)
11 ii.g 3 ltJ h6.
Now Black threatens ... g4 and
...ll'lxd4, so White must react:
161
A ttacking C h e s s : Th e King 's I n dian, Vol u m e 1
a) 12 dS Ci:Je7 13 Ci:Jd2 and now both
13 ... Ci:Jg6 and 13 ...fs are covered in note
'c' to White's 12th move in Line B11 of
Chapter 9.
b) 12 dxes dxes also transposes to
the Gligoric after 13 cs, which is con­
sidered in note 'a' to White's 12th move
in Line A of Chapter 10, or 13 h3, which
is note 'b' to White's 12th move in Line
B of Chapter 10.
10 fs
...
This is consistent. Black plays the­
matically and hopes to show that
White has been wasting time with his
bishop moves. I prefer this to the open
10...exd4 and the unprovoked 10...Ci:Jh6.
162
11 ii.g s
The bishop continues to bounce
around. Other moves also do not lead
to much:
a) 11 exfs exd4!? (11 ...ii.xfs 12 dS
Ci:Jd4 13 Ci:Jxd4 exd4 14 ii.xg4 dxc3 1S
ii.xfs :xfs 16 bxc3 ii.xc3 17 .l:f.b1 b6 is
fine too for Black) 12 Ci:Jbs (after 12 ii.gs
'it'd7 13 Ci:Jbs I like 13 ....l:f.xfs! when
White has little better than 14 ii.cl .l:f.f8,
which is certainly fine for Black)
12 ... .l:f.xfs 13 h3 Ci:Jges 14 Ci:Jfxd4 l:tf7 (or
14 ... Ci:Jxd4 1S Ci:Jxd4 .l:f.f8) 1S ii.e3 a6 16
Ci:Jxc6 bxc6 17 Ci:Jd4 cs 18 Ci:Jc2 ii.b7 with
good counterplay in Bu Xiangzhi­
T.L.Petrosian, Tiayuan 20os.
b) 11 dS Ci:Je7 (Black can also play
11...Ci:Jd4!? 12 Ci:Jxd4 exd4 13 Ci:Jbs Ci:Jes
14 exfs d3 transposing to the Gligoric!)
12 Ci:Jgs Ci:Jf6 looks really stupid for
White.
The position is the same as the
main lines of the Bayonet, but White
has not played b4 or .l:f.e1. Still, when
playing the white pieces one can some­
times get away with a few things and
maintain equality. Indeed, after 13
 Th e M a r de/ Plata Va r i a t io n : Wh ite 's E ig h t h M o ve De viations
exfs! ? ll'lxfs! (better than 13 ... gxfs 14 f4
or 14 'ifb3 !?) the position is fairly level.
c) 11 dxes ll'lgxes 12 exfs ii.xfs 13
ii.e3 (after the hasty 13 ll'lds?! ll'lxf3+
14 ii.xf3 ll'ld4 Black has some initiative)
13 ...'it'f6 (or 13 ... 'it'd7) 14 .U.c1 (14 ll'lds
ll'lxf3+ 1S ii.xf3 'it'f7 16 'it'd2?! would
give Black the initiative after 16 ...ll'les
17 ii.e2 c6) 14....l:f.ae8 1s b3 h6 16 'it'd2
gs 17 ll'ld4 ll'lxd4 18 ii.xd4 and now
18 ...g4 19 ..t>h1 'it'g6 gave Black good
play in J.Granda Zuniga-J.Polgar, Aruba
1992. A promising alternative is
18 ... ll'lf3+ 19 ii.xf3 'it'xd4 20 'it'xd4 ii.xd4
21 .l:f.fd1 (21 ii.xb7? ii.d3) 21 ...ii.cs!
when 22 ii.xb7 ii.a3 win s the exchange.
Returning to 11 ii.gs and here Black
has a pleasant choice:
BU: 11 'ife8
...
822: u. ..J:.f6
. as in B.Lalic-V.Zivkovic, Zagreb 2007,
There is even a third option in
11 ...'it'd7!?, which looks funny but is not
bad:
a) 12 exfs exd4 13 ll'lb s l:txfs ! is note
'a' to White's 11th move, above.
b) 12 dxes ll'lgxes (or 12 ...fxe4 13
ll'lxe4 ll'lgxes 14 ll'lxes ll'lxes with the
idea of ...'it'f7) 13 exfs 'it'xfs 14 'it'd2
ii.e6 1s ll'lds 'iff7 16 ll'lxes ll'lxes 17
.l:f.ac1 .l:f.ae8 was about equal in R.Vera­
R.Slobodjan, Havana 1999.
c) 12 dS ll'ld4!? 13 ll'lxd4 exd4 14
ll'lbs fxe4 1s ll'lxd4 .l:f.xf2 16 ll'le6 .l:f.xf1+
17 ii.xf1 (17 'it'xf1 'it'xe6 18 dxe6 ii.d4+
19 ..t>h1 ll'lf2+ is a perpetual) 17 ...ll'lf6 18
ll'lxg7 'it'xg7 19 'it'd4 ll'lg4 20 'it'xg7+
..t>xg7 21 ii.d8 c6 reaches an unclear
ending.
821) 11...'it'eS
Black keeps pieces on the board.
12 dxes
Others:
a) 12 exfs gxfs (12 ... e4!? is possible,
while after 12 ...ii.xfs 13 h3 ll'lf6 14 dS,
14... ll'ld4!? could be investigated) 13
dxes dxes 14 h3 ll'lf6 1s ll'lds 'iff7 16
ii.e3 h6 gave Black an active position in
B.Kantsler-V.Bologan, Moscow 1991.
b) 12 ll'lds 'it'f7 13 dxes (the creative
13 ii.e7 ll'lxe7 14 ll'lgs 'it'e8 1s ll'lxc7
1 63
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

'it'd8 16 ll'lce6 - 16 ll'lxa8 exd4 is less 13 ...ll'lf6

clear - 16 ... ii.xe6 17 ll'lxe6 'it'd7 18 exfs
gxfs 19 ll'lxf8 .l:f.xf8 20 h3 was S.Lputian­
V.Bologan, Azov 1991, and here
20 ... ll'lf6 looks much better for Black)
13 ... ll'lgxes (or 13 ... dxes, leading to the
note to White's 13th move, below) 14
exfs ii.xfs 1s 'it'd2 l:tae8 was level in
P. S an Segundo Carrillo-V.Topalov, Ma­
drid 1997.
c) 12 dS ll'ld4? 13 ll'lxd4 exd4 14 ll'lbs
'it'es 1s f4! 'it'xe4 16 ii.xg4 fxg4 17 l:te1
'iffs 18 ll'lxd4 ii.xd4+ 19 'it'xd4 was better 14 ii.d3
for White in C.Toth-G.Hemandez, Merida This move has been criticized, but
1993, but 12 ... ll'le7 and 12 ... ll'ld8!? both White does not have any advantage:
look very playable for Black. a) 14 cs ll'lxe4 1s ll'lxe4 fxe4 16 ll'ld2
12 ...dxes (after 16 'it'ds+ 'iff7 both 17 'it'xe4 ii.ts
Preferring the dynamic pawn struc­ and 17 'it'xf7+ .l:f.xf7 18 ll'ld2 ii.e6 give
ture that arises to 12 ... ll'lgxes 13 exfs Black good play) 16 ... ii.e6 17 ll'lxe4 h6
ii.xfs 14 'it'd2 'iff7 with equality in gives Black a very active position.
R.Vera-J.Hebert, Montreal 1997. b) 14 ii.xf6 is the safest. After
13 h3 14...ii.xf6 1s ll'lds Black can try 1s...fxe4,
White can also play 13 ll'lds 'iff7 14 1s ...'it'd8 or the ambitious 1s ...ii.d8!?.
ii.d2 (instead 14 ll'ld2 h6 1s ii.h4 gs 16 14.. .ii.e6 15 �el 'iff7 16 cs
ii.xg4 fxg4 17 ii.g3 ll'ld4 gave Black a Instead 16 exfs gxfs 17 ll'lxes fails
good position in V.Burmakin-V.Bologan, to 17 ... ll'lxes 18 .l:f.xes ll'le4.
Sochi 2004) 14...ll'lxf2! ? (more enterpris­ 16...ll'ld7 1
ing than 14...ll'lf6 1s ll'lgs 'it'd7 16 exfs
gxfs) 1s l:txf2 fxe4 16 ii.e3 exf3 17 l:txf3
'it'd7 18 .l:f.xf8+ ..t>xf8 19 'iff1+ ..t>g8 20
.l:f.d1 (20 ll'lf6+ hf6 21 'it'xf6 looks like a
better try, although Black should be fine
here) 20 ...ll'ld4 21 ii.xd4 exd4 22 ll'lf6+
hf6 23 'it'xf6 'it'e8 24 ii.f3 'if e3+ 2S ..t>f1
ii.e6 26 'it'xd4 'it'xd4 27 .l:f.xd4 cs 28 .l:f.e4
'itf7 gave Black slightly the more pleas­
ant endgame in O.Csoli-G.Pirisi, Hungar­
ian League 2001.

1 64
 The M a r de/ Plata Var i a t i o n : Wh ite 's Eig hth M o ve D e v i a t i o n s
Black already has some initiative.
11 ii.bs
After 17 exfs gxfs 18 ii.e3 h6 Black
also stands very well.
11...ll'ld4!
White is in trouble, since 18 ii.e3
!Lixf3+ 19 'it'xf3 c6 20 ii.e2 f4, with the
idea of ...ll'lxcs, is much better for Black.
Instead 18 ii.xd7 ii.xd7 19 ii.h4 ii.c6 20
!Ligs 'it'f6! 21 exfs gxfs 22 ll'le2 .l:f.ad8 23
�xd4 .l:f.xd4 24 �5? h6 and 0-1 was
l.Sokolov-A.Shirov, FIDE World Champi­
onship, Las Vegas 1999. Here 25 ll'lf3
loses to 2s ...ii.xf3.
822) 11 ii.f6
...
This move was Kasparov's choice
and it was also played with success by
Anand... against Kasparov!
12 ii.xf6 ll'lxf6
Also possible is 12 ...'it'xf6 !? 13 h3
(the alternatives 13 dxes ll'lgxes, 13
exfs ii.xfs and 13 dS ll'ld4!? do not look
dangerous) 13 ...ll'lh6 14 dS (14 dxes
dxes is relatively unexplored, but Black
has done well in practice), but now af­
ter 14 ...ll'ld8 15 cs ll'ldf7 16 cxd6 ll'lxd6
17 ll'ld2 White had some edge follow­
ing both 17 ...ll'lhf7, E.Scarella­
D.Valerga, Villa Martelli 2002, and
11 ... ii.d7 18 .l:f.c1, P.Ricardi-G.Kasparov,
Buenos Aires (simul) 1992 (a game
which Kasparov lost). I think Black can
improve, however, with 14...ll'ld4!.
13 exfs
Instead 13 dS ll'le7 leaves Black with
the active ideas of .. .fxe4 and ...ll'lfs or
.. .f4, and 14 exfs ll'lxfs is certainly okay
for him.
Another possibility is 13 dxes dxes
14 'it'xd8 .l:f.xd8 15 ll'lds (15 exfs can be
met by 15 ... ii.xfs with equality or Kas­
parov's suggestion 1S ...e4) 1s ...ll'lxe4!
16 ll'lxc7 .l:f.b8 17 .l:f.fd1 (better is 17
.l:f.ad1 ii.d7 18 .l:f.fe1) 11 ... ii.d7 18 ii.d3
l:tbc8 (Kasparov suggested 18 ... ll'lf6 and
18 ... ll'lcs! looks good too) 19 ll'lds ll'lcs
20 ii.fl ii.e6 , which is already some­
what better for Black. Indeed, after 21
b4? ii.xds 22 cxds ll'lxb4 23 ll'lxes ll'le4
24 ii.c4 ll'lc3 25 .l:f.d2 ll'lbxds 26 g3 ..t>g7
27 ..t>g 2 ll'le3+ White resigned in
L.Portisch-G.Kasparov, Linares 1990.
13 ...ii.xts
1 65
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
This is more solid than 13 ... gxfs 14
dxes dxes 1 5 'it'xd8 .U.xd8 16 ll'lds ll'le8
17 l:tad1 when White had some pres­
sure in A.Mikhalchishin-K.Hulak, Por­
toroz/Rogaska Slatina 1993.
14 d5
Instead 14 dxes dxes is fine for
Black. His active pieces and dark-square
control compensate for the isolated es­
pawn.
14...ll'le7
If White cannot generate any initia­
tive, the exchange of dark-squared
bishops will favour Black positionally.
1s ll'lgs
After 15 'it'd2 ..t>g7 16 ll'lh4 ii.d7 17
f4 exf4 18 'it'xf4 ll'lfs 19 ll'lxfs+ ii.xfs 20
g4 ii.d7 21 gs ll'lg4 22 'it'xf8+ 'it'xf8 23
.l:f.xf8 .U.xf8 24 ii.xg4 ii.xg4 25 ll'lbs l:tf4
Black was taking over the initiative in
G.Kasparov-V.Anand, Geneva (rapid)
1996.
1s...h6
Instead 1 s ... c6 16 ii.d3 ii.g4 (proba­
bly better is 16 ...'it'b6 17 ii.xfs ll'lxfs 18
ll'le6 .l:f.f7 with the idea ...ll'ld4) 17 'it'd2
1 66
'it'b6 18 h 3 ii.d7 19 dxc6 bxc6 20 .l:f.ae1
.l:f.ad8?! (20 ... l:tae8) 21 ll'la4 'ifC7 22 f4
gave White the initiative in D.Rajkovic­
J.Nunn, German League 1990.
Black has also tried 1s ... 'it'c8?!, but
this is a strange choice of square. After
16 ii.d3 ii.xd3 (16... c6 17 l:tc1!) 17 'it'xd3
White has a slight edge. However, a
better version of this idea is 1 s ... 'it'd7!?
16 ii.d3 c6 17 ii.xfs ll'lxfs 18 ll'le6 .l:f.fe8,
with the idea of ... ll'lg7 or perhaps even
capturing twice on dS to undermine
the e6-knight.
16 ll'le6 ii.xe6 17 dxe6 ll'lfs
Alternatives are 17 ... c6 and 17 .....t>g7.
18 ii.d3 ll'ld4 19 f4
19 ii.xg6 ll'lxe6 is unclear.
19 .....t>g1 20 'it'e1 ll'lxe6 21 'ifg3 gs
The position is unclear. Now 22
fxes?! dxes 23 'it'xes?? 'it'xd3 24 'it'xe6
.U.ae8 25 'iffs ll'le4! 0-1 was the sudden
end of P.Van der Sterren-G.Kamsky,
Wijk aan Zee 1994. Instead 22 fxgs
hxg s with the idea of . . .ll'lhs is unclear.
Black's kingside is rather loose, but his
knights are active.
 Part II
The Classical· Variation
1 d4 �6 2 c:4 g6 3 lbc3 .ig7 4 e4 d6
s ttJf3 o o 6 .ie2 es without 7 o-o
..

1 d4 ltJf6 2 C4 g6 3 ltJc3 Ji.g7 4 e4 d6 5 Here we look at those lines where

0if3 o-o 6 ii.e2 es White does not play 7 o-o. There are
three different systems to consider and
they are vastly different. With the Gli­
goric Variation (7 ii.e3) White main­
tains the tension, which makes it more
difficult for Black to formulate a con­
crete plan. The Petrosian Variation (7
dS) does quite the opposite - White
fixes the centre immediately. Lastly
there is the Exchange Variation (7
dxes), where White resolves the ten­
sion in a rather extreme fashion by
trading queens.

167
 Chapter 9
Gligoric Variation
7 i.. e 3 ttJg4 8 ..tgs f6 9 ..tc1 ttJc6 and 9 i.. h4 ltJc6
1 d4 ll'lf6 2 c4 g6 3 ll'lc3 ii.g7 4 e4 d6 5
ll'lf3 0-0 6 ii.e2 es 7 ii.e3
The Gligoric Variation is a pretty
flexible continuation for White and
this makes it very dangerous. White
keeps the tension in the centre and
does not disclose what he intends to do
with his king. It is not easy for Black
because White can adjust his plans
depending on how Black chooses to
continue. The upside to all of this is
that the Gligoric is not easy to play for
White because Black has eight (!) main
continuations. Because of this, White
168
has to be ready to handle the theory of
several lines, whereas Black only really
needs to know one. White also has to
be prepared to play with several differ­
ent pawn structures. This makes it dif­
ficult to play compared to, for example,
the Petrosian Variation, where White
really only has to understand one
structure. Because of this demand on
the white player, the Gligoric is more
popular amongst titled players than it
is at other levels.
Before we move on to my principal
recommendation, 7 ...ll'lg4, here is a
summary of Black's other options:
a) 7 ...ll'lc6?! is played far too often.
After 8 dS ll'le7 9 ll'ld2 White has the
best of all worlds: an active bishop on
e3 that is typical of Chapter 1; a knight
on d2 that is ready to jump to c4 after
c4-c5 as we saw in Chapter 4; and the
flexibility to castle on either side of the
board (or not at all), as we shall see in
the Petrosian Variation of Chapters 11
and 12.
 Gligo ric Va ria t i o n : 7 ii.. e 3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c 1 tlJ c 6 a n d 9 ii.. h 4 tlJ c 6
b) 7...tlJa6 i s perfectly respectable.
After 8 o-o (not so dangerous is 8 dS
.'Llg4) we are in the main lines of 7 0-0
.'Lla6 8 ii..e 3 which is outside of our rep­
ertoire. This approach has its positive
side, however, because Black only has
to know the 8 ii..e 3 line (8 .l:f.e1 is an­
other major variation after 7 o-o tlJa6).
c) 1 ...tlJbd7 similarly is likely to
tran spose to the line 7 o-o tlJbd7 8 ii.. e 3
when 8....l:f.e8!? 9 dS tLJhs 10 g3 ii..f8 has
recently become fashionable. This is
Bologan's DVD recommendation, but
he does not even consider 8 dS which
leads to what I consider to be a good
variation of the Petrosian variation for
White after 8...tlJg4 (otherwise 9 tlJd2) 9
�d2 ! (better than the more common 9
�g s) 9 .. .fs 10 tLJgs tLJdf6 (10 ... tLJcs 11
b4) 11 exfs gxfs 12 h 3 tlJh6 13 g4 when
White has the initiative.
There is a great trap concealed here
too, as after 13 ... ..t>h 8? 14 tlJe6 ii..xe6 1S
gs! White is down a whole knight, but
Black cannot avoid losing two pieces!
d) 1 ...'ike7 was played by Kasparov
against Karpov in their 1990 match,
but nowadays Black's exchange sacri­
fice 8 dxes (8 dS is also possible)
8 ... dxes 9 tLJds 'ikd8 (9 ... tLJxds 10 cxds
c6 11 d6 is also considered to favour
White) 10 ii.cs tLJxe4 11 ii.. e 7! 'ikd7 12
ii..xf8 ..t>xf8 is considered insufficient
after 13 'ikd3 !.
e) 7 ...exd4 is one of Black's main
continuations. Aft er 8 tlJxd4 l:te8 9 f3 c6
the fi nesse 10 ii..f2 ! is considered to
favour White (but 10 o-o dS 11 cxds
tLJxds! equalizes instantly and 10 'ikd2
dS 11 exds cxds 12 cs tlJc6 13 o-o
.l:f.xe3! is another Kasparov exchange
sacrifice).
f) 7...c6! ? is a fashionable line, al­
though here Black has to know a couple
of different structures. One point of his
move order is that 8 o-o exd4 9 tlJxd4
can be met with 9 ... .l:f.e8 10 f3 dS when
Black has snuck back into the equaliz­
ing line of 7...exd4. Instead White can
play 9 ii..xd4 with some chances of an
edge, but he often plays 8 dS instead.
The blocked position arising after
8...tlJg4 9 ii.g s f6 10 ii..h4 cs remains
topical.
g) 7 ...h6!? was a favourite of John
Nunn's. The point is that after 8 o-o
Black can play 8 ...tlJg4 9 ii..c1 tlJc6 10 dS
tlJe7 11 tLJe1 fS 12 ii..xg4 fxg4 with an
unusual and double-edged pawn struc­
ture. Instead 8 h3 is well met by 8 ... exd4!
9 tlJxd4 .l:f.e8 when 10 f3 would combine
poorly with 8 h3 due to White's dark­
square weaknesses on the kingside. 8 dS
is also considered harmless due to
8... tlJg4 9 ii..d2 fS (equalizing instantly
1 69
A ttacking C h e s s : Th e King 's I n d i a n , V o l u m e 1
according to Nunn), but after 10 h 3 !
(not, though, 10 exfs ii.xfs! 11 h3?
ll'lxf2! 12 ..t>xf2 e4 when Black wins back
the piece with a strong attack) 10 ... ll'lf6
11 exfs gxfs (11 .. .ii.xfs 12 ll'lh4) 12 g3!
ll'la6 13 'ifcl ..t>h7 14 'it'c2 White's clever
manoeuvres gave him a pleasant ad­
vantage in E.Perelshteyn-B.Lopez, US
League 2009. Here Black's ...h6 move has
created weaknesses on the kingside and
hurt his position. We will discuss this
idea again in Chapter 11.
7 ...ll'lg4
This is the main line and it is has
been Radjabov's frequent choice. It is
both principled and forcing. White
cannot very well allow his bishop to be
captured, so he invariably plays:
8 �gS f6
Now White has a choice of retreats.
k t .A.cl. ·
B: 9 JUi4
Instead 9 ii.d2 cuts off White's con­
trol of d4 and makes little sense. After
9 ... ll'lc6 10 dS ll'ld4 Black is doing well.
1 70
A) 9 ii.cl
This move makes a funny impres­
sion, undeveloping White's pieces.
Black has essentially played the moves
...ll'lg4 and ...f6 'for free' and now he
has the move as well. However, these
free moves do not automatically help
Black's position.
9...ll'lc6!
This is the most popular move and
fits in well with our repertoire. Even so,
some players are sceptical about its
value and in New in Chess the strong
Georgian Grandmaster Mikheil
Mchedlishvili wrote that '9 ... ll'lc6 just
won't do for Black'. While I agree with
the ideas presented by Mchedlishvili, I
will dare to challenge his assessment of
9 ... ll'lc6 ...
Also possible is 9 ...ll'ld7 when 10 o-o
ll'lh6 transposes to the line 7 o-o ll'lbd7
8 ii.e3 ll'lg4 9 ii.gs f6 10 ii.cl ll'lh6. Black
has two other independent continua­
tions that are considered respectable in
9 ... exd4 and 9 ... fs, but we will not con­
sider them here.
1o d s
 G lig o ric Va r i a t i o n : 7 ii.. e3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJc6 a n d 9 ii.. h 4 tlJc6
Instead 10 o-o transposes t o the line
7 o-o tlJc6 8 ii..e 3 tlJg4 9 ii.gs f6 10 ii.. c 1
which was covered in Line B2 of Chap­
ter 8, while after 10 h3 Black should
avoid 10 ...tlJh6 11 dS! tlJe7 12 h4!,
which transposes to the note to his
10th move, below, and play 10...exd4
11 tlJxd4 tlJges 12 tLJxc6 (12 ii..e3 fS)
12 ...tLJxc6 intending 13 .. .fs.
10 tlJd4!
...
Instead 10...tlJe7 seems natural. It
looks as though Black is doing well -
compared to lines in the Mar del Plata,
Black is up a few tempi and ready to
play .. .fs. However, White has not cas­
tled and because of that he has extra
options. Another problem is the e7-
knight. If this knight cannot contribute
to Black's kingside play he often suf­
fers. This is a difficult variation for
Black and it is this line for which
Mchedlishvili condemned 9 ...tlJc6. The
specific reason is that 11 h 3 ! tlJh6 12
h4! causes Black major problems.
Some examples:
a) 12 ...fs 13 tLJgs! tlJg4 (13 .. .fxe4 14
g4! controls fS and gives White a big
advantage as shown in several games,
while 13 ...f4 14 g4!, with the idea of
14_.fxg3 1S fxg3, is also much better
for White) 14 h S (or 14 ii..xg4 fxg4 1 S
ii.. e 3) 14...tlJf6 1 S hxg6 hxg6 1 6 ii..f3 and
White has an advantage across the
whole board.
b) 12 ...tlJf7 13 h s cs (after 13 ... gs 14
'it'c2 h6 1s g4! c6 - 1s ... ii..xg4 16 tLJxgs!
ii..x e2 17 tlJe6 wins for White - 16 tLJh2
the battle can only be decided on the
queenside, where White has a space
advantage) 13 ... cs 14 hxg6 hxg6 1 S
ii.. e 3 ii..d7 16 g 3 ! ? (with the idea of ..t>f1-
g2 and 'it'g1-h2) 16 ...a6 17 ..t>f1 fS 18
..t>g2 fxe4 19 tlJxe4 tLJfs 20 ii.gs tLJxgs
21 tlJfxgs 'ife7 22 .l:f.h7 tlJd4 23 ii..g4 .l:f.fs
24 �1 .l:f.xgs 2 s ii..xd7 .l:f.fs 26 ii..xfs
gxfs 27 tlJd2 and White was better in
M.Mchedlishvili-G.Kacheishvili, Tbilisi
2007.
c) 13 .. .fs 14 hxg6 tlJxg6 1s 'it'c2
(White pressures fS and prepares to
castle queenside) 1S .. .f4 16 ii..d2 cs 17
o-o-o a6 18 g3 ii..d7 19 l:LJM tlJxh4 20
l:txh4 fxg3 (a better try is 20...bs!?) 21
fxg3 tLJgs 22 .l:f.dh1 bs 23 'it'd3 (23 .l:f.h s
also brings Black to the brink of defeat)
23 ...bxc4 24 'it'e3 h6 2S .l:f.xh6 ii..xh6 26
.l:f.xh6 tlJf7 27 .l:f.g6+ ..t>h7 28 .l:f.xd6! tlJxd6
29 �6+ ..t>g8 30 'it'g6+ ..t>h8 was
A.Maric-1.Gaponenko, Moscow 2001.
Here the simplest is 31 'it'xd6 when
White has tremendous compensation
for the two exchanges.
Now we return to 10...tlJd4!.
This move is eminently logical.
White has not been developing his
1 71
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
pieces and, although this has granted
him some flexibility, it could clearly
become a problem should the position
open up. This move was not considered
by Mchedlishvili, but I believe it com­
pletely solves Black's problems after 9
ii.cl.
11 ll'lxd4 exd4 12 ll'lbs
After 12 ii.xg4 dxc3 13 ii.xc8 'it'xc8
14 bxc3 Black is already better following
14 ...'if e7 or 14...fs, while 12 'it'xd4 fS is
very risky. One high-level example went
13 'it'd1 fxe4! 14 ii.xg4 'ifh4 1S ii.e2
ii.xc3+ 16 bxc3 'ifxf2+ 17 ..t>d2 ii.g4! 18
cs .l:f.fs! 19 'it'e1 .l:f.xds+ 20 ..t>c2 'it'xg2 21
l:tg1 'it'xe2+ 22 'it'xe2 ii.xe2 23 cxd6
ii.d3+ 0-1, E.L'Ami-L.Nisipeanu, Euro­
pean Championship, Budva 2009.
10...ll'ld4 must have been a shock to
L'Ami, who is usually very well prepared.
12 fs 13 0-0
...
Instead 13 h3 is crushed by
13 ... ll'lxf2 14 ..t>xf2 fxe4+, while 13 exfs
also runs into 14 ... ll'lxf2! 14 ..t>xf2 and
now 14 ... a6! 1s ll'la3 ii.xfs 16 ii.f3 d3
gives Black a strong attack.
The only other option is 13 ii.xg4
1 72
fxg4 14 o-o (14 ll'lxd4 'ife7! 1s 'it'd3 ii.ts
gives Black the initiative), which re­
mains untried. 14...'ife7 looks good
enough after 1S 'it'd3 (1S .l:f.e1 'it'f6!)
1 s...l:te8 (or 1s...cs!? 16 dxc6 bxc6 17
ll'lxd4 ii.b7, with ideas like ...cs and
. ...l:f.ae8) when the e4-pawn gives White
problems. White will be forced to play
f3, either immediately or after 16 .l:f.e1
ii.ts, when ... gxf3 will give Black good
compensation for the falling d4-pawn.
13 ll'les
...
Black should not get carried away
with 13 ... ll'lxh2? 14 ..t>xh2 fxe4 1 S ll'lxd4
�4+ 16 ..t>g1 ii.es 17 f4 exf3 18 ll'lxf3
.l:f.xf3 19 l:txf3 �2+ (or 19 ... ii.g4 20
ii.e3) 20 �1 �1+ 21 �2 'ifh4+ 22
..t>e3 when the white king was safe
enough and the material advantage
counted in J.Ulko-D.Sokolov, Moscow
200S.
14 exfs
White takes the pawn. Instead 14
ll'lxd4 causes no problems at all for
Black. After 14_.fxe4 1 S ii.e3?! (safer is
1 S ll'le6 ii.xe6 16 dxe6, but Black has
several possibilities here, the surest
 G lig o ric Va ria t i o n : 7 ii.. e3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJ c6 a n d 9 ii.. h4 tlJ c6
being 16 ... c6) 1 S ... �4 16 .l:f.c1 (it is al­
ready too late for 16 tlJe6 ii..xe6 17 dxe6
because of 17 ...tlJf3+! 18 gxf3 ii.es 19 f4
.l:f.xf4, as pointed out by Golubev)
16 ... tlJg4 17 ii..xg4 ii..xg4 18 'ifd2 ii..e S
Black already has a stron g attack: 19 f4
(19 h 3 ii..x h3 and 19 g3 � 3 are no bet­
ter) 19 ... exf3 20 tlJxf3 ii..xf3 21 gxf3
.l:f.ae8 and Black was clearly better in
A.Rychagov-M.Vachier Lagrave, Russian
Team Championship 2009.
14 d3!
...
Black cannot save the d4-pawn, so
he uses to it to grab the bishop-pair
and gain time for development.
15 ii..x d3 tlJxd3 16 'it'xd3 ii..xfs
Black's powerful bishops give him
excellent compensation for the pawn.
White will also have some difficulties
developing his queenside, and the c4-
pawn may become vulnerable.
17 'it'dl
White has several alternatives, but
Black always has enough play:
a) 17 'it'e2 a6 18 tlJc3 'ifh4 19 ii..e 3
.l:f.ae8 gave Black good compensation in
R.Vidonyak-D.Stets, Lvov 2009.
b) 17 'it'b3 a6 18 tlJc3 ii..d3 19 l:td1
�4 20 ii..e 3 ii..xc4 was good for Black
in E.otero-R. Leitao, Cuba 1993. If 21
'it'xb7 �ab8 22 'ifxe7 .l:f.xb2 and Black
has the initiative.
c) 17 'ifg3 prevents ...�4. but is an
odd square for the queen.
Black has:
c1) 17...a6 18 tlJc3 ii.es ! ? 19 f4 ii..g 7
20 ii..d2 'it'f6 21 b3 bS! 22 cxbs 'it'd4+ 23
:tf2 axbs 24 .l:f.c1 was F.El Debs-D.Flores,
Campinas 2010, and here 24...'it'cs! ?
gives Black good play.
c2) 17 ... 'it'd7 18 tlJc3 .l:f.ae8 19 ii..d2
ii.es 20 f4 ii..d4+ 21 ..t>h1 a6 22 .l:f.ae1 bS
and Black had counterplay in A.David­
D.Stets, Fourmies 2010.
c3) 17...ii..e s immediately is also
possible: 18 ii.g s!? 'it'd7 19 'it'b3 a6 20
tlJa3 (after 20 tlJc3 ii..d 3 21 .l:f.fd1? 'iffs
threatening ...ii..c 2 is good for Black,
while after 21 .l:f.fe1 Black should play
21 ...'it'fs with compensation, rather
than allow 21 ...'it'g4 22 .l:f.xes! when
White gets counterplay) 20 ... 'ifg7 21
.l:f.ae1 l:tae8 22 tLJc2 (not 22 'it'xb7 ii..d 3)
22 ...ii..xb2 23 tlJe3 ii..c 8 (23 ...ii..d 4!?) 24
1 73
A ttacking C h e s s : The Kin g 's I n dian, Vo l u m e 1

ii.h6 'it'xh6 2 5 'it'xb2 .l:f.e4 was level in active and White's queenside is weak.
D.Yevseev-A.Khruschiov, Peterhof 2009. 21 ii.e3
11 'it'f6
••• This is sensible. White prioritizes
This active move leaves the Cl-pawn developing his pieces.
to its fate, but Black will gain sufficient 21 b6 22 ii.d4 ii.xd4 23 'it'xd4 .l:f.xc4
•••

play. After 17...'it'M White should go for

18 ll'lxe7! as in E.Kanter-A.Savina, Dimi­
trov 2007. However, 17...a6 is plausible:
18 ll'ld4 �4 19 ii.e3 .l:f.ae8 20 'it'd2 ii.xd4
(Black basically plays for a draw against
his much higher-rated opponent with
opposite-coloured bishops, rather than
depending on the bishop-pair for long­
term compensation) 21 ii.xd4 .l:f.e4 22
ii.c3 .l:f.fe8 23 f3 .l:f.e2 24 'it'd4 'it'xd4+ 2 5
ii.xd4 .l:f.c2 2 6 .l:f.fc1 .l:f.ee2 27 .l:f.xc2 .l:f.xc2
28 b3 .l:f.d2 29 ii.c3 .l:f.c2 30 ii.f6 rt;f7 31 The endgame is equal. A couple of
ii.d8 .l:f.b2 32 .l:f.e1 .l:f.b1 33 rtlf2 l:txe1 34 examples are 24 'it'd3 l:te4 25 b3 .l:f.e2 26
..t>xe1 c6 and Black held in L.Fressinet­ 'it'd4 d5 27 a4 'ife4 28 .l:f.ad1 'it'xd4 29
D.Sharavdorj, Las Vegas 2009. .l:f.xd4 .l:f.d8 30 .l:f.fd1 .l:f.b2 with equality in
L.Andonovski-M.Jazadzievski, Struga
2008, and 24 'it'd2 .l:f.e4 25 .l:f.fe1 .l:f.xe1+
26 .l:f.xe1 'it'xa2 27 f3 d5 28 .l:f.e7 .l:f.f7 29
.l:f.e8+ .l:f.f8 30 .l:f.e7 with a draw in
A.Gavrilov-A.Zontakh, Lipetsk 2009.

B) 9 ii.h4

18 ll'lxc7 .l:f.ac8 19 ll'le6

After 19 ll'lb5 .l:f.xc4 Black has excel­
lent compensation for the pawn.
19 ii.xe6 20 dxe6 'it'xe6
•••

White has managed to hold on to

the pawn and eliminate Black's bishop­
pair, but Black's major pieces are all

1 74
 G l ig o ric Va riati o n : 7 ii.. e3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJ c 6 a n d 9 ii.. h4 tlJc6
This is much more challenging.
White keeps his bishop active and hin­
ders the advance of Black's f-pawn. Be­
cause this line is critical for 7...tlJg4, we
will consider two main lines. The move
9 ...tlJc6 is 'consistent', but most players
prefer the modem 9 ... g s, which we will
consider in the next chapter.
9 tlJc6
...
Black sometimes plays 9 ... tlJd7 10
o-o tlJh6, which is a solid line generally
reached from the move order 7 o-o
tlJbd7 8 ii..e3 tlJg4 (8 ... c6 is more com­
mon here, while 8 ... l:te8 is fashionable)
9 ii.g s f6 10 ii..h4 (and here 10 ii..d 2 and
10 ii.cl are seen more frequently)
10 ... tlJh6. The older 9 ... 'it'e8 is not so bad
either, but I have complete confidence
in my two recommendations !
1o d s
Only this move is critical. Instead 10
o-o leads us back to a position consid­
ered in the notes to White's 10th move
in Line B2 of Chapter 8. others:
a) 10 h 3 tlJh6 11 dS (11 dxes dxes
makes little sense because Black plays
... tlJh6 anyway, but this could still
transpose to variation 'b', below)
11 ... tLJe7 and here:
al) 12 'if c2 cs 13 g4 tlJf7 is fine for
Black. The plan is to play ...ii..d7, ...tlJc8
(to protect the f6-pawn), and then
... ii..h 6, activating the bishop. If White
castles queenside, then ... a6 and ... bs
will be enticing. One example was 14
o-o-o a6 1s ..t>b1 ii..d 7 16 tlJd2 'it'as 17
l:tdf1 bS 18 ii..g 3 ii.. h6 with a good posi­
tion in V.Gefenas-J.Van Oosterom, cor­
respondence 1996.
a2) 12 tlJd2 gs 13 ii.. g 3 fs 14 exfs (or
14 f3 tlJg6 with counterplay) 14 ...tlJhxfs
1 S tlJde4 tlJg6 16 'it'd2 tLJf4 17 ii.fl tlJd4
18 o-o-o a6 19 ..t>b1 ii..fs 20 f3 cs 21
ii..d 3 bS was another good example of
Black's possibilities in F.Gheorghiu-
1 .Sokolov, Kava la 1990.
b) 10 dxes dxes (even simpler is
10 ... tLJgxeS; then 11 tLJxes dxes is equal
and 11 o-o can be met with 11 ... ii..e 6,
intending ...'it'd7 and ...fs, or by
11...tLJxf3+ 12 ii..xf3 ii..e 6, again intend­
ing ...'it'd7 and ...fs), and here 11 o-o
tlJh6 is again comfortable for Black.
Others:
bl) 11 'it'xd8 tlJxd8 is equal (but not
11 ... l:txd8? 12 h3 tlJh6 13 tLJds).
b2) 11 tLJds tlJh6 (intending to con­
quer d4 with ... ii..g4) 12 h3 and here
12 ...tlJd4?! is met by 13 tlJxd4 exd4 14
es!, while 12 ... ii..e 6 13 'ifb 3 ! 'it'c8 14
0-0-0 gave White some advantage in
V.Korchnoi-V.Nevednichy, Paks 2004.
However, Black can instead play 12 ... g S
13 ii..g 3 f S with counterplay, V.Shish­
kin-C.Carmaciu, Bucharest 2006.
1 75
A t tacking Ch ess: Th e K i n g 's I n d i a n, Vo l u m e 1
10 ll'le7
...
This retreat is more viable than it
was in Line A, because it is not possible
for White to launch his h-pawn for-
ward with his bishop on h4. The main
concern for Black i s still that the e7-
knight has trouble finding a useful
role.
It would be nice if the untried
10 ...ll'ld4 worked here too, but it looks a
bit too speculative after 11 ll'lxd4 exd4
in view of both 12 'it'xd4 g s 13 ii.g3 fS
14 'it'd1! ll'lf6 (14 .. .f4? 1 s ii.xg4) 1s exfs
(1S f3 fxe4 16 fxe4 ll'ld7 gives Black
some play) 1 s...ii.xfs 16 o-o ll'le4 17
ll'lxe4 ii.xe4 18 'it'd2 'it'f6 19 l:tae1 'it'xb2
20 'it'xg s and 12 ll'lbs .l:f.e8 13 h3 ll'lh6
(13 ... ll'les 14 o-o cs 1s dxc6 bxc6 16
ll'lxd4 'ifb6 17 ll'lb3 also look insuffi­
cient) 14 f3 cs 1S dxc6 bxc6 16 ll'lxd4
'it'b6 17 'it'd2. Well, I tried.
11 ll'ld2
This natural move unleashes an at­
tack on the g4-knight. Instead 11 h 3
tran sposes back to 1 0 h 3 , while 1 1 o-o
will transpose to lines with 12 o-o after
11 ... ltJh6 12 ltJd2 OT 11 ...h S 12 ltJd2.
1 76
After 11 ll'ld2 Black can retreat his
knight or maintain it on g4 for a move.
81: 11 lbh6
•.•
82: 11 h s...
Instead 11 .. .fs 12 ii.xg4 fxg4 13 ii.gs
:tf4! ? (perhaps Black could try 13 ...h6
14 ii.e3 gs 1s cs ll'lg6) is possible. Th is
looks a bit too specul ative, though, af­
ter 14 o-o (there is no point in playing
14 ii.xf4 exf4 because Black will spend
a tempo to force the capture anyway)
14...h 6 1s ii.xf4 exf4 16 ll'lb3 gs 17 ll'ld4
ii.xd4 18 'it'xd4 ll'lg6, as in Y.Shulman­
A.Poluljahov, Gausdal 1994, and here
19 cs would test Black's concept.
Bl) 11 ll'lh6
•••
This is Black's main continuation.
12 f3
White stays flexible. other moves
show his hand too early:
a) 12 g4 ll'lf7 13 'it'c2 cs 14 f3 ii.d7 is
an easy-enough-to-handle set-up for
Black, especially with White committed
to g4.
 G ligo ric Va ria tio n : 7 ii.. e 3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJ c 6 a n d 9 ii.. h 4 tlJ c 6
b ) 12 b4 does really not do that
much, other than rule out castling
queen side. Black can just begin his play
on the other side of the board with
12 ... g s 13 ii.. g 3 fs (or 13 ... tlJg6) 14 f3
tlJg6.
c) 12 o-o gs 13 ii..g 3 is committal
too:
cl) 13 ... tlJg6 14 l:tel (after 14 l:tcl fS
1S exfs ii..xfs 16 tlJde4 tLJf4 17 cs ii.. g6
18 .l:f.e1 tLJxe2+ 19 'it'xe2 tLJfs 20 'it'bs g4!
21 'it'xb7 h S ! Black had a kingside ini­
tiative in L.Van Wely-A.Shirov, Tunja
1989) 14...tlJf4 1 S ii.fl (or 1S tlJf1 fS)
1S ... fs 16 exfs tLJxfs (16 ...ii..x fs! ?) 17
tlJde4 tlJd4 and Black has counterplay.
c2) 13 .. .fs 14 exfs (14 f3 f4 1S ii..f2 cs
also gives Black sufficient play)
14 ... tlJhxfs 1S ii..h s (to prevent
1S ... tlJg6) 1S ...tlJd4 16 f3 bS!? (a creative
idea; instead both 16 ... cs 17 tlJde4 h6
and 16 ... ii..fs 17 tlJde4 tlJg6 look good
enough) 17 cxbs l:tb8 18 tlJde4 tLJxbs 19
'i'd2 (19 tLJxgs .l:f.fs) 19 ...h6 was unclear
in A.Onischuk-V.Bologan, Poikovsky
2004.
12 gs
...
This is the sharpest continuation.
Black initiates play on the kingside
immediately. There is a serious alterna­
tive in 12 ... cs. Black tries to block the
queenside before beginning his king­
side campaign. This has been popular
amongst the top players, but I think it
actually gives White quite a pleasant
choice.
Here 13 a3 prepares 14 b4, but
surely 13 :tb1 is a better way of doing
this. Instead 13 ii..f 2 can be met with
13 .. .fs, while after 13 g4 tlJf7 it is hard
for White to come up with a good plan.
If the bishop moves from h4, then
... ii..h6 is possible, but if the bishop
maintains its post, then White cannot
play h4 and Black can play ...ii..d7,
...l:tb8, ...a6, ...tlJc8, (now f6 is protected),
and ...ii..h6. So White should either ini­
tiate his queenside play immediately or
else change nature of the position:
a) 13 .l:f.b1 has been played fre­
quently by Gelfand:
al) 13 ...tlJf7 14 o-o fS 1S b4 b6 16
a4! ? ii..h6 17 ii..f 2 ..t>h8 18 as tlJg8 19
'it'c2 tlJf6 20 .l:f.b2 tlJhs 21 .l:f.a1 ii..d7 22
axb6 axb6 23 l:txa8 'it'xa8 24 .l:f.a2 'it'd8
2s bxcs bxcs 26 .l:f.a7 and White had
maintained the initiative in B.Gelfand-
1.Nataf, Cannes (rapid) 2002.
a2) 13 ... gs 14 ii..f2 fs 1S b4 b6 16
bxcs bxcs 17 h3 (17 o-o tlJg6) 17 ...tlJg6
18 g4 fxg4 19 hxg4 l:LJM and Black had
counterplay in B.Gelfand-L.Dominguez,
FIDE World Championship, Moscow
2001.
b) 13 dxc6 is also critical, changing
1 77
A ttacking C h e s s : Th e King 's I n d i a n, Vo l u m e 1
the contours of the position. Black's
centre may look imposing, but White is
not really threatened by it and he will
have good squares for his pieces after
13 ...bxc6 14 b4 ii.e6 and then :
bl) 1S ll'lb3 dS 16 ll'lcs (instead 16
cxds cxds 17 exds ll'lxds 18 ll'lxds
ii.xds 19 o-o ll'lfs 20 ii.f2 ii.h6! was to­
tally fine for Black in B.Gelfand­
T.Radjabov, Astrakhan 2010) 16 ... ii.f7
17 0-0 d4 18 ll'l3a4 d4 18 ll'l3a4 ll'lc8 19
ltJ b3 ll'ld6 20 ll'las 'it'e8 21 ll'lcs was
pleasant for White in G.Giorgadze­
T.Shaked, Linares 1997.
b2) 1S ii.f2 !? dS 16 ll'lb3 ! (16 cxds
cxds 17 exds ll'lxds 18 ll'lxds ii.xds 19
o-o ll'lfs 20 ll'lb3 is Gelfand-Radjabov
above) 16 ...d4 17 ll'la4 was V.lvanchuk­
T.Radjabov, Monaco (rapid) 2007, when
White's position is easier to play.
14 C5
Instead 14 o-o ll'lg6 1S cs will trans­
pose to 1S o-o below, but White does
have another option in 14 h3 ll'lg6 1S
g4 fxg4 (after 1S ... f4 White will have a
free hand on the queenside, while
1S .. .fxe4 1S ll'ldxe4 looks very good for
1 78
him) and now:
a) 16 hxg4 ll'lh4 (or 16 ... ll'lf4 intend­
ing ... cs) looks unclear. Black can play
... a6 and _J:tb8 and maybe ... l::tf4, even
as an exchange sacrifice.
b) 16 fxg4 ll'lf4 17 l::th2 cs 18 ll'lfl a6
19 ll'le3 l::tb8! 20 ii.fl ii.d7 was unclear
in J.Gustafsson-V.Nevednichy, Sarajevo
2010. White has a space advantage,
but his king does not have an ideal
home.
14...ll'lg6
This is more flexible than 14... g4 1S
fxg4 ll'lxg4 16 ii.xg4 fxg4 17 o-o ll'lg6
when 18 cxd6 �ess clear is 18 ll'lc4
dxcs!? 19 ii.xcs l::txf1+ 20 'it'xf1 b6 21
ii.e3 ii.a6) 18 ... cxd6 19 ll'lc4 looks a little
better for White.
15 g3
White prevents the knight from
coming to f4, but he can also allow it by
playing 1S 0-0 ll'lf4 16 cxd6 cxd6 17
ll'lc4 .U.f6 !? (17 ... g4 18 fxg4 ll'lxe2+ 19
'it'xe2 ll'lxg4 is also possible) 18 a4 l::t g 6
19 exfs ll'lxfs (not 19 ...ii.xfs 20 g4!,
crippling the h6-knight, S.Shipov­
D.Anagnostopoulos, Cappelle la Grande
 G ligo ric Va ria tio n : 7 ii.. e3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJc6 a n d 9 ii.. h 4 tlJ c 6
1997) 20 ii..d3 .l:f.h6 21 ii..xfs ii..xfs 2 2
�4 .l:f.g6 23 g4 and here 23 ...ii..xe4 24
fxe4 was a little better for White in
S.Shipov-M. Pavlovic, Athens 1997. but
Black could consider 23 ...ii..c8! ? with the
idea of ...h s.
1s ...g4
Black has an interesting alternative
in 1S ...fxe4 16 fxe4 (this has always
been played; after 16 tlJdxe4! ? Black
can play 16 ... tLJfs, which should be
somewhat better for White, or else sac­
rifice a pawn with 16 ... ii..h 3 17 ii..e 3!
0ifs 18 ii..xg s 'it'd7 when he has at least
activated his pieces, but it is not clear if
it is enough for a pawn) 16 ... ii..h 3 17
�e3 b6 and here:
a) 18 cxd6 cxd6 19 tlJc4 'it'd7 20 'ti'b3
(after 20 ii..xg s bS 21 tlJe3 b4 22 ii..x h6
ii..x h6 23 ii..g4 't\Vb7! Black has serious
counterplay, and here 21 tlJd2 b4 also
gives Black the initiative) 20 ...tlJg4 21
ii.. x gs tlJf2 22 .l:f.g1 ii.. g 4 23 ii.. e 3 ii..xe2 24
ii..xf2 was Y.Shulman-B.Socko, Saint
Petersburg 1997. Now 24...ii..d3 ! ? 2 5
.l:f.d1 ii..xc4 26 'it'xc4 ii..h 6 would give
Black good play for the pawn.
b) 18 c6! ? and now 18...a6 is cer­
tainly possible, but more enterprising
is 18 ... 'it'c8 when 19 ii..xg s ii.. g 2 20 .l:f.g1
�3 21 'it'c2 'it'xh2 22 o-o-o 'it'xg3 is
very murky and 19 tlJc4 tlJg4 20 'it'd2
tLJxe3 (20 ...h 6 20 ... ..t>h8) 21 tLJxe3 (21
'it'xe3 'it'd8 regroups) 21 ... a6! ? is un­
clear.
16 fxg4
After 16 exfs gxf3 17 ii..xf3 tlJxfs
Black activates his pieces and has no
problems.
16...tLJxg4
Black has also tried sacrificing a
pawn with 16 .. .f4, but I have trouble
believing that this can be sound after
17 tlJc4.
17 ii..xg4 fxg4 18 o-o
White can also delay this move: 18
tlJc4 (18 cxd6 cxd6 19 tlJc4 tlJh8 trans­
poses, but here too Black could con­
sider 19 ... b6) 18...tlJh8 (possibly not
best; again there is 18...b6!?, while an-
other possibility is 18...ii..d7! ? with the
idea of 19 cxd6 cxd6 20 tlJxd6?! .l:f.xf2 !)
19 cxd6 cxd6 20 o-o tlJf7 21 ii..e 3 ii..d7
22 a4 when White's space advantage
1 79
A ttacking C h e s s : Th e King 's I n d i a n , Vo l u m e 1
brings him a slight but pleasant advan­
tage. Indeed, he maintained an edge
after 22 ... ii.f6 23 '1i'e2 ii.gs 24 'lth1 'ii' e7
2S tllb s ii.xe3 26 tllx e3 in J.Ehlvest­
L.Van Wely, Groningen 1993.
18 ... b6!?
This move forces White to resolve
the queenside tension while preparing
a possible ... ii.a6. Instead 18 ... tllh 8 19
cxd6 cxd6 20 tllc4 leads back to the
previous note.
19 cxd6
White could also try 19 c6, although
this limits his queenside possibilities.
Black could continue with 19 ... ii.a6, but
this might tempt White into sacrificing
the exchange with 20 '1i'xg4!?. Instead
after 19 ... hs White could play 20 .lii . e3,
so perhaps the best move is 19 ...ii.h6 ! ?.
19 ... cxd6 20 ii.e3 ii.d7 21 '1i'e2 '1i'e7 22
tllc4 l:txfl+ 23 l:txfl l:tf8 24 l:txfB+ '1i'xf8
25 ii.d2
2s ... tt:Jhs!
Now that he has developed his
pieces and consolidated his position a
bit, Black begins a standard manoeu­
vre; the knight is bound for gs.
180
26 a4 ii.h6 27 tllb s ii.xd2 28 tllx d2
After 28 '1i'xd2 ii.xbs 29 axbs '1i'f6,
with ideas like ... '1i'g6 and ...tllf7-g S,
Black has no problems.
28... a6 29 tllc 3 bS 30 axbs axbs 31 tllf1
Or 31 tll x bs '1i'b8.
31 ... 'ii'cs 32 'ltg2 'ii'cs 33 tll e 3 hs
Black has gradually taken over the
initiative and soon won in C.Csiszar­
P. S inkovics, Hungarian League 1996.
e2) 11 hs
.••
Fedorov's favourite move is ambi­
tious but quite risky. Black hopes to
gain some space on the kingside before
retreating his knight, but now White
can open the h-file, which can lead to
trouble for Black if he is not alert. On
the plus side, White's bishop blocks the
h-file for the time being and some of
White's dark squares on the kingside
will be weakened if he prises open the
kingside. Whether or not Black can
make use of this is a different matter.
12 h3 tllh 6 13 g4!
This is the only real test of Black's
ambitious play.
 G l ig o ric Va riatio n : 7 il e3 tlJg4 8 il g 5 f6 9 il c1 tlJ c 6 a n d 9 il h 4 tlJ c 6
13... hxg4 14 hxg4 tlJf7
Black would love to play ... 1'.. h 6,
...'ltg7, and ...�h8, but that is not so
easy because White's bishop attacks
the f6-pawn. Therefore Black needs to
move his clumsy e7-knight, but this too
is not that easy. The typical plan for
Black is ... cs, ...1'..d7 and ...tlJc8, but
while Black is making these manoeu­
vres White can start feeding his pieces
towards the kingside.
15 '1i'c2
This move prepares castling queen­
side and is the most popular, but oth­
ers are dangerous too:
a) lS '1i'b3 also prepares castling,
but from here the white queen may
slide along the third rank to threaten
the black king: 1s ... cs 16 tiJd1 ! ? 1'..d7
(Black could also consider the immedi­
ate 16 ...l:te8 or an evacuation plan
starting with 16 ... tiJh8) 17 '1i'h3 (17
'1i'xb7 l:tb8 18 '1i'a6 tlJc8 19 b3 il h6 20
�b2 'ltg7 21 o-o-o '1i'C7 gave Black
some compensation for the pawn due
to the clumsy position of the white
queen in J.Ehlvest-A.Fedorov, Calcutta
1999) 17 ... l:te8 18 tlJe3 looked quite
threatening in J.Ehlvest-V.Spasov, Ko­
caeli 2002, but actually here Black
could try to create some counterplay
on the queenside, starting with
18...'ii' a s!?.
b) lS tiJfl cs 16 tlJe3 (Golubev also
mentions 16 '1i'd3 1'..d7 17 '1i'h3 l:te8 18
'1i'f3 tlJc8 19 tlJe3 a6 20 a4, but while
White may have some advantage, this
does not look fatal for Black; his king is
safe enough for now and he can try to
prepare ... bs) 16 ... 1'..d7 17 1'..d3 was
Z.Azmaiparashvili-A.Fedorov, Elis ta
Olympiad 1998, and here Fedorov sug­
gests 17 ...tlJc8 with the idea of ... il h6.
15 ...c5
After 1s ...1'..d7 16 o-o-o tlJc8 17 ilg3
il h6 White has a thematic device
available.
Indeed, 18 f4! (much stronger than
18 l:th2 'ltg7 19 z:tdh1 l:th8 20 'ltb1 V2-V2,
E.Magerramov-A.Fedorov, Dubai 2001)
18 ... 1'.. xf4 19 ilxf4 exf4 20 es! (the g6-
pawn is the target) 20... tZJxes 21 l:th6
'1i'e7 22 �dhl '1i'g7 23 CZJf3 tlJe7 24
tZJxes fxes (24... dxes 2S cs!) 2S 1'.. d3
1 81
A ttacking C h ess: Th e King 's I n d i a n , Vo l u m e 1
ii.xg4 26 ii.xg6 gave White a winning
attack in F. Peralta-V.Spasov, Lorca
2004.
16 o-o-o
16 ... a6
By playing this immediately, Black is
ready to play ...bs at any time. The al­
ternative is to play 16 ... ii.d7 first, fol­
lowed by either ...a6 or ... tllc8. Some
examples:
a) 17 �dfl and here:
al) 17 ... a6 18 �h 2 tllc8 19 ii.g3 ii.h6
20 f4 exf4?! (an exchange of mistakes;
Black should play 20 ... ..txf4) 21 ii.xf4?!
(21 es!) 21 ... .Lf4 22 �xf4 was
B.Avrukh-E.lnarkiev, Moscow 2002.
Here 22 ... tlle s looks solid enough.
a2) 17 ... tllc8 18 ii.g 3 ii.h 6 19 f4
..txf4! 20 ii.xf4 exf4 21 l:i.xf4 tlle s (but
not 2 L .'ltg7? 22 �hf1 tlles 23 g s ! fs 24
exfs ii.xfs 2s tllce4 tllb6 26 '1i'c3 with a
big advantage in A.Huzman-V.Sara­
vanan, Biel 2000) 22 l:i.hfl gs 23 �4f2
'ltg7 and Black is okay. If White protects
the g4-pawn with a rook, Black will be
able to play ... tllc 8-e7-g6.
b) 17 �h2 tllc8?! (this abandons the
1 82
g6-pawn, so Black must consider
17 ... a6, which could transpose to the
note to White's next move) 18 ii.g3
ii.h6 19 f4! (instead 19 lldhl 'ltg7 20
g S ! ? fxgs 21 'ltb1 gives White compen­
sation, but this is less challenging), and
now Black has problems:
bl) With 19 ... exf4 Black wants to
keep the bishops on the board, but this
fails tactically: 20 es 'ltg7 21 exf6+
'1i'xf6 22 tllce4 '1i'e7 (or 22 ...'ii' es 23 ii.h4
with a winning advantage in F.Peralta­
A.Suarez Real, Leon 2008) 23 ilh4 gs 24
ii.xg s ii.xgs 2s �h7+ 'ltxh7 26 tllxg s+
with a mating attack in T.Enkhbat­
Y.Shulman, Connecticut 2002. An im­
pressive attack especially considering
that Shulman is a strong proponent of
the Gligoric with White!
b2) 19 ...ii.xf4 20 Lf4 exf4 21 es
tllx es 22 l:i.h6 '1i'e7 23 �dhl '1i'g7 24 tl:Jf3
gives White a strong attack similar to
Peralta-Spasov above.
17 �dfl
Instead 17 l:th2 can be met by the
immediate 17 ...bs. After 18 �dhl Black
might try 18 ...b4 or 18 ... ii.d7, with the
 G l ig o ric Va ria tio n : 7 il e3 tlJ g 4 8 il g 5 f6 9 il c 1 tlJ c6 a n d 9 il h 4 tlJ c6

idea of 19 ilg3 '1i'c8 when the pressure Black is very solid.

on g4 will deter White from playing f4 18 ...1'.. h 6 19 f4 b4 20 tiJd1 exf4 21 1'..xf4
for a while. If then 20 f3 b4 21 tiJb1 as 1'..xf4 22 l:txf4 'ltg7
with counterplay. Instead, long after I
had first drafted this section, 19 'ltbl
was seen in B.Gelfand-H.Nakamura,
Amsterdam 2010 (Nakamura had
opted for the move order 17 ...1'.. d7 18
l:i.dhl bS), and after 19 .. �b8 20 1'.. g 3
'1i'c8! (again we see this instructive
idea) 21 f3 �b7 22 'ital '1i'e7 23 f4 one
way to obtain counterplay was
23 ... exf4!? 24 1'..xf4 b4 2S tiJdl as, but
Nakamura preferred 23 .. �fb8 24 fxes
dxes!?, maintaining a tough, unbal­ Black has achieved everything he
anced and approximately level middle­ could hope for and stands well. After
game. the further moves 23 �f2 l:th8 24 l:tfh2
11 bs 1s 1'..g3
... �xh2 2 S �xh2 1'..d7 26 '1i'd3 tlJg8 27
Fedorov suggests 18 tiJd1 ! ? with un­ CZJf2 tlJgh6 28 '1i'g3 '1i'e7 29 tiJf3 �e8 30
clear play. One possibility is 18 ... b4 19 '1i'f4 as 31 �hl a4 32 1'..d 3 a3 33 b3 �h8
ilg3 1'.. h6 20 f4 ilxf4 21 1'..xf4 exf4 22 Black's initiative had become quite se­
l:i.xf4 'ltg7 (but not 22 ...tZJes 23 'llt'b 3 ! rious in B.Gelfand-A.Fedorov, Batumi
with the idea of 'ii'h 3) 23 �hfl tZJgs and 1999.

1 83
 Cha pter 10
Gligoric Variation

1 i.e3 lt:Jg4 s i.gs f6 g i.. h 4 gs

1 d4 tllf6 2 c4 g6 3 tll c 3 ii.g7 4 e4 d6 S

tllf3 0-0 6 ii.e2 es 7 ii.e3 tllg4 8 ii.gs f6
9 ..th4 gs

A: 11 cs
B: 11 h3
This somewhat ugly move really C: 11 dxe5
makes the g7-bishop look bad, but its D: 11 dS
counterpart on g3 will have rather lim­
ited scope as well and it may even be­ A) 11 cs
come a target. This is currently the This aggressive move is quite tricky.
main line of the Gligoric Variation. 11...tllc 6!?
10 ii.g3 tll h6 More common is 11 ...g4 12 tllh4
It is too late for 10 ... tll c6, because af­ tllc6 (White is clearly better after
ter 11 dS tlle7 12 tllxg s! tllxf2 13 ii.xf2 12 ... exd4 13 '1i'xd4 and 12 .. .fs 13 tllxfs,
fxg s 14 ilg4 White is clearly better. but 12 ... dxcs with the idea of 13 dxes
Now White has a broad choice: ii.e6!? could be considered) 13 cxd6 (13

184
 G lig o ric Va r i a t i o n : 7 il e 3 tiJ g 4 8 ilg5 f6 9 il h 4 g5
dxes dxes 1 4 o-o reaches note 'a' to
White's 12th move, below) 13 ... cxd6 14
dxes (of course after 14 ds Black can
play 14...tiJd4) and now Black has a
choice:
a) 14 ... dxes looks fairly solid, but
with the c-pawn s exchanged Black
cannot control the ds-square. Indeed,
White maintains a pull while Black is
devoid of counterplay after 15 1'..c4+
ltih8 16 '1i'xd8 l:txd8 17 o-o tiJd4 and:
al) 18 f4 gxf3 19 tiJxf3 tiJxf3+ 20
:.xf3 looks great for White. After 20.. .fs
21 exfs tiJxfs 22 �afl tiJd6 23 1'..b3 1'..g 4
24 l:i.f6! he was winning in M.Roiz­
B.Socko, Austrian League 2009.
a2) 18 �adl is also good. After
18 ... 1'..d7 19 f4 l:i.ac8 20 1'..d3 tiJf7 21
fxes tiJxes 22 1'..b1 1'..e6 23 1'..f2 tiJec6 24
�fs White maintained the initiative
deep into the endgame in A.Onischuk-
1.Smirin, World Team Championship,
Bursa 2010.
b) 14.. .fxes is more combative, but
it's certainly ugly. Black's justification
for his strange pawn structure is the
floating knight on h4.
White has tried several moves in
this position:
bl) 15 tiJfs tiJxfs 16 exfs 1'..xfs 17
ilxg4 tiJd4 with equality in Z.Ksieski­
B.Socko, Glogow 2001.
b2) 15 tiJbs a6 16 tiJxd6 'ii'a s+ 17
'lt>f1 as played in 5.Slugin-1. Belov, Rus­
sian Team Championship 2009, could
be met with 17 ...tiJd4 with compensa­
tion for the pawn.
b3) 15 tiJds 1'..e6 16 o-o tiJd4 17 tiJe3
�c8 18 b3 'ii' d 7 19 ilc4 ilxc4 20 tiJxc4
1'..f6 21 f4 '1i'e7 22 fxes dxes 23 tiJfs
tiJhxfs 24 exfs h s left Black with the
initiative in A.Poluljahov-R.Anton­
iewski, Koszalin 1999.
b4) 15 o-o!? tiJd4 (worse is 1s ...1'..f 6?!
16 tiJfs tiJxfs 17 exfs 1'..xfs 18 1'.. x g4,
which was pleasant for White in
L.Fressinet-D.Stellwagen, European
Team Championship, Novi Sad 2009,
but Black should consider 1s ... 1'.. e6) 16
.�c4+ 'lth8 17 tiJe2 1'.. e6 (White also re­
tains an edge after 17 ... tiJxe2+ 18 1'.. x e2
1'..f6 19 tiJfs tiJxfs 20 exfs 1'..xfs 21
1'..xg4) 18 1'.. x e6 tiJxe6 19 '1i'd2 tiJf4 20
l:i.adl l:tc8 21 f3! and White had the
initiative in A.Giri-M.Vachier Lagrave,
Biel 2010.
bS) 15 1'..c4+ is the most obvious
move and is probably best: 1s ... 'lth8 16
'1i'd2 tiJd4 17 o-o (White also kept an
edge after 17 tiJe2 tiJxe2 18 1'..xe2 1'.. e6
19 o-o '1i'e7 20 b3 �ad8 21 l:i.adl in
A.Onischuk-E.lnarkiev, Poikovsky 2009)
17 ... 1'.. d7 (this is Black's latest try;
11 ...'1i'e7 18 b3 'ii'as 19 �fd1 1'..d7 20
tiJe2 '1i'xd2 21 l:i.xd2 also left White with
1 85
A ttacking Ch ess: Th e King 's I n dia n , Vol u m e 1
a slight edge in E.L'Ami-G.Agamaliev,
Hoogeveen 2008) 18 tlle2 �c8 19 l:tacl
�xc4 (Black tries to solve his problems
tactically, but White's positional ad­
vantage persists) 20 �xc4 tllxe2+ 21
'1i'xe2 .tbs and Black will win back the
exchange.
However, White is still better: 22
�dl (more complicated but still favour­
ing White is 22 '1i'c2 ii.f6 23 tllfs tllxfs
24 exfs h S 2 s f3 gxf3 26 l:txf3 ii.xc4 27
'1i'xc4 dS 28 '1i'e2 e4 29 �fl, S.Slugin­
Lyaskovsky, Russian Team Champion­
ship 2010) 22 ...'1i'd7 23 �dcl '1i'e6 24 a4
Lc4 2S '1i'xc4 '1i'xc4 26 J:1xc4 ii.f6 27
'ltfl and White eventually won the
ending in S.Volkov-M.Krylov, Moscow
2010.
12 cxd6
After 12 dxes dxes we reach a posi­
tion that is often reached via the move
order 11 dxes dxes 12 cs tllc6, al­
though in that case 12 ... ii.e6 is another
viable option for Black. Here Black's
knight is already on c6, but he can still
hold the balance.
White has:
186
a) 13 o-o g4 14 tllh4 ii.e6 lS '1i'a4
'1i'd4! gave Black good play in
V.lvanchuk-T.Radjabov, Odessa (rapid)
2007.
b) 13 ilc4+ 'lth8 14 '1i'xd8 tllxd8 lS
o-o-o (lS tllb s ii.e6 was also satisfac­
tory for Black in the game Y.Shulman-
1.Umanskaya, St Petersburg 1994)
1S ... tlle6 16 b4 as 17 a3 axb4 18 axb4
c6 19 'ltb2 tllf4 20 ii.xf4 exf4 21 'ltb3
1/2-1/2, P.San Segundo Carrillo-A.Romero
Holmes, Vendrell 1996.
c) 13 '1i'xd8 is the most dangerous
move according to Bologan, who now
gives 13 ... �xd8! (rather than 13 ... tllxd8
14 tllb s �f7 lS o-o-o ii.e6 16 h4 g4 17
tllg S! when White was much better in
V.Grabliauskas-C.Krogh, Ringsted 1992;
here 1S ... tlle6 16 ilc4 is unpleasant but
may be a better try) 14 tlld s g4 l S tlld2
tlld4 16 tllxe7 tllc2+ 17 'ltdl tllxal 18
ilc4+ (no better is 18 tllxa8 ii.e6)
18 ... tllf7 19 tllx a8 ii.h6 20 ii.ds ii.d7 21
tllc7 llc8 22 tlle6 ii.xe6 23 ii.xe6 �xcs
when Black has equal chances in a
complicated endgame. Note that 24
ii.xg4? fails to 24...�c2 with the initia­
tive.
12 cxd6 13 dxes
...
Instead 13 dS just helps Black for­
mulate a plan:
a) 13 ... tlle7 14 o-o tllg 6 lS �el fS 16
exfs tllxfs 17 tll d 2 tllf4 18 ii.fl g4 19
�cl h S ! and Black had already seized
the initiative in E.Levin-S.Klimov, St Pe­
tersburg 2010.
b) 13 ... tlld4!? is also possible: 14
tllxd4 exd4 l S tllb s (not l S '1i'xd4 fS)
 Glig o ric Varia tio n : 7 il e 3 tlJ g 4 8 ilg5 f6 9 il h 4 g5
lS.. .fS! 16 .i.xd6 fxe4 (worse is 1 6. . .l:i.e8
17 es!) 17 1'.. xf8 '1i'xf8 with some com­
pensation. Here 18 tlJxd4? fails to
18 ...'llt'b 4+.
13 ...dxes
14 1'..c4+
White should seize the diagonal
while he can. Black had no probl ems
after 14 o-o 1'.. e 6 in P.H.Nielsen­
E.Mortensen, Festuge 1991.
14...'lthB
This is better than 14... CZJf7 lS h3!
when Black is in a pin and has commit­
ted his knight too early (but lS o-o ilg4
successfully fights for the d4-square).
15 h3
White prevents ...1'..g4, so Black
must find a way to create some coun­
terplay.
1s ... tlJd4 16 o-o g4! 17 tlJxd4 exd4 18
tZJds
Instead 18 tlJe2 gxh 3 19 '1i'xd4 was
P.San Segundo Carrillo-M.Pavlovic,
Ubeda 1996, and now the simplest is
19 ...'1i'xd4 20 tlJxd4 fS with counterplay.
18 ... gxh3 19 '1i'xd4 hxg2 20 �fdl
The position is very complicated. In-
stead 20 'ltxg2 looks critical, but Black
has enough play after 20 .. .fs. One pos­
sible line is 21 es f4 22 Lf4 1'.. e 6 23
ilg3 CZJfs 24 '1i'e4 tZJxg3 2s fxg3 'ii'g s 26
�xf8+ �xf8 27 �el bS! 28 il b3 (not 28
1'.. x bs '1i'd2+ 29 �e2 .i.xds) 28 ...'1i'd2+ 29
�e2 il fS! 30 '1i'e3 (30 �xd2 1'..xe4+ with
the idea of ...1'.. x es is equal) 30 ...'1i'xe3
31 �xe3 �e8 32 tiJC7 �xes 33 �xes
1'.. x es 34 tlJxbs and with such limited
material on the board, Black's good
bishops should ensure a draw.
20 ...1'..g4 21 'ltxg2
White sacrifices an exchange to
fight for the initiative. After 21 �d3 or
21 l:td2 Black has counterplay with
21 ... fs.
21... .i.xdl 22 l:i.xdl fS 23 es
The position is unclear, but fairly
balanced. White has a strong central
presence, but his king is not completely
comfortable and Black is up an ex­
change for a pawn.
2 3 ...f4!?
other interesting possibilities are
23 ... �g8, 23 ...�c8 and 23 ...tlJg4.
24 CZJxf4
187
A ttacking C h e s s : Th e King 's I n d i a n , Vo l u m e 1
Instead 24 ii.xf4 '1i'h4 2 S ii.g3 '1i'xd4
26 �xd4 tllfs is similar.
24...'1i'xd4 2 s �xd4 tllfs 26 �e4 tllxg3
Black could also play 26 ... �ac8.
21 'ltxg3 z:tfs 28 tlld3 �gs+ 29 'lth2 �fB
30 f4 l:tg6 31 .tbs �dB 32 ii.c4 l:tfs 33
ii.bs z:tds 34 ii.c4
Now Black declined a repetition of
moves against his higher-rated oppo­
nent by playing 34... a6! ? in
J.Gustafsson-A.Shimanov, European
Championship, Rijeka 2010, and even­
tually went on to win.
B) 11 h3
This move looks a little funny be­
cause Black usually voluntarily retreats
his knight in any case, but it prevents
... ilg4 and the advance ... g4 is discour­
aged because it would open the h-file
for White.
11 ...tllc6
Both 11 ... exd4 and 11 ... tt:Jd7 are pos­
sible as well, but the text move fits in
best with our repertoire.
12 dS
White has an alternative in 12 dxes
1 88
dxes (I do not like 12 .. .fxes 13 cs!: for
example, 13 ... dxcs 14 '1i'xd8 followed by
lS tllx gs) and now:
a) 13 '1i'xd8 tllxd8 14 tllds tlle6 lS h4
(otherwise, lS tlle 7+ 'ltf7 16 tllxc8
l:taxc8 is level, while lS o-o-o l:i.e8 16
h4, as in G.Timoscenko-E.Hagara, Slo­
vakian League 2008, could be met with
16 ... g4, intending ... c6 and then either
... tlld4 OT... tllc s) 1S ... g4 16 tllh 2 tlld4
(16 .. .fs also gives Black counterplay) 17
tllxo �b8 18 ii.d1 fs 19 exfs tllhxfs 20
tllxg4 tllx g3 21 fxg3 e4! and Black had
the initiative in S.Emst-D.Stellwagen,
Dutch Championship, Leeuwarden
200S.
b) 13 o-o ii.e6 (the solid 13 ...'ii'xdl
14 l:i.fxdl ii.e6 also looks playable) 14
'1i'c1 '1i'e8 lS tllds l:i.c8 16 b4 (Yer­
molinsky instead suggests 16 !tel g4
17 hxg4 ii.xg4 18 tlle 3 ii.xf3 19 ii.xf3
tlld4 20 ii.dl '1i'g6 21 ii.c2 which looks
unclear) 16 ... g4! 17 hxg4 ii.xg4 18 bS
ii.xf3 19 ii.xf3 tlld 4 20 '1i'a3 'lth8 21
l:i.adl (or 21 '1i'xa7 fS 22 exfs tllhxfs
with counterplay) 21 -.fS and Black had
kingside counterplay in A.Yermolinsky­
J.Fedorowicz, US Championship, Los
Angeles 1991 .
12 ... tlle 7
The speculative 12 ...tlld4 13 tllxd4
exd4 14 '1i'xd4 fS lS '1i'd2 f4 16 ii.h2
tllf7 has even been tried by Kasparov
(in a rapid game), but it is probably
objectively unsound.
(seefollowing diagram)
 Gligoric Varia t i o n : 7 11.. e 3 tlJ g 4 8 11.. g s f6 9 1'.. h4 g5
13 '1i'd2!?
An independent try. After 13 tiJd2
we tran spose to 9 ... tlJc6 10 h3 tiJh6 11
dS tlJe7 12 tiJd2 gs 13 1'..g3 which was
discussed in note 'a2' to White's 10th
move in Line B of the previous chapter.
13 ... tiJf7
After 13 ...tlJg6 ! ? 14 o-o-o White has
scored well, but following 14... a6 1S
'iit b 1 1'.. d 7 the position is not clear.
14 tiJh2
Or 14 1'..h 2 cs 1S 1'..g 1 a6 16 tiJh 2 fS
17 CZJf1 which looks rather comical.
Now 17 ...fxe4 18 1'..g 4 CZJfs 19 tZJxe4
bS looked okay for Black in S.Volkov­
A.Shomoev, Ulan Ude 2009, although
White managed to win. Instead the
flexible 17 ... 1'..d7 is natural, while
17 ...l:tb8 and 17 ...'ii'a s!? are also possible.
14 ...f5 15 exf5
After 1S f3 Black could force the
pace with 1S .. .f4 or maintain the ten­
sion with 1s ... tlJg6.
15 ...tZJxf5 16 o-o tiJh8
Black repositions his knight. Similar
is 16 ...tlJd4 17 1'.. g 4 tiJh8 when 18 tlJe2
CZJfs 19 11..h s was S.Shipov-V.Golod,
Minsk 1993, and here Black could con­
sider 19 ... tlJg6 20 1'.. xg6 hxg6, intending
... e4! ? with counterplay.
11 1'.. h 5 CZJd4
Black does not really have to rush
this move and might prefer 17 ... 1'.. d7 ! ?.
18 tlJe4
White could also consider 18 tZJe 2.
Then 18 ...tZJfs would leave Black a
tempo down on the note to his 16th
move, or he could play 18 ... tZJxe2+ 19
'1i'xe2 11..fs, intending ... tlJg6 and maybe
even ... e4. White has a grip on the light
squares, but his g3-bishop is not much
better than its g7-counterpart.
18 ... h6 19 �fe1 1'..f5 20 �acl tlJg6
189
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
Now White played 21 cs? dxcs 22
l:i.xcs b6 and Black won a pawn in
U.Boensch-B.Gelfand, Dortmund 1990.
Instead 21 ii.xg6 ii.xg6 leaves Black
better with his bishop-pair, so White
should settle for 21 ii.g4 with approxi­
mate equality.
C) 11 dxes
White exchanges in the centre,
which often leads to an exchange of
queens.
11 ...dxes
Now White has two main continua­
tions. He can exchange queens imme­
diately with 12 '1i'xd8 or play a little
finesse first by going 12 'ii'd s+.
C1: U �8
cz� u trds+
Instead 12 h3 tllc6 transposes to the
notes to White's 12th in Line B, while
12 o-o c6 is level and 12 '1i'b3 does not
achieve much After 12 ... c6 White can­
not castle long. Note too that 12 cs tllc6
was covered in the notes to White's
190
12th in Line A, but here Black can also
consider 12 ...ii.e6 ! ? covering the a2-g8
diagonal.
A sharp try is 12 h4!?, but Black has
enough resources: 12 ... g4 13 tllh 2 tllc6
14 '1i'xd8 �xd8 1s tlld s tlld4! (Black has
serious counterplay) 16 l:i.c1 (instead 16
ii.d1 fS was Li Wenliang-F.Nijboer,
Groningen 2002, but 16 ...ii.d7 ! ? looks
simple and good) 16 .. .fS! ? (or just
16 ... c6 17 tlle7+ 'ltf7 18 tllxc8 tllxe2 19
'ltxe2 .!:i.axc8 with equality) 17 exfs (17
tllxa �b8 18 exfs tllhxfs also gives
Black good play) 17 ...tllhxfs 18 ii.xg4 c6
19 tlle3 tllx g3 20 fxg3 ii.h6 and Black
had the initiative in l.Gulkov­
S.Dyachkov, Voronezh 1999.
Cl) 12 '1i'xd8 �xd8
13 tlld s
This allows White to gain the
bishop-pair, but that does not trouble
Black Instead 13 tlld2 ii.e6 14 o-o-o
tlla6 1S f3 c6 16 tllb3 tllf7 17 ii.f2 ii.f8
18 �xd8 tllxd8 19 l:i.d1 b6 is equal and
after 20 tll a1 tllb7 21 tll c2 �d8 22 l:i.xd8
tllxd8 the players agreed to a draw in
 G l ig o ri c Va ria tion : 7 ile3 tlJ g4 8 ilg5 f6 9 il h 4 g5
V.lvanchuk-V.Bologan, Foros 2006.
13 ...tlJa6
More ambitious is 13 ... �d7, but this
is obviously riskier. Black clogs up his
own development and the g4-pawn is
weakened. Moreover, he should not
fear the loss of his light-squared bishop
because he will get counterplay on the
dark squares.
14 0R7+
Instead 14 o-o-o 1'..e6 is similar to
lvanchuk-Bologan, above.
14 '>tf7!
.••
Black's king will either have to take
a square away from his knight (on f7)
or his bishop (on f8). From a positional
standpoint, it makes more sense to go
to f7 because the h6-knight will not
necessarily be in a hurry to go to f7, but
the g7-bishop will almost certainly
want to go to f8. There is also a tactical
point, as we shall see.
That said, 14...'lt>f8 is also playable:
15 tZJxc8 �axc8 16 a3 tZJcs 17 tiJd2 �d4
(instead Black should play 17 ... �d7
with equality) 18 f3 l:i.cd8 19 1'..f2 ! (but
not 19 �dl? tlJa4!) gave White a slight
advantage in G.Timoscenko-E.Pinter,
Hlohovec 1998. The point is th at
19 ...l:i.xd2 20 1'..x cs+ is check.
15 tZJxc8 �axes 16 �cl
White wants to safeguard his
queenside. The point of Black's 14th
move i s seen in the line 16 a3?! tZJcs 17
tiJd2 l:i.d4 18 f3 �cd8 19 �dl (after 19
1'..f2 �xd2 20 1'..x cs it's not check and
Black can just play 20 ... �xb2) 19 ... tlJa4!
when White's position falls apart.
White tried 16 �dl tiJcS 17 tiJd2 as
18 b3 1'..f8 19 o-o in R.Akesson-G.Jones,
Klaksvik 2008, and now 19 ... �d4 20 f3
�cd8 2 1 tiJbl tlJe6 would be good for
Black because of his grip on the dark
squares.
16 ... tZJcs
The immediate 16 ...1'..f8 also looks
very good. The bishop threatens a dis­
ruptive check on b4. Now 17 0-0 would
be met by 17 ... g4 18 tiJel �d2, while 17
a3 tZJcs also favours Black.
11 tiJd2 as 18 f3 .i.fB 19 1'..f2 �d7
Black is ready to double rooks on the
d-file and to increase his grip on the
dark squares with ... b6, ... tlJe6 and
191
A ttacking C h e s s : The King 's I n dian, Vo l u m e 1
... ii.cs. White cannot even castle, so he
tries to stir up some counterplay.
20 h4 g4! 21 ..11i. e3
Black has a strong initiative after 21
fxg4 l:i.cd8 22 �c2 tlld 3+ 23 ..11i.xd3 l:i.xd3
or 22 ..txcs ..txcs.
21 �cdB 22 �c2 gxf3 23 gxf3 b6 24
•..
�gl
Now Black erred with 24 ... tlle 6? 2S
cs! which unleashed White's light­
squared bishop in M.Roiz-1.Smirin, Tel
Aviv 2002.
Instead 24 ... tlld 3+ 2S ..11i. x d3 �xd3 26
tllfl and here either 26...tllg 8 or
26 ...'lte6 are comfortable for Black,
while another idea is 24... tllg8, aiming
to bring the knight to d4. After 2S tllfl
tlle7 Black need not fear 26 ii.xcs bxcs
followed by ...tllc6-d4 and ... ..11i. h6 with a
pleasant position for him.
C2) 12 'ii' d S+
This is White's most ambitious try.
Now Black does not want to exchange
queens himself, because after
12 ... 'ii' xds 13 cxds White will have
pressure on the c-file.
192
12 'lthB
•••
Another possibility is 12 ... tllf7 13 h4
(or 13 o-o-o 'ii' e7 14 h4 c6 lS 'ii'a s tlla6
16 hxg s fxgs, reaching the same posi­
tion) 13 ... '1i'e7! (13 ... g4 14 tllh 2 hS lS f3
wins a pawn) 14 hxgs fxgs lS o-o-o c6
16 'ii'a s tlla6 17 tlle1 tllcs 18 '1i'a3 �e8
with unclear play.
13 0-0-0
Instead 13 cs?! c6 14 '1i'xd8 �xd8 lS
tlld2 ..11i.f8 16 tlla4 was V.Korchnoi­
J. Fedorowicz, World Team Champion­
ship, Lucerne 1989, and now 16 ...�d4!
17 b3 tlld7 18 .!:!.cl bS! 19 cxb6 ..11i.b4! 20
l:tc2 axb6 is much better for Black, as
pointed out by Mikhalevski.
White can also exchange queens
with 13 '1i'xd8 l:txd8, although this
should not be dangerous:
a) 14 tlld2 cs!? (or 14 ... ..11i. e6) lS f3
tllc6 16 o-o-o tlld4 17 ..11i. d3 ..11i.e6 18 tllb3
�ac8 was equal in V.lvanchuk­
T.Radjabov, Sochi 2007.
b) 14 tlld s tlla6 1s o-o-o (White
cannot even grab the bishop-pair here
because, of course, lS tlle7 is not check)
1s ... ..11i. e 6 16 tlld2 (16 a3 c6 17 tllc3 tllcs
18 b4 tllb 3+ 19 'ltc2 tlld4+ 20 tllxd4
exd4 21 tlla4 was S.Volkov-A.Shomoev,
Krasnoyarsk 2007, and here 21...b6
looks best, while Black can vary earlier
as well, perhaps with 17 ... tlle7! ? or
18 ... tlld 7) 16 ... �d7 17 f3 c6 18 tllc3 ..11i.f8
19 tllb3 b6 20 �xd7 ii.xd7 21 ..tf2 ..11i. e6
22 �dl tllf7 and the position was bal­
anced in E.Perelshteyn-M.Beelby, Tulsa
2008.
c) 14 h4 g4 lS tlld2 and now
 G lig o ric Va ria ti o n : 7 il e 3 tlJg4 8 il gs f6 9 il h 4 g5
15 ... cs!? is possible, but in practice
Black has preferred to develop his
queen knight:
cl) 1S ... tlJa6 16 o-o-o 1'.. e6 17 CZJb3
b6! (the inferior 17 ... c6?! 18 tZJas �xdl+
19 l:txdl �b8 20 a3 gave White an edge
in A.Onischuk-A.Shomoev, Sochi 2007,
but Mikhalevski's 17 ...CZJf7 18 tZJas
l.h6+ 19 'ltbl tiJd6 looks solid enough)
18 f4 gxf3 19 gxf3 c6 20 'ltbl CZJf7 21 f4
exf4 22 1'..xf4 fs and Black had enough
play in V.Bhat- B.Sambuev, Montreal
2009.
c2) 1S ... tlJc6!? (Black logically takes
aim at the d4-square, but there are
some tactical issues) 16 tZJds (16 CZJb3 is
possible as well, but the text move is
certainly critical) 16 ...tlJd4!? (going for
it, but 16 ... l:td7 17 CZJb3 �f7 is safer) 17
!Uxo tZJc2+ (not 17 .. Jlb8 18 �cl with a
healthy extra pawn) 18 'ltdl tlJxal 19
tlJxa8 CZJfs! (a clever resource) 20 tZJo
(not 20 exfs 1'..h 6!) was played in
5.Feller-M.Al Modiahki, Paris 2010. Now
Black could justify his play with
20 ... tZJxg3 21 fxg3 �d4! (but not
21...1'..h6 22 tZJds, blocking the d-file) 22
tZJbs (22 'Itel 1'.. h 6 23 �dl �xe4 24 1'..fl
1'..d7 looks fine for Black) 22 ...�d8 23
'Itel 1'..h 6 24 l:i.dl a6 25 tlJc3 �d4 with a
strange situation. The al-knight is
trapped, but White can hardly win it.
He may exchange it indirectly with 26
'ltbl, but then 26 ... .i.xd2 27 'ltxal?!
ilxc3 28 bxc3 l:i.xe4 29 �d8+ 'ltg 7 cer­
tainly favours Black
13 '1i'e7
••.
14 'ii' as
Slow is 14 h3 and 14... tlJa6 15 'ii' as
c6 was fine for Black in P.Wells-L.Kritz,
Biel 2004.
Instead 14 h4 g4 15 tiJel tlJc6 16
tZJc2 fs 17 exfs was K.Sundararajan­
R.Shetty, Visakhapatnam 2004, and
here the obvious 17 ... CZJxfs gives Black
counterplay.
14 ...c6
After the natural 14... tlJc6 White has
15 tZJds ! tZJxas 16 tZJxe7 1'..e6 17 tZJds
l:tac8 18 tiJd2 with an edge,
E.lturri zaga-5.Feller, Moscow 2010.
15 tiJd2 b6 16 'ii' a4 1'.. d 7 17 f3 cs!?
Black gives up the ds-square, but he
will get d4 for his knight in return.
193
A ttacking C he s s : Th e King 's In dian, Volu m e 1
18 '1i'a3 tllc6 19 tlld s •t1 20 ii.d3 ii.e6
The position was unclear in
A.Motylev-T.Radjabov, Wijk aan Zee
2007. Black's kingside set-up may look
funny, but he has possibilities to initi­
ate play with ...f5 or ...b5.
D) 11 d 5
Closing the position is White's most
popular option. After Black breaks with
.. .f5 White will gain the e4-square for
his pieces. Black will have to either
fight for control of this square or gen­
erate enough counterplay to offset
White's grip on the light squares.
Here a decision must be made.
194
Black may break immediately or he can
develop first in order to better fight for
the e4-square.
D1: 11...f5
02: 11....�7
D1) 11 ...fs 12 exfs
12...g4!?
This invention of Kasimdzhanov
commits Black to a pawn sacrifice. The
alternative is 12 ... tllxf5 (but not
12 ... ..txf5?! 13 h4 g4 14 tllg 5 ! ) 13 tlld2
tlld4 and here:
a) After 14 tlld e4 g4! (intending ...h 5
with counterplay) White could still play
15 ii.xg4 ii.xg4 16 '1i'xg4 tllc2+ 17 'ltd2
tllxal 18 �xal with some compensa­
tion, although Black is probably not
worse.
b) 14 ii.d3 was suggested by Bolo­
gan. After 14 ... ii.f5 (?!) 15 ii.xf5 l:txf5 16
tllde4 White is better, but Black should
prefer 14... tlla6, as in E.Karavade­
T.Abrahamyan, Yerevan 2006.
c) 14 o-o tllx e2+ (14...ii.f5 15 tllde4
tlld7 16 ilg4 was a little better for
 G l ig o ric Varia tio n : 7 ile3 tb g 4 8 il g5 f6 9 i.. h4 g5
White in S.Gross-V.Golod, Ceske Bude­
jovice 1993) lS '1i'xe2 '1i'e8 16 tl:Jde4
1t'g6 17 f3 ilfs (instead 17 ... tl:Ja6 18 a3
g4 19 b4 keeps the knight sidelined)
was Ki.Georgiev-V.Bologan, Gibraltar
2006. Here White could play 18 cs im­
mediately with the initiative.
13 tl:Jd2
After 13 tl:Jh4? Black wins a piece
with 13 ... 1'..f6. White does not obtain
enough compensation after 14 h3
ilxh4 lS hxg4 1'..x g3 16 fxg3 tl:Jf7 with
the idea of ... 0,g s.
13 ... 1'..xfs
14 tl:Jfl
This move wins a pawn and is
therefore critical. 14 tl:Jde4 is less dan­
gerous, leading to 14 ... tl:Jd7 and then:
a) lS 1'..d 3 'lth8 (1s ...'1i' e8? !) 16 o-o
as 17 '1i'e1 1'..g 6 18 '1i'e2 '1i'e7 was about
equal in L.Gofshtein-R.Kasimdzhanov,
Hoogeveen 1999.
b) lS 0-0, after which the following
play is very thematic: 1s ...'1i'e8 16 1'..d3
'1i'g 6 17 '1i'c2 'lth8 18 �adl a6 19 b4 tl:Jf6
20 tl:Jxf6 '1i'xf6 2 1 tl:Je4 1'..x e4! (a typical
idea; Black will target the g3-bishop) 22
ilxe4 tbfs 23 '1i'd3 hS 24 f3 (or 24 cs as!
with counterplay) 24 ... tl:Jxg3 2S hxg3
'ii'g s 26 'ltf2 �f6 (simpler is 26...1'..h 6)
27 '1i'e3 '1i'xe3+ 28 'ltxe3 1'..h 6+ 29 'ltf2
'ltg7 30 �h l �h8 31 'lte2 1'..g s?! (Black
can equalize with 31 ... gxf3+ and then
32 gxf3 iJ..g S OT 32 iJ..xf3 aS!) 32 fxg4
hxg4 33 �xh8 'ltxh8 34 a4 'ltg7 3S as
and White kept an edge in R.Hungaski­
D.Lemos, Riobamba 2007.
14...tl:Jd7
Another idea is 14 ... 'ii'g s lS tl:Je3
i..d7 16 tl:Je4 '1i'g6 17 1'.. d3 tl:Jfs with
counterplay. Here Huzman gives 18
tl:Jxg4?! tl:Jxg3 (or 18...hs!?) 19 tl:Jef6+
ilxf6 20 1'..x g6 tl:Jxhl 21 1'..fs ! i..g 7! 22
1'..e6+ (Black is also doing well after 22
ilxd7 tbxd7 23 '1i'e2 hS 24 o-o-o �f4!)
22 ...1'.. xe6 23 dxe6 tl:Jc6 24 'ii'd s hS 2S
e7+ �f7 and because White will not be
able to round up the hl-knight, Black is
much better.
15 tl:Je3
1s ...e4
Th is looks like the best move order.
Instead 1s ... tl:Jcs can be played when 16
o-o e4 transposes to the main line, but
195
A ttacking C he s s : The Kin g 's In dian, Vo l u m e 1
White can also try:
a) 16 b4 e4 17 �cl tlld 3+ 18 ii.xd3
exd3 19 tllxfs tllxfs gives Black some
counterplay: for example, 20 o-o (much
worse is 20 '1i'xg4? '1i'f6 with the idea of
21 o-o d2 22 l:i.c2 tlle 3) 20 ... tt:Jd4 (other
possibilities are 20 ... h s and 20 ...'1i'f6) 21
'1i'xd3 hS, targeting the g3-bishop.
b) 16 tllxfs tllxfs 17 ii.xg4 e4 and
now:
bl) 18 ii.xfs ii.xc3+ (White is much
better after 18 ... tlld 3+ 19 'ltfl l:i.xfs 20
'1i'g4 '1i'f6 21 tllxe4 '1i'xb2 22 �dl '1i'c2 23
'1i'e2 '1i'xe2+ 24 'ltxe2) 19 bxc3 l:txfs
(Black's compensation looks insuffi­
cient after 19 ... tlld 3+ 20 'ltfl l:txfs 21
'1i'g4+ 'ii'g s 22 'ii'x gs+ �xgs 23 'lte2) 20
o-o 'ii'g s with some compensation.
b2) 18 o-o tllx g3 (Huzman prefers
18 ... ii.xc3 19 bxc3 'ii'g s with some
compensation) 19 hxg3 ii.xc3 20 bxc3
'ii'g s 21 ii.e6+ tllx e6 22 dxe6 �fs 23
'1i'd4 was agreed drawn in A.Huzman­
B.Avrukh, Pula 2000. Instead of 23 '1i'd4,
Huzman suggested 23 'ii' bl. This has
since been seen in practice and al­
though White eventually won after
23 ... �es 24 �ael 'ii'fs 2s e7 �xe7 26 f3
l:tae8 27 fxe4 '1i'g6?! 28 'ii' ds+ l:te6 29
'1i'xb7 in H.Grooten-N.Vanderhallen,
Vlissingen 2006, Black could have
maintained the balance with the move
27 ...'ii'g s ! .
16 o-o tllc s
Also interesting is 16 ...ii.xc3 17 bxc3
'ii'g s with an unclear position in
F.Atakisi-N.Gavrilakis, correspondence
2006.
196
17 l:tcl
White lends the c3-knight some ex­
tra protection and creates the possibil­
ity of playing b2-b4. Others:
a) 17 ii.xg4 is untried, but Black
should have enough for the pawn after
17 ... ii.xc3 18 bxc3 tllxg4 (or perhaps
18 ... ii.xg4 19 tllxg4 'ii'g s) 19 tllxfs (19
tllxg4 'ii'g s 20 tlle 3 tlld3 gives Black
counterplay) 19 ... l:txfs 20 '1i'xg4+ 'ii'g s.
b) 17 tllxg4 ii.xg4 18 ii.xg4 ii.xc3 19
bxc3 '1i'f6 gives Black some compensa­
tion. After 20 ii.e6+ (or 20 .!:!.cl tllxg4 21
'1i'xg4+ '1i'g6) 20 ... tllxe6 21 dxe6 '1i'xe6
22 cs tllfs 23 '1i'g4+ 'lth8 24 cxd6 cxd6
2S l:i.fel the game was agreed drawn in
R.Markus-N.Djukic, Herceg Novi 200S;
2s ... l:tae8 would be unbalanced, but
level.
c) 17 '1i'd2 'ii' gs (or 17 ...ii.d7 !? 18 b4
tlla4 forcing White to make a reason­
able exchange sacrifice) 18 l:i.adl a6 19
b4 tlld7 20 cs tlle s (not 20...dxcs 21 bxcs
tllxcs 22 ..txa) 21 ii.xes ii.xes 22 tllxfs
'ii'xfs (and not 22 ...'1i'xd2 23 l:txd2 tllxfs
24 tllxe4 when White's extra pawn
means more than it does in the game)
 G lig oric Va riati o n : 7 il e 3 tlJ g 4 8 ilg5 f6 9 il h 4 g5
23 '1i'xh6 1'..xc3 24 cxd6 �f6! 2S '1i'h4
(White really had to try 2S '1i'e3 or 2S
'1i'c1) 2s...cxd6 26 1'..xg4 'ii'e s. Now Black
held without much difficulty against his
higher rated opponent: 27 f4 exf3 28
J:1xf3 �xf3 29 ilxf3 �f8 30 'lthl '1i'f4 31
'1i'xf4 l1xf4 32 a3 1/2-1/2, A.Huzman-
1.Sidorenko, Ramat Aviv 2004.
11 ...as
Black secures the cs-knight, but also
possible is Mikhalevski's 17 ...'1i'd7!? to
cover fS and g4. After 18 b4 (18 �el
gives Black time to play 18 ... as and 18
ilf4 CZJd3! 19 CZJxfs! 'ii'xfs 20 1'..xh6
1'.. xh6 21 1'.. xg4 '1i'g6 22 l:tc2 e3! gives
him counterplay according to Mik­
halevski) 18 ... CZJd3 19 .i.xd3 exd3 20
CZJxfs 'ii'xfs looks okay for Black.
18 tlJxg4 tlJxg4 19 1'.. xg4 'ii'g s 20 1'.. xfs
l:i.xfs 21 '1i'c2 e3 22 �eel exf2+ 23 .i.xf2
.:!.af8
24 tlJe4!?
White return s the pawn to neutral­
ize Black's activity. 24 .lii.xcs is also pos­
sible, but after 24 ... dxcs 2s �xfs �xfs it
will not be easy for White to make pro­
gress.
24...tlJxe4 25 '1i'xe4 1'.. es
Black threatens ... 1'..xh2+. Instead
2s ... 1'..xb2 26 1'.. e3 l:txfl+ 27 l:txfl �xfl+
28 'ltxfl is also possible and after
28 ... 'ii'hs Black should hold.
26 1'.. e3 l:i.xfl+?
This is a mistake. Instead 26 ...'ii'h S!
holds after 27 l:txfs (no better are 27 g 3
l:txfl+ 2 8 l:i.xfl �xfl+ 29 'ltxfl '1i'xh2
and 27 h3 �xfl+ 28 �xfl 1'..h2+ 29
'ltxh2 l:txf1) 27 ... 1'.. x h2+.
27 l:txf1 �xfl+ 28 'ltxf1 'ii' h s 29 h3
1'.. xb2 30 '1i'e6+ 'ltg7 31 '1i'd7+ '1i'f7+ 32
'ii'xf7+ 'ltxf7 3 3 a 4
This was M.Roiz-T.Radjabov, Saint Vin­
cent 200S. White has a clear advantage
in the bishop ending because Black's
pawns are fixed on dark squares and
White can create an outside passed
pawn on the kingside .
D2) 11 ...tiJd7
Black makes this move in prepara­
tion for .. .fs, so that the knight can
quickly come to f6 to fight for the e4-
square.
12 tiJd2 fs 13 exfs
197
A ttacking C h es s : The Kin g 's I ndian, Vo l u m e 1
The alternative is 13 f3 tllf6 (intend­
ing ... g4) 14 h3.
Here Black has:
a) 14...tllh S lS ii.f2 tllf4 16 g3 tllxe2
17 '1i'xe2 cs (17 .. .fxe4!? looks better, as
after 18 tlldxe4 Black can activate his
knight with 18...tllfs) 18 g4! f4 was
B.Gelfand-1.Nataf, Cap d'Agde 2002.
Here 19 h4 gives White an obvious ad­
vantage.
b) 14 ... cs lS ..tf2 (White's play has
been too slow to consider lS dxc6?!
bxc6 with ...ds coming) 1S .. .f4 16 �bl
(White tried the slower 16 a3 in
E.Bacrot-T.Radjabov, Morelia/Linares
2006) 16 ...b6 17 b4 cxb4 18 �xb4 tlld7
19 a4 tllcs 20 as ii.d7 was solid for
Black in B.Gelfand-T.Radjabov, Turin
Olympiad 2006.
c) With all of White's pawn moves,
it is logical to open the position with
14 .. .fxe4 and Black has scored well with
this move: lS tlldxe4 (after lS fxe4
Black can play 1S ... g4 or 1 S ... tlld7, in-
tending ...tll cs and perhaps ...l:tf4!?)
1s ...tllfs (Black can also play 1s ...tllxe4
immediately when 16 tllxe4 tllfs trans-
198
poses and 1 6 fxe4 can be met with
16 ... g4 17 hxg4 tllx g4! 18 ii.xg4 'ii'g s) 16
ii.f2 tllxe4! (this weakens White's con­
trol of dS) 17 tllxe4 (17 fxe4 tlld4)
17 ... cs 18 '1i'd2 (18 dxc6 bxc6 19 '1i'd2 dS
20 cxds cxds 21 tllcs tlld4 22 ii.xd4
exd4 is unclear according to Bologan)
18 ... h6 19 o-o-o (or 19 o-o tll d4 with a
good position) 19 ... a6 20 g4 tllh4!
(20...tt:Jd4 21 h4!) 21 ii.xh4 gxh4 22 ii.d3
(better is 22 gs with complications)
22 ...�f4! 23 'ltb1 l:i.b8 (or 23 ... bs!? im­
mediately) 24 '1i'e2 '1i'e7 2S a3 ii.d7 26
'lta2 bS and Black had the initiative in
A.Huzman-V.Bologan, Moscow 2006.
Moreover, Black can even improve on
this with Bologan's 19 ... bS!? 20 cxbs a6
21 bxa6 (21 b6 �b8 22 'ltbl '1i'xb6 23 h4
gxh4) 21 .....txa6 22 ii.xa6 �xa6 23 'ltbl
'1i'a8 24 tllc3 �b8 with a strong initia­
tive.
13 tllf6
...
This is the point of Black's play.
14 tllde4
Instead 14 o-o ii.xfs is not danger­
ous, but White can play 14 ii.d3 tllxfs
lS tllde4 and here:
 Gligoric Va ria tio n : 7 il e 3 tlJ g 4 8 il g 5 f6 9 il h 4 g 5
a) 1S ...1'.. h6 1 6 o-o 'lth8 17 cs g4 18
tiJxf6 '1i'xf6 19 tZJbs '1i'e7 20 '1i'e2 and
then:
al) 20... tZJxg3 21 fxg3 �xfl+ 22 .!:i.xfl
dxcs 23 '1i'e4! threatens 24 d6! and
gives White a winning attack
a2) 20 ... 1'.. g7 21 cxd6 cxd6 22 '1i'xg4
tlJxg3 23 '1i'xg3 1'.. d7 was V.Kramnik­
T.Radjabov, Wijk aan Zee 2007. After 24
'1i'e3 ! Black would not have enough for
the pawn.
a3) 20 ... a6 21 1'..xfs �xfs (not
21 ... .i.xfs 22 cxd6 cxd6 23 tlJxd6) 22
tiJc3 �f8 23 tlJe4 'ltg8! 24 '1i'c4 1'..f s is
okay for Black according to Bologan.
b) 1S ...tZJxe4 16 tZJxe4 tiJh4! 17 o-o
(or 17 ilxh4 gxh4 with the idea of
...1'..fs) 17 ... g4 18 0.d2 (if 18 cs then
18 ... tlJg6 intending ... tiJf4 or ...h s)
18 ...'ii' g s 19 tlJe4 '1i'e7 20 tiJd2 h S ! ? 21 f3
CZJfs 22 1'..f2 1'..h6 23 '1i'e2 '1i'g7 24 1'..xfs
1'..xfs 2S 1'..e3 l:tf7 (Bologan suggests
2s ... 1'..f4!?) 26 �ael l:i.af8 27 1'.. xh6
'1i'xh6 28 '1i'e3 '1i'xe3+ 29 �xe3 gxf3 30
l:tfxf3 was Wang Yue-V.Bologan, Ji­
angsu Wuxi 2008. Here 30 ... 'ltg7 would
leave Black with a slight advantage in
the endgame.
14 tZJxe4 15 tZJxe4 1'.. xfs 16 1'.. d3
•••
A good example of Black's possibili­
ties is 16 f3?! g4 17 1'..f 2? (better is 17
o-o) 17 ...gxf3 18 gxf3 tlJg4 19 1'.. d3 CZJxf2
20 CZJxf2 'iih4 21 o-o e4 22 tZJxe4 1'..h 3
23 �el 1'..e s 24 '1i'e2 'lth8 2 s 'lth1 �xf3
26 '1i'xf3 ilg4 27 '1i'f2 ilf3+ 28 'ltgl
1'..x h2+ 0-1, J.Donner-S.Gligoric, Eersel
1968.
16 ...g4 17 0-0
White can also prepare to castle
long, but Black has resources: 17 '1i'e2
'lth8 18 o-o-o '1i'e8 19 'ltbl '1i'g6 20 h3
gxh3 21 gxh3 tlJg8! (a typical regroup­
ing) 22 1'.. h4 1'.. xe4! 23 1'.. xe4 �6 24
ilg3 CZJf6 2S 1'..f 3 e4 26 1'.. g4 tlJxg4 27
hxg4 '1i'g6 28 'ital as and Black had the
initiative in J.Dorfman-1.Nataf, Mon­
dariz 2000.
White has control of the e4-square,
but Black has gained some space on
the kingside. White will be able to initi­
ate queenside play with c4-cS, but
Black's main concern is the activation
of his h6-knight.
17 'lthB
•••
The king is a little safer here and
Black can consider regrouping with
... tlJg8 at some point. Nataf has used
this move frequently, but Black has
some choice here and it is not clear
which move is best. There are two
other logical tries:
a) 17 ... 1'..xe4 really tries to force
things when 18 1'..xe4 'ii'g s protects the
g4-pawn and intends ... tZJfs and ...hs.
White has:
199
A t tacking C h e s s : Th e Kin g 's Indian, Vo l u m e 1
al) 19 'itb3 !Iab8 20 cs lLifs 21 'it'a3
was suggested by Mikhalevski, but
19 ...b6 looks like a better try.
a2) 19 'it'd3 lLifs 20 !Iae1 !If7 21 cs
!Iaf8 22 cxd6 cxd6 23 !Xcel a6 24 b3 hS
gives Black counterplay. After 2S f4
(forced) 2s ... lLixg3 26 fxgs !Ixf1+ 27
!Ixf1 !Ixfl+ 28 'it'xf1 lLixfl 29 @xfl Black
had to work a little bit to draw in
R.Markus-V.Sikula, Hungarian League
2004. Instead Black could have tried
2s ...'i'h6 26 ..ixfs !Ixfs 27 fxes !Ixf1+ 28
!Ixfl !Ixfl+ 29 @xf1 h4! with equality.
a3) 19 'it'c1 'it'hs (19 ...'it'xcl followed
by 20 ...lLifs is also possible) 20 cs !If7 21
cxd6 cxd6 22 'it'c2 lLifs (after 22 ... ..if6 23
!Iad1 !Iaf8 24 'it'a4 lLifs 2s 'it'xa7 lLid4 26
'itb6?? ..id8 White was facing ...lLie2xg3
and resigned in J.Vanheste-J.Piket, Am­
sterdam 1988, but of course he can im­
prove with 26 'it'a3) 23 ..ixfs 'it'xfs 24
'it'xfs !Ixfs 2s !Iac1 e4! gave Black
enough counterplay in D.Kuljasevic­
D.Popovic, Zupanja 2008.
b) With 17 ... 'it'e8 Black brin gs his
queen to g6 before addressing his side­
lined knight.
2 00
White has:
bl) 18 !Iel 'if g6 19 'if d2 (19 cs!
transposes to variation 'b2') 19...!If7 20
b4 !Iaf8 21 ..ih4 ..ixe4! 22 ..ixe4 lLifs 23
cs ..ih6 24 'it'd3 ..igs 2s ..ixg s 'it'xg s 26
..ixfs !Ixfs gives Black good play in a
position with all the major pieces. The
following game serves as a good illus­
tration of Black's chances in this type of
position: 27 !Ie2 h S 28 c6 b6 29 'it'a6?
h4 30 'it'xa7 !Ixf2 ! 31 !Ixf2 'it'e3 and 0-1
in P.Smirnov-D. Kryakvin, Belorechen sk
2009.
b2) 18 cs looks more to the point:
18 ... 'it'g6 19 !Iel lLif7 20 ..ih4 (White
was also better after 20 !Icl lLig s 21
lLixgs ..ixd3 22 lLie6 in Yang Shen­
Xiaowen Zhang, Jiangsu Wuxi 2008)
20 ...!Iae8 (20...'it'h6 21 g3 is not such a
help to Black) 21 !Ic1 dxcs 22 !Ixcs lLid6
23 'ifa4 ..ixe4 24 ..ixe4 'it'h6 (perhaps
better is 24 ... lLixe4 2S 'it'xe4 - but not
2S !Ixe4? !If4! - 2S ...'it'xe4 26 !Ixe4 !If4
27 !Iec4 !Id4 28 @f1 bS 29 !Ic3 a6 as
suggested by Mikhalevski, but Black is
still worse) 2S ..ig3 'it'd2 was L.Aronian­
T.Radjabov, Morelia/Lin ares 2006. Now
26 !Id1 'it'e2 27 ..id3 'it'xb2 28 !Ixe7
would leave White much better.
18 .l:i:el
Instead 18 'it'e2 ..id7 19 f3 'it'e7 20
fxg4 and Vi-Vi B.Gelfand-T.Radjabov,
Russian Team Championship 2006, was
not very revealing, while 18 f3 'ife7 19
'it'c2 ..ig6 20 fxg4 lLixg4 21 'it'e2 lLih6 22
..if2 lLifs 23 !Iael ..ih6 gave Black
enough play in A.Kovalyov-1.N ataf,
Montreal 2008.
 G l ig o ric Varia t i o n : 7 j_ e 3 0i g 4 8 j,_95 f6 9 j_ h 4 gs
18 'ife7
.••
After 18 ... j_d7 19 cs 'if e7 20 l:i:c1
0ifs 21 cxd6 cxd6 22 .l:f.c7 White won
material in Y.Shulman-1.Nataf, Mont­
real 2008. The alternative, 18 ... 0ig8,
was played by Radjabov, but he did not
repeat it when given the chance. After
19 f3 (perhaps 19 J:!.c1 OT 19 CS!? later
worried the Azeri) 19 ... gxf3 20 'it'xf3
0ih6 21 'it'e2 j_g6 22 j_f2 0ifs 23 �f1
the game was drawn in A.Onischuk­
T.Radjabov, Mallorca Olympiad 2004.
19 !k1 b6 20 b4 as!
Black creates some play for himself
on the queenside.
21 a3 axb4 22 axb4
22 j_g6
•••
Instead 22 .. Jfa3 23 cs bxcs 24 bxcs
j_g6 (24.. JHa8! ? is also possible) 2S
j_b1 .l:f.a1 26 'it'e2 �fa8 27 lUC3 j_xb1 28
�xb1 !Ixb1 29 !Ixb1 'if gs was Wang
Yue-T. Radjabov, Elista 2008. White still
seems a bit better here, but Black
gradually e qualized and drew.
23 cs bxcs 24 bxcs !Ias
Black eyes the dS-pawn.
2S 'it'e2
After 2S cxd6 cxd6 26 0ic3 j_xd3 27
'it'xd3 the g4-pawn is not under attack
and Black can play 27 ... 0ifs with coun­
terplay.
2s l:i:a4 26 0ic3 j_xd3 27 'it'xd3 l:i:b4
•••
Not 27 ... l:i:d4?? 28 'it'xd4.
28 cxd6 cxd6 29 !Ie4 l:i:xe4 30 0ixe4
£Ufs
At last Black activates the knight.
Now ... hs is coming and it becomes
clear that the g3-bishop is more of a
problem for White than Black's 'bad'
bishop on g7, which may become ac­
tive with ... j_h6. The game V.Akobian-
1.Nataf, Montreal 2008, concluded as
follows:
31 'it'd1 hs 32 h3 j_h6 33 !Ic6 h4 34
hxg4?
White had to give up a piece with
34 j_xh4 'it'xh4 3 S 'it'xg4.
34 0ixg3 3S fxg3 j_e3+ 36 @h2 'ifh7
•••
Black can also win with 36 ...hxg3+
37 @xg3 j_f4+ 38 @f3 'i'h4.
37 'it'c2
Instead 37 �c4 loses to 37 ... hxg3+
38 @xg3 j_f4+ 39 @f3 'it'h2.
37 hxg3+ 38 @xg3 j_g1! 0-1
•••
201
 Chapter 11
Petrosian Variation
7�d-S aS withOut 8 it.gs
1 d4 lLif6 2 c4 g6 3 lLic3 ..ig7 4 e4 d6 5
lLif3 o-o 6 ..ie2 es 7 ds
This variation was pioneered by the
ninth World Champion. Although it
appears to be a simple line, the Petro­
sian Variation leads to a real heavy­
weight strategic struggle. Both sides
often have the chance to play on either
side of the board. Although it is not
considered critical nowadays, it is still a
popular choice. The positions are gen­
erally not as sharp as they are in the
Mar Del Plata Variation, but the theory
does run deep in some lines. The Petro-
2 02
sian Variation was once a big favourite
of Kramnik and he used this system to
defeat Kasparov on more than one oc­
casion, so it certainly has to be taken
seriously. Personally I have been happy
to play the Petrosian with both colours.
1 as!
...
This prophylactic move is the main
line and it has been played more than
twice as often as all other moves com­
bined. A couple of Black's obvious
moves run into difficulties: for exam­
ple, 7 ... lLie8 8 h4! or 7... lLih s 8 g 3 ! fs 9
exfs ..ixfs (9 ... gxfs 10 lLixes!) 10 lLig s
lLif6 11 g4! and White has an unshake­
able grip on the e4-square.
The old main line was 7...lLibd7 8
..ig s (8 ..ie3 may be even better) 8 ... h6
g ..ih4 gs 1 0 ..ig3 lLih s 11 h4!. In this
line Black will have to depend com­
pletely on piece play, because a subse­
quent .. .fs will hand the e4-square to
White. A respectable line is 7 ... lLia6
(which was Gallagher's recommenda­
tion in his 2004 book) 8 lLid2! and now
 Petrosian Varia tio n : 7 dS a s with o u t 8 j,_9 5
8 ... c6 9 a 3 will sideline the a6-knight,
while Gallagher's 8 ... cs looks too stodgy
to me after 9 h4 h S 1o lLif3!.
With 7...as Black intends to play
... lLia6, from where it can hop to cs at
an appropriate moment to pressure
the e4-pawn. From a6 the knight will
also help to hold up both White's b2-b4
and c4-c5 advances without interfering
with Black's development (compared to
7...lLibd7). After White makes the nec­
essary preparations for the b4-advance
(commonly !Ib1, b3 and a3), Black will
often play ... lLics, in order to meet b4
with ... lLia4!. Of course the knight will
need the proper support to make this
leap, but we will see that the moves
...'it'e8 and ...j_d7 are often an integral
part of Black's plans.
After 7 ... as, White has a diverse
choice.
A: 8 lbd2
B: 8 .i.e3
C: 8 0-0
D: 8 h3
Lines A and B are seen occasionally
but are not very dangerous, while Line
C is a kind of transpositional device.
Line D was fashionable for a while, but
8 j_g5 is the absolute main line and
will be considered in Chapter 12.
There are a couple of other options:
a) 8 h4 lLia6 9 lLid2 is Line A.
b) 8 a3 j_d7! (8 ... lLia6 allows 9 b4
because of the pin on the a-file; this is a
common theme in the Petrosian) 9 j_e3
(or 9 !Ib1 a4! intending ...'it'e8 and
... lLia6) 9 ... lLig4 10 j_g5 f6 11 j_h4 lLia6
12 o-o 'it'e8 13 lLid2 lLih6 14 f3 lLif7 1 5
@h1 f S gave Black equal chances in
l.lvanov-J.Fedorowicz, USA 1989.
c) 8 g3 was another specialty of the
late GM Igor Ivanov that I have even
tried myself. My friend GM Jesse Kraai
called this 'a move from outer space'
when he saw it and thought chess
must be a great game if such moves
were playable. 8 g3 is not very danger­
ous theoretically, but Black should be
aware of White's ideas after 8 ...lLia6 9
lLih4!. This looks strange, but it is part
of White's plan, as we shall see.
203
A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1
Now we have:
c1) 9 ... lLie8 10 o-o fs 11 exfs gxfs 12
f4 e4 13 ..ie3 intending ..id4 and lLig2-
e3 was D.Vig orito-A.Matikozian, Las
Vegas 200S, when we can see White's
idea take shape, although even here
Black is not without resources.
c2) 9 ... lLics 10 f3 lLih s 11 o-o (it is
probably better to play 11 lLig2 ..ih3 12
0-o fs 13 exfs gxfs with an unclear po­
sition) 11 ...lLif4! (taking advantage of
the knight's position on h4) 12 lLig2
lLixe2+ (after 12 ... lLih3+ 13 @h1 fS 14
exfs gxfs 1s f4 the h3-knight's position
looks a bit precarious) 13 'it'xe2 ..ih3 (or
simply 13 .. .fs) 14 ..ie3 b6 1s @h1 fs 16
exfs gxfs 17 !Ig1 'it'd7 18 lLid1 !Iae8 19
lLif2 ..ixg2+ 20 !Ixg2 @h8 21 !Iag1 'it'f7
22 b3 lLid7 23 g4 and now, instead of
23 ... !Ig8?! 24 gxfs 'it'xfs 2 s lLie4 when
the g-file and e4-square were under
White's control in l.lvanov-B.Gelfand,
New York 1989, Black could play 23 ...f4!
24 ..id2 e4! 2s lLixe4 lLics with good
counterplay.
A) 8 lLid2
2 04
With the centre closed White hopes
to stir up trouble on the kingside.
8 lLia6 9 h4
...
Instead 9 o-o transposes to Line D,
while 9 g4 lLics 10 h4 is considered in
the notes to White's 10th move, below.
9...lLic s
T h e most dynamic. Black h a s no
reason to fear the advance of the White
h-pawn. 9 ... hs is also possible, but it is
a bit stodgy:
a) 10 lLif1 lLics 11 ..ig s 'it'e8 12 lLid2
..id7 13 f3 lLih7 14 ..ie3 fS gave Black
counterplay in Y.Teplitsky-A.lstratescu,
Yerevan Olympiad 1996; White does
not have an ideal home for his king.
b) 10 f3 c6 (10 .....ih6 is given as
equal by Nunn and Burgess) 11 lLif1
lLics 12 ..ie3 and now rather than
12 ... cxds 13 cxds, which hands the c4-
square to White, Black should play
12 ... ..id7, intending ... a4, ...'it'as, ... cxds
and ... bs.
c) 10 lLif3 ! ? heads for the weakened
gs-square: 10...lLics 11 lLig s c6 12 ..ie3
..id7 and a complicated strategic
struggle lies ahead.
10 h s
Instead 10 g4 is somewhat unnec­
essary, but it has been seen often
enough in practice. Following 10 ... a4
(10... c6 is also possible) 11 hS gxh S ! ? 12
gs (after 12 gxhs ..ih6! intending
... @h8 the opening of the kingside is
welcomed by Black) 12 ...lLig4 13 lLif1 fS
we have:
a) 14 exfs ..ixfs 1s f3 a3! gives Black
good play. One possibility is 16 fxg4
 Petro sian Va riatio n : 7 dS as with o u t 8 j,_95
..i.xg4 17 j_xg4 hxg4 18 'it'xg4 axb2 19
..i.xb2 lLid3+ 20 @e2 lLixb2 21 'it'hs l:i:f4
22 'it'xh7+ @f8 with a messy position
that looks very playable for Black.
b) 14 f3 lLif2 (Black could also play
14.. .fxe4 1 S fxg4 j_xg4 16 j_xg4 lLid3+
17 @d2 'it'xg s+ 18 @c2 lLib4+ 19 @b1
"ihg4 with an unclear position) 1 S
ltixf2 fxe4 1 6 @g2 a3 ! 17 !Ixhs exf3+ 18
..i.xf3 e4 19 j_e2 j_fS gave Black excel­
lent compensation for the piece in the
famous game B.Kouatly-G.Kasparov,
Evry (simul) 1989.
10 c6
...
Black looks for queenside counter­
play. This move was relatively untested
when Nunn and Burgess wrote in the
1990s, but now there are many exam­
ples and Black has scored well.
11 g4
Others:
a) 11 h6 limits White's possibilities.
After 11...j_h8 12 f3 lLihs 13 lLif1 lLif4
14 g3 lLixe2 1 S 'it'xe2 fS Black had the
initiative in A.Bykhovsky-1.Caspi,
Herzliya 2006.
b) 11 'it'c2 (it looks odd to place the
queen on the c-file) 11 ... a4 12 lLif1 cxds
13 cxds j_d7 14 hxg6 fxg6 (14...hxg6 is
also possible - Black is not about to get
mated on the h-file) 1s f3 lLihs 16 g3
bS! gave Black good play in
A.Dunnington-Z.Lanka, Paris 1990. If
White takes on bS then f3 will be hang­
ing.
11 cxds
...
Often is it better to delay this ex­
change because White gains access to
the c4- and bS-squares, but here there
is a neat tactical point. Black has also
avoided the exchange in practice with:
a) 11 ... 'it'b6 is a common idea. The
queen eyes the b2- and f2-squares,
thereby creating some tactical possi­
bilities: 12 f3 (not 12 gs? lLixhs 13
j_xhs lLid3+ 14 @e2 lLif4+) 12 ...j_d7 13
lLif1 cxds 14 cxds !Ifc8 1s lLie3 a4 16
lLic4 'ife7 17 lLia3 ! was A.Aleksandrov­
R.Leitao, Yerevan Olympiad 1996.
White's knight manoeuvring has been
very time-consuming, but he has man­
aged to grip the bS-square. Instead of
the game's speculative 17 ... bs, I would
suggest 17 ... 'it'b6 when Black looks
okay.
b) 11 ...a4 is logical and then:
bl) 12 g s should backfire: 12 ...lLixhs
13 j_xhs lLid3+ 14 @f1 'it'b6 1 s !Ih2 and
now, instead of 1s ...gxhs?! 16 'it'xhs
h6? 17 gxh6 j_f6 18 lLif3 with a win­
ning attack in A.Aleksandrov­
V.Kotronias, Pula 1997. Black should
play Huzman's 1s ... a3 ! 16 b3 'it'd4! 17
lLidb1 lLixc1 18 'it'xc1 gxhs when he has
a huge advantage.
205
A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1
b2) 12 f3 ..id7 13 lLidb1! 'it'b6 14
'it'd2!? (White makes some unusual
moves, but they are all designed to
shut down Black's counterpl ay)
14 ... !Ifc8 1s lLia3 'it'as 16 @f1 l:!.ab8 17
'it'e1 'it'd8 18 @g2 cxds 19 cxds lLie8 20
lLiabS and White was better in
C.Gabriel-E.Grivas, Corfu 1999.
12 cxds
Instead 12 exds e4! gives Black ob­
vious counterplay:
a) 13 g s lLixh s 14 ..ixh s gxh s 1s
lLicxe4 l:i:e8 16 'it'xh s .ifs was clearly
better for Black in D.Shchukin­
E.lnarkiev, St Petersburg 2001.
b) 13 h6 ..ih8 14 gs lLih s 1s lLidxe4
lLixe4 16 lLixe4 l:i:e8 17 l:i:h4 .ifs 18 ..if3
..ixe4 19 !Ixe4 l:i:xe4+ 20 ..ixe4 'if e7 21
f3 !Ie8 22 'it'd3 lLig3 won back the pawn
and left Black with some initiative in
A.Voll-D.Petrovic, correspondence 2004.
12 bs!
...
Black takes advantage of the loose
pawns on e4 and g4. Instead 12 .....ixg4
does not work, but the idea is worth
keeping in mind, despite here 13 ..ixg4
lLid3+ 14 @fl 'it'b6 1S l:i:h2 lLixb2 16
206
'it'b 3 ! being good for White. 1 2 ... a4 and
12 ... ..id7 are possible, but if Black
wants to play more slowly it is better to
delay the exchange on dS.
13 gs
Altematively:
a) 13 f3 b4 14 lLia4 lLifd7 is comfort­
able for Black. Even ... ..if6-gS is a possi­
bility.
b) 13 hxg6 fxg6 14 f3 b4 1S lLia4
lLifd7 is similarly pleasant for Black.
c) 13 a3 and here:
cl) 13 ...'it'b6 14 f3 ..id7 1S lLib3 b4
16 lLixcs 'it'xcs (instead 16...bxc3 17
lLixd7 lLixd7 18 bxc3 fS looked too
speculative in l.Tsesarsky-M.Tseitlin,
Givataim 2000, although Black won) 17
lLia4 'ifa7 looks fairly level.
c2) 13 .....ixg4! now works.
After 14 ..ixg4 b4! both 1S lLie2?
lLixg4 and 1S lLia4? lLixa4 hang pieces,
which leaves:
C21) 1S l2Jcb1 l2Jd3+ 16 @fl jVb6 17
!Ih2 lLixg4 18 jVxg4 lLixc1 wins.
c22) 1S axb4 axb4 16 !Ixa8 'it'xa8 17
lLibs lLid3+ 18 @f1 lLixg4 19 jVxg4 'it'b8
and both the knight on bS and bishop
 Petrosian Va ria tion : 7 ds a s with o u t 8 j,_95
on cl are hanging.
c23) 15 lLia2 lLid3+ 16 ..tf1 bxa3 17
bxa3 jVb6 18 !Ih2 lLixg4 19 'it'xg4 'it'd4
20 lLib3 'ifc4 21 lLid2 'it'c2 22 'it'e2 lLif4
intending ....l:f.ab8 looks like fun for
Black.
13...lLixhs
There is no reason to play the pas­
sive 13 ...lLie8. Black must exploit his
tactical possibilities.
14 j_xhs
14... b4! 1s lLia4
Avoiding the challenge is worse. Af­
ter 15 lLib3 bxc3 16 lLixc5 cxb2 17 j_xb2
dxc5 18 j_e2 'it'xg5 White's position
was a mess in B.Kohlweyer-D.Stets,
German League 2005.
1s...lLid3+
15 ...lLixa4 16 'it'xa4 gxh 5 is probably
even better. After 17 !Ixh5 j_g4 18 !Ih4
j_d7 Black will play ...'it'xg5 with a clear
advantage.
16 @fl lLif4!
Now 17 j_f3? j_a6+ 18 @g1 'it'xg5+
19 @h2 was R.Kozlov-P.Gnusarev, As­
tana 2007, when 19 ...j_e2 wins in­
stantly. Better is 17 j_e2, although after
17 ... 'it'xg 5 18 @e1 'it'g2 19 .l:f.f1 j_h3 20
j_f3 'it'g 5 Black has a strong initiative
for the sacrificed material.
As in the Gligoric Variation, White
allows this bishop to get chased around
a bit.
8 ...lLig4
Worse is 8 ... lLia6 9 lLid2 when White
has the ideal piece set-up and total
flexibility with his king position.
9 j_g5 f6 10 j_h4
Instead 10 j_d2 is not very danger­
ous after 10...f5 (however, 10...lLia6 11
h3 lLih6 12 h4 is a bit annoying) and
now:
a) 11 o-o lLia6 (or 11 ... lLif6) 12 lLie1
lLif6 13 exf5 j_xf5!? was V.Gallego
Jimenez-J.Gallagher, Benidorm 1991.
Taking with the bishop is often a de­
cent choice when Black has good con­
trol of the e4-square, but 13 ...gxf5 is
also fine.
b) 11 exf5 gxf5 (again 11...j_xf5 is
fine, as in A.Huss-J.Gallagher, Samnaun
2008) 12 lLig5 lLif6 is variation 'c' below.
207
A ttacking C h e s s : Th e King 's I n di a n , Vo l u m e 1
Instead 12 ... 'it'e8 13 cs!? dxcs 14 'ifb3
@h8 1S 0-0 gave White compensation
in B.Zlotnik-A.Morozevich, Moscow
1991, although he soon went wrong
and lost.
c) 11 lLigs lLif6 12 exfs gxfs (this
time 12 ... ..ixfs 13 g4! ..id7 14 lLige4 is
to White's advantage) 13 f4 e4 is fairly
level.
After 10 ..ih4, White wants to play
lLid2, f3 and ..if2 with a flexible and
harmonious position. In most cases
Bl ack should maintain his knight on g4
until forced to retreat. White generally
cannot play f3 without allowing ... lLig4-
e3, so Black usually waits for White to
play h3, which will cost a tempo and
somewhat weaken the white kingside.
10 lLia6
...
10 ... h s 11 lLid2 lLia6 transposes.
Another idea is 10 ...lLih6. The point
is to play ...gs and .. .fs before White
gets in f3 and ..if2. After 11 lLid2 gs 12
..ig3 lLia6 (12 .. .fs !?) 13 a3 ! ? (or 13 f3 fS
when Black gets counterplay following
14 ..if2 g4 or 14 exfs lLixfs 1s ..if2 lLid4)
we have:
208
a) 1 3 ... ..id7 1 4 o-o fS (Black should
consider 14 ... g4 or 14... 'it'e8) 1S exfs
lLixfs 16 lLide4 looks somewhat better
for White.
b) 13 ... 'it'e8 14 f3 (14 b4 fS 1S f3
axb4 16 axb4 ..id7 17 l:i:b1 g4 gives
Black counterplay according to Avrukh)
14... ..id7 1S ..if2 and now, instead of
1S ... a4? 16 lLibS! which was very good
for White in E.Bacrot-E.Relange, French
Team Championship 200S, Avrukh
suggests 1s .. .fs 16 exfs lLixfs 17 o-o
lLid4 with sufficient play.
11 lLid2
No one has tried 11 h3 lLih6. Pre­
sumably Black gets a favourable ver­
sion of the note to his next move. If
Black really does though prefer to play
with a pawn on hS, he could avoid this
possibility altogether by playing 10 ... hS.
11 hs
...
Instead 11...lLih6!? allows White to
reach his desired set-up with 12 f3, but
Black still looks fine following 12 ... lLif7
(or 12 ... ..id7 13 o-o 'it'e8) 13 a3 ..id7 14
l:i:b1 lLics 15 b4 axb4 16 axb4 lLia4.
Black is also not bothered by 12 a3 ..id7
 Petrosian Varia tion: 7 ds as witho u t 8 ii. g s
1 3 !Ib1 lLics (13 ... 'it'e8 14 b4 sidelines
the a6-knight) 14 b4 axb4 1S axb4 l2Ja4
with the idea of ... gs and ... fs.
12 a3
The immediate 12 h3 lLih6 has
scored well for Black: 13 g4 (instead 13
a3 ii..d7 transposes to the main line,
while after 13 f3 lLif7 Black will play
... ii..h 6) 13 ... hxg4 14 hxg4 lLif7 with the
idea ...ii..h 6, ... @g7 and ...!Ih8, which
gives Black good play.
12 ...ii..d 7
13 h3
White finally tires of having Black's
knight hanging around. This move i s
generally criticized, but the alterna­
tives are also unimpressive:
a) 13 0-0 and now 13 ...lLih6 is com­
pliant, wh ile 13 ...lLics allows 14 b4, so
we have:
al) 13 ...'it'e8 14 b3 fs 1S l:i:b1 lLics 16
b4 axb4 17 axb4 lLia4 and here White
should avoid 18 'it'c2 lLixc3 19 'it'xc3
fxe4 20 lLixe4, since Black has
20 ... lLixh2! when 21 @xh2 fails to
21 ...J:i.f4 and 21 lLixd6 cxd6 22 @xh2
.l:f.f4 gives Black the initiative, as ...e4 is
in the air. Better is 18 lLixa4 ii..xa4 with
equality in J.Speelman-A.Strikovic,
Oviedo (rapid) 1992.
a2) 13 ...'it'e7 14 .l:f.b1 lLics 1S b4 axb4
16 axb4 l2Ja4 17 lLixa4 !Ixa4 18 h3 lLih6
19 f3 was J.Speelman-J.Piket, Tilburg
1992, when Black should just play the
obvious 19 ....rf.fa8 with equality, as in­
dicated by Speelman.
b) 13 .l:f.b1 lLics and here:
bl) 14 b3 g S ! 1S ii..g 3 fS 16 h3 lLif6
gives Black the initiative after either 17
ii..xh s lLid3+ 18 @f1 lLif4 (or 18...lLixhs
19 'it'xh s lLif4), or 17 f3 'it'e8 18 'it'c2
fxe4 19 fxe4 'it'g6, V.Loginov-A.Fedorov,
St Petersburg 1997.
b2) 14 b4 axb4 1S axb4 lLia4 16
lLixa4 (after 16 'it'c2 Black again seizes
th e initiative with 16... gs 17 ii.. g3 fS)
16...ii.. xa4 17 'it'c1 ii..d7 and Black had a
good position in G.Danner-V.Spasov,
Leon 2001.
13 ...lLih6 14 .l:f.b1 lLics 15 b4 axb4 16
axb4 lLia4
17 'it'c2
Black also had fine play after 17 lLixa4
ii..xa4 18 'it'c1 gs 19 ii..g3 h4 20 ii..h 2 ii..d7
209
A ttacking C he s s : Th e Kin g 's In dian, Vo l u m e 1
21 f3 f5 22 ..ig1 g4 in l.Morovic Feman­
dez-G.Kamsky, Las Palmas 1994.
17 ...lLixc3 18 'it'xc3 gs 19 ..ig3 h4! 20
..ih2 fs 21 cs g4 22 c6 ..ic8
Black keeps the tension. Kasparov
gives 22 ... bxc6 23 dxc6 ..ie6 24 ..ic4 as
unclear.
23 hxg4 fxg4
23 ...lLixg4!? looks good too.
24 cxb7 ..ixb7 2s o-o 'it'gs
Black had the initiative in E.Bareev­
G.Kasparov, Tilburg 1991.
C) 8 0-0
This move is played rarely nowa­
days, but it is still important, especially
for transpositional reasons.
8...lLia6 9 'it'c2
We will take this as the main line,
but there are several other moves. 9
..ig 5 is actually the most common
move, and after 9 ... h6 10 ..ih4 we are in
Chapter 12. Instead after 9 ..ie3 lLig4 10
..ig5 f6 White has a less flexible version
of Line B. White's two remaining alter­
natives to 9 'it'c2 are the typical knight
retreats:
210
a) 9 lLie1 lLic5 and here:
al) 10 'it'c2 transposes to note 'a' to
White's 10th move in our main line,
although this way White has managed
to avoid 9 'if c2 lLih5.
a2) 10 f3 has scored terribly for
White after 10 ...lLih5 ! . Indeed, Black has
a pleasant choice after both 11 ..ie3
(between 11 ...b6, 11 ...lLif4 and 11 ... f5),
and 11 lLid3 (between 11 ...b6 and
11 ...lLixd3 12 ..ixd3 lLif4), so White usu­
ally prevents the knight hop with 11 g3.
Then 11.. ...ih3 12 !If2 f5 (or 12. . ...id7)
13 lLid3 lLixd3 14 'it'xd3 (14 ..ixd3 f4 15
g4 lLif6 is similar) 14.. .f4 15 g4 lLif6
(15 ... lLig3 is possible but more risky af­
ter the cool 16 ..id1) 16 ..if1 ..ixf1 17
'it'xf1 lLid7, with the idea .....if6, gives
Black good play.
b) 9 lLid2 lLic5
and here:
bl) 10 'it'c2 transposes to note 'b' to
White's 10th move in our main line,
again with White avoiding 9 'it'c2 lLih 5 .
b2) 10 f3 lLih5 11 lLib3 b6 has done
well for Black.
b3) 10 !Ib1 will transpose to varia-
 Petrosian Va ria t i o n : 7 ds as with o u t 8 j,_95
tion 'b4' after 10...lLie8 11 b3, while
10 ... j_h 6 is also sensible.
b4) 10 b3 lLie8 (10 ... j_h6 has also
scored well for Black after 11 'it'c2 lLie8
or 11 ...j_d7) 11 a3 fS 12 l:i:b1 lLif6 and
now:
b41) 13 b4 is a strange pawn sacri­
fice. After 13 ... axb4 14 axb4 lLicxe4 1S
�dxe4 lLixe4 16 lLixe4 fxe4 17 j_e3
Black is doing well after all of 17 ....rf.a2,
17 ...j_d7 and 17 ... 'it'M. The extra pawn
may not mean much itself, but Black
certainly has no problems.
b42) 13 f3 j_d7 (preparing for
White's queenside advance) 14 b4 axb4
1S axb4 lLia4 (we will see this idea
countless times) 16 'it'c2 lLixc3 (Black
could also play 16 ...j_h6 or 16 ... lLih s) 17
"if'xc3 and now Black has several prom­
ising ideas, such as 17 ... lLihs , 17 ... j_h6
and 17 .. .fxe4 18 lLixe4 lLixe4 19 fxe4
�xf1+ 20 j_xf1 'ifh4.
Now we return to 9 'ifc2:
9 lLics
...
After this move we reach a position
that has occurred hundreds of times,
although it usually arises after 7 o-o
lLia6 (or 7... lLibd7) 8 dS lLics 9 'it'c2 as.
With the move order we are going by,
however, Black has another promising
option in 9...lLih s.
White has not been able to show
anything here. On the contrary, Black
has done very well. Some examples:
a) 10 g3 fs 11 exfs j_xfs 12 'it'd1 (or
12 lLie4 lLif6 13 lLifd2 lLib4 14 'it'b1 c6)
12 ... lLif6 13 lLih4 j_h3 14 lLig2 'it'd7 gave
Black some initiative in S.Hamann­
R.Hiibner, Skopje Olympiad 1972.
b) 10 !Ie1 lLif4 11 j_f1 is a bit trick­
ier. After 11 .. .fs 12 j_xf4 exf4 13 es
White has good chances, so Black
should play 11 ...j_g4!.
c) 10 lLie1 lLif4 11 j_f3 fS 12 j_e3 b6
13 exfs j_xfs 14 j_e4 lLics 1s j_xfs !Ixfs
16 g3 lLih 3+ 17 @g2 'it'd7 18 lLid3?!
(better is 1 8 j_xcs bxcs 19 lLie4, al­
though Black is fine after 19 ... l:i:af8 or
19 ... a4) 18 ... lLixd3 19 'it'xd3 lLif4+! 20
gxf4 exf4 gave Black a strong attack in
J.Nikolac-G.Ligterink, Wijk aan Zee
1976.
Even though 9 ...lLih S looks promis­
ing, we will also examine 9 ...lLics in
211
A ttacking C he s s : The Kin g 's I n d i a n , Vo l u m e 1
detail because the position can arise in
various ways and doing so provides
Black with a couple of good options.
10 ..tgs
After 10 b3 lLihs Black is even better
off than after 9 ...lLihs above. 10 ..ie3
gives Black a choice between the solid
10...lLifd7 or 10...b6 and the more en­
terprising 10...lLig4 11 ..ixcs dxcs 12 h3
when he has the extra option of
12 ... lLih6! ? compared to 11 ... lLig4 in the
note to his next move.
Again there are also the knight re­
treats:
a) 10 lLie1 lLifd7 (this retreat keeps
control of both cs and es, but Black has
decent options too in 10...b6, 10.....id7
and 10 ... lLie8) 11 ..ie3 fS 12 f3 (after 12
exfs gxfs 13 f4 exf4! 14 ..ixf4 lLies
Black has good piece play) 12 .. .f4 13
..if2 gs and Black is doing rather well.
White still has to prepare b4 and Black
should get in ... g4 rather easily -
White's queen being on c2 (as opposed
to d1, from where it controls g4) helps
Black.
b) 10 lLid2 gives Black a pleasant
212
choice between 10...lLie8 11 lLib3 fS (or
11 ...b6) 12 lLixcs dxcs! 13 ..ie3 b6, in­
tending ...lLid6, and 10.....ih6 11 lLib3
..ixc1 12 l:i:axc1 lLifd7 (12 ... lLixb3 and
12 ... b6 are satisfactory as well) when
Black has no problems.
10. h6 11 ..ie3 b6
..
This so lid move has gradually
evolved to become the main line. Black
has also tried 11 ... lLifd7, 11 ...lLihs and
11 ... 'ife7, while after the forcing
11 ... lLig4 12 ..ixcs dxcs 13 h3 lLif6 14
lLixes lLixds 1s cxds ..ixes 16 f4 ..id4+
17 @h2 White has a big centre, so I pre­
fer the text, which keeps more tension
in the position.
12 lLid2
12.....ig4!?
This unusual-looking move has be­
come popular. It looks rather anti­
positional to offer a trade of light­
squared bishops, but Black hopes that
White must make some concession to
make this exchange. The more natural
12 ... lLig4 is also viable. After 13 ..ix94
..ixg4 14 a3 (14 f3 ..id7 planning . ts ..
gives Black counterplay, so White tries
 Petros ian Va riation: 7 ds as with o u t 8 il.. g s
to ignore the bishop) 14 ..lLia6 15 !Iab1
fS ! (1s ... i.d7 16 b4 axb4 17 axb4 fS has
also been played, but the text is
sharper) 16 f3 (16 exfs gxfs 17 f4 e4 is
unclear) 16 ... i.h s 17 b4 axb4 18 axb4
f4 19 i.f2 g s Black has counterplay. The
break ... g4 is coming and if White plays
h3, Black will retreat the hS-bishop and
play ...h s and ... g4 with even greater
effect.
13 f3
This natural move gains a tempo,
but Black has a reason for provoking
this advance. The natural 13 i.xg4
lLixg4 14 i.xcs bxcs is fine for Black
despite the exchange of light-squared
bishops. It is not easy to come up with
an active plan for White and Black will
play .. .fs or ...h s , activating the bishop
on h6. Instead after 13 'it'd1 Black
should just retreat with 13 ... i.d7, con­
tent th at White has undeveloped his
queen. Others:
a) 13 h3 i.d7 (happy to have in­
duced a kingside weakness, Black re­
treats, but 13 ... i.xe2 14 lLixe2 lLih s has
al so been tried) 14 b3 lLih7 15 a3 fS 16
exfs (or 16 b4 lLia6 and 16 f3 f4 looks
good too) 16 ... gxfs 17 f4 exf4 18 i.xf4
'ifh4! 19 !Iae1 lLigs 20 @h 2 l:i:ae8 21
'it'c1 i.es 22 i.xes l:i:xes 23 b4 axb4 24
axb4 f4! gave Black a winning attack in
Y.Zilberman-1.Smirin, European Club
Cup, Panormo 2001. White cannot
avoid detonation on h3 - a far reaching
consequence of his 13th move!
b) 13 !Iae1 is more challenging.
Then 13 ...i.xe2 (Black should consider
the untried 13 ... i.d7 ! ? when it is not
clear what White's rook is doing on el)
14 l:i:xe2 lLig4 15 i.xcs bxcs and we
have:
bl) 16 lLibs hs 17 a4 i.h6 18 'ifc3 fs
19 f3 lLif6 20 lLib3 ?! (winning the as­
pawn at great cost; instead 20 exfs
gxfs 21 lLib3 h4 22 lLixas lLih s with
counterplay is given by Panczyk and
llczuk) 20 .. .fxe4 21 fxe4 lLig4 with a
strong initiative for Black in A.Korobov­
N.Kuren kov, Moscow 2007. White's
knights are too far away from the ac­
tion.
b2) 16 g3 looks silly, but the follow­
ing play shows the strategic complexity
inherent in the Petrosian Variation.
With his rook on e2, White wants to
play on the kingside him self: 16 ...h s 17
f3 lLih6 (or 17 ...lLif6 18 f4) 18 f4 fS 19
lLif3 and White had the initiative in
A.Korobov-P.Czarnota, Cappelle la
Grande 2004.
13 i.d1 14 b3 lLihs!
...
This is precisely why Black provoked
f2-f3.
15 .l:f.fel
213
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1

For the pawn Black has accelerated

his kingside play.

1s ... ..tf6!
Both 1s ...t2lf4 16 ..if1 and 1s .. .fs 16
a3 look too slow, so Black offers a pawn 19 exfs gxfs 20 a3 a4 21 b4 lLib3
to activate his dormant bishop. Black had excellent play in
16 ..ixh6 l.Naumkin-1.Smirin, lschia 199S.
After 16 a3 ..igs White cannot play
17 ..if2? because of 17 .....ixd2 18 'it'xd2 D) 8 h3
l2lxb3. With his bishop now activated,
Black also has good play after 16 !Iab1
..tgs 17 ..tf2 fs.
16 ..tgs! 11 ..ixgs
...

White cannot take the rook. He gets

mated after 17 ..ixf8 ..ie3+ 18 @h1 (or
18 ..tf1 'ifh4 19 t2ld1 l2lg3+) 18 ... l2lg3+
19 hxg3 'it'xf8 20 l2lf1 'it'h6+ 21 l2lh2
@g7.
11 'it'xgs 18 tlJfl
...

A quick finish was seen after 18 ..if1

fS 19 a3 !If7 20 !Iab1 t2lf4 21 b4? (bet­
ter is 21 g3 !Ih7, although Black has
reasonable compensation for the
pawn) 21 ...l2lcd3 and 0-1 in R.Wukits­
G.Timoscenko, Oberwart 2003. White is
losing the exchange because 22 !Ied1
walks into 22 ...l2lh3+ 23 @h1 l2ldf2
mate.
1s fs
...
This is a tricky move that has been
favoured by Zvjaginsev. 8 h3 in the Pet­
rosian Variation is very similar to the
Makogonov Variation (S l2lf3 0-0 6 h3)
and there are several transpositional
possibilities: for example, 1 d4 l2lf6 2 c4
g6 3 t2lc3 ..ig7 4 e4 d6 s l2lf3 0-0 6 h3 es
7 dS as 8 ..igS l2la6 9 lLid2 (or 9 ..ie2)
9 ...'it'e8 10 ..ie2 transposes to the Petro-

214
 Petros ian Varia ti o n : 7 dS as with o u t 8 j,_95
sian. Here we will look at lines where
White has played an early j_e2, while
in the Makogonov Variation (which will
be covered in Volume 2), we will look at
positions where White delays or omits
th is move.
There are a couple of ideas behind
the subtle 8 h3. White is ready to play
the sequence 9 j_g5 h6 10 j_e3 because
now Black cannot play ...l2lg4. White
then hopes that ... h6 will weaken the
black kingside - the h6-pawn itself may
be weak, but more importantly the di­
agonal b1-h7 will be weakened after a
subsequent .. .fs. This may seem minor,
but in practice this detail can create big
problems for Black. Another idea for
White is to play g2-g4. Then if Black
plays .. .fs, White can pl ay gxfs, opening
the g-file, and exchanging twice on fS
could grant White the e4-square for his
knights. If Black does not play ... fs then
White may tum his attention to the
other wing.
The downside to 8 h3 is that it
weakens White's kingside, so castling
on that side of the board becomes less
appealing for the first player. Also the
g4-advance creates weaknesses and
Black's .. .fs break can be very danger­
ous for White if he is unable to keep
the position under control.
Generally Black has two different
ways to deploy his pieces. He can play
... tLJcs, ... a4, ... c6, ... j_d7 and ... 'it'as to
create play on the queenside. However,
Black should be careful concerning the
move ... cxds. If it is played too early,
White will seize control of c4 and bS
and it will be difficult for Black to cre­
ate counterplay, so Black should gen er­
ally delay this exchange until he can
successfully control the bS-square. lfhe
can do this, then the advance ...bs will
not only control c4, but will give him
the possibility of playing ...b4.
Black's other set-up is seen in the
main lines of both Lines D1 and D2,
below. Black will play ... 'it'e8, ...t2ld7 and
... tlJdcs. This will usually allow him to
get in ...fs because the diagonal for the
c8-bishop remains open. It is in these
lines that we see the importance of
leaving out the weakening move ... h6.
Because the gs-bishop is hanging in
the air, Black may drive it away first
with .. .f6 before playing .. .fs, or play .. .fs
straightaway. Having both knights on
the queenside may also give Black
some tactical possibilities. The moves
... j_d7 and ... a4 are also thematic (espe­
cially if Black does not play .. .fs), but in
general I think th at Black should avoid
playing ... c6 in this set-up.
8 l2la6
...
Instead 8 ...tlJh S is not a bad idea,
but White can play either 9 g3 (as he
would after 7 ...tlJh s), or the exotic 9
l2lg1 ! ? intending 10 j_f3.
An interesting idea is 8 ...l2lfd7!?. In­
deed, Panczyk and llczuk give this as
their main line even though it has only
been played a handful of times. Now 9
g4 tLJcs 10 j_e3 l2lba6 intending .. .fs
looks pretty comfortable for Black. In­
stead g h4! l2lf6 (9 ... fs 10 h S ! ) 10 t2ld2 ! ?
215
A ttacking C he s s : Th e King 's In dian, Vo l u m e 1
leaves White a tempo up on 8 lLid2 lLia6
9 h4, although this is not exactly fatal
in this kind of position.
9 ..tgs
This pin is a typical weapon in both
the Petrosian and Makogonov varia­
tions. White tries to entice Black to
weaken his kingside. Other moves are
less dangerous:
a) 9 ..ie3 lLics (9 ...lLih s is also possi­
ble) 10 lLid2 lLie8 11 g4 fs 12 gxfs gxfs
13 exfs (13 'it'c2 lLia6! threatens ... f4
and then 14 exfs .�xfs 1s .�d3? is not
possible because of 1s ...lLib4) 13 ... ..ixfs
gives Black excellent play. Note that
Black's h-pawn would be much more of
a target on h6 here, which is one point
of White's 9 ..igs.
b) 9 g4 is not well timed: 9...lLics 10
'it'c2 c6 11 ..ie3 a4 (better than 11 ...cxds
12 cxds when bS is weak) 12 o-o-o
(White cannot even take a pawn be­
cause 12 ..ixcs dxcs 13 lLixa4 runs into
13 ...lLixe4!) 12 ... cxds 13 cxds 'ifas 14
lLid2 ..id7 1S gs (this runs into a great
shot, but after 1S lLic4 'ife7 Black threat­
ens ... bs and 16 lLia3 can be met by
216
16...lLicxe4 17 lLixe4 'it'xc2+ and
18 ... lLixe4 with an extra pawn)
1S ... lLifxe4! 16 lLidxe4 (or 16 lLicxe4
lLixe4 17 lLic4 'i't'b4 18 'it'xe4 !Ifc8 19
..id2 'it'cs 20 ..ie3 and here, instead of
repeating moves with 20 ... 'i't'b4, Black
can play 20 ... 'it'bs! with the idea of
... .ifs) 16 ...lLixe4 17 'it'xe4 (17 lLixe4 loses
to 17 ...!Ifc8 18 lLic3 a3!) 17 .....tfs 18 'i'h4
!Ifc8 19 @d2 !IxC3! 20 bxc3 'it'xds+ and
Black had a winning attack in L.Yurtaev­
Y. Shulman, Vladivostok 199S.
9 'it'e8
...
Instead 9 ... h6 10 ..ie3 shows one
idea behind White's 8th move (there is
no ...lLig4), and Black has weakened his
kingside in order to break the pin. I pre­
fer the text move.
The move ... 'it'e8 is very typical in
the Petrosian Variation. Black breaks
the pin without weakening his kingside
and observes the a4-e8 diagonal. Here
he hopes to show that h3 cost a tempo
and weakened the white position,
while the gs-bishop may also be mis­
placed. White has two important con­
tinuations:
 Petros i a n Variati o n : 7 dS as with o u t 8 j,_ 9 5
01: 10 g4
02: 10 �d2
Dl) 10 g4
This lunge can be the prelude to an
attempted kingside attack, but often
White is just hoping to limit Black's
possibilities.
10 lLid7!
•••
Black can also play 10...lLics, but af­
ter 11 lLid2 but the f6-knight is in the
way. By playing the king's knight to cs,
Black keeps the possibility of playing
...fs and ... j_xfs. Black has also tried to
move the knight to g8, but not surpris­
ingly, this is less dynamic. Indeed,
10 ...@h8 11 lLid2 (or 11 l:!.g1 lLig8)
11 ... lLig8 12 h4 f6 13 j_e3 fs 14 f3 (after
14 gxfs gxfs 15 exfs j_xfs 16 lLide4 lLif6
Black had decent play in T.Polak­
S.Novikov, Pardubice 2007) 14 ...f4 15
j_f2 is better for White, because all the
play will shift to the queenside, where
White has a significant space advan­
tage.
11 !Igl
Others:
a) 11 lLid2 lLidcS 12 lLif1 (in stead 12
h4 could be met by 12 .. .fs or 12 ...f6, 12
a3 should be met by 12 ...j_d7, and 12
!Ig1 transposes to the notes to White's
12th move in our main line, below) is
the same as variation 'a1' in the notes
to White's 12th move below, but with­
out l:i:g1 and ... @h8. Here the position
looks even better for Black after 12 .. .f6
OT 12 .. .fS.
b) 11 'it'd2 lLidcS 12 l:i:g 1 lLib4! cre­
ates tactical problems for White. Black
intends ... @h8 and .. .fs, and White can­
not play a3 because of the fork on b3.
c) 11 a3 and now 11 ...lLidcS? is met
by 13 b4 because of the pin along the
a-file, but Black has 11 ...lLib6 ! .
This i s an unusual square for the
black knight, but White has made a lot
of pawn moves and so Black adapts to
the situation. He plans .. .fs or ...j_d7
and ... lLia4. After 12 lLid2 j_d7 13 h4
practice has seen:
c1) 13 ... lLics 14 hs fs (14 ... lLiba4!?)
15 hxg6?! (better is 15 gxfs gxfs 16
j_e3) 1s ...'it'xg6 16 j_e7 fxe4 17 j_xfg
217
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
!Ixf8 gave Black compensation in
B.Thorfinnsson-E.Berg, Reykjavik 2009,
but 16 .. .fxg4! 17 ..ixf8 !Ixf8 would have
been even stronger.
c2) 13 ... lLia4! 14 lLibs lLi4cS 1s hs fs
16 gxfs (not 16 hxg6 'it'xg6, hitting the
bishop) was agreed drawn here in
L.Nisipeanu-V.Nevednichy, Sovata
1999. After 16 ... gxfs 17 l:i:g1 @h8 Black
has good counterplay.
11 lLidcs
...
11 ... @h8 12 'it'd2 lLidcS (J.Alvarez
gives 13 ..ih6 ..ixh6 14 'it'xh6 f6 1S h4
as better for White, but I prefer Black
after 1S ... ..id7) 13 o-o-o ..id7 (or 13 ... a4)
is a good alternative.
12 h4
Instead 12 'it'd2 lLib4! is note 'b' to
White's 11th, above. White's main al­
ternative is 12 lLid2 @h8 and here:
a) 13 lLif1?! is slow and a bit clumsy
after 13 .. .fs 14 gxfs gxfs 1s lLig3 fxe4!
and now:
al) 16 lLigxe4 .ifs 17 ..ihs lLid3+!
(even better than 17...'it'd7) 18 @f1 (al­
ternatives are worse: 18 'it'xd3 'it'xhs or
18 @d2 lLixb2, with a clear advantage
218
to Black in both cases) 1 8...lLixb2 19
'it'e2 'it'd7 is good for Black. The knight
can escape with ...lLia4 and 20 'it'xb2 is
met by 20.....ixe4 21 lLixe4 'it'xh3+, win­
ning.
a2) 16 lLihs with a further divide:
a21) 16 ... 'iff7 is tricky: 17 ..ih4 ..ih6!
(Black should avoid the complications
of 17 ... l:i:g8 18 !Ixg7 !Ixg7 19 ..if6 'it'xf6
20 lLlxf6 !Ig1+ 21 ..if1 ..ixh3 22 'it'hs) 18
..if6+ 'it'xf6 19 lLixf6 l:i:xf6 gave Black
compensation for the queen in
A.Aleksandrov-1.Zakharevich, St Peters­
burg 1994.
a22) 16 ...'it'g6! 17 lLixg7 (one point
of Black's play is that after 17 .l:f.g3 'iff7
the move 18 ..ih4 does not protect f2,
but Black has even better in the shock­
ing 17 ... ..ig4!! 18 ..ie7 'it'xhs 19 ..ixg4
'iff7 20 ..ixf8 l:i:xf8 with a strong attack)
17 ... 'it'xg7 18 'it'd2 lLib4 (Black does not
miss his dark-squared bishop because
his knights are tremendous) 19 o-o-o
lLixa2+! 20 @b1 lLixc3+ 21 bxc3 .ifs
was winning for Black in O.Jovanic­
Z.Kozul, Ljubljana 2004.
b) 13 a3 and now:
 Petro s i a n Va ria tion: 7 ds as with o u t 8 j,_95
bl) 13 ...j_d7 14 h4 f5 15 gxf5 (worse
is 15 h5 fxg4) 15 ... gxf5 16 h5 j_f6 17
�xf6+ !Ixf6 18 'it' c2 'it'f8 19 o-o-o fxe4
20 lLicxe4 l:i:f4 21 f3 j_f5 22 !Ig2 'it'h6
with counterplay, M. lvanov-0.Cvitan,
Cappelle la Grande 1995.
b2) 13 ...a4 is simple enough: 14 'it'c2
f5 15 gxf5 gxf5 16 o-o-o fxe4 17 lLidxe4
�f5 gives Black good play.
12 ...@hB!
A typical reaction to White's rook
move. Instead 12 ... c6 is inconsistent
and looks panicky to me, but Black may
be able to get away with it: 13 h5 cxd5
14 h6 j_h8 15 cxd5 (after 15 lLixd5 f6 16
..i.e3 Black should avoid 16 ... lLixe4 17
�b6 j_e6 18 lLixa8 'it'xa8 19 lLid2, when
White is much better according to
Piket, and instead play 16 ... j_e6!)
15 ...j_d7 16 a4 lLic7! (16...bs? 17 axb5
�c7 18 b6 lLib5 19 lLixb5 j_xb5 20 lLid2
gave White a clear advantage in J.Piket­
J.Polgar, 1st matchgame, Aruba 1995)
17 lLid2 b5 18 axb5 lLixb5 gave Black
counterplay in S.Giddins-R.Leitao, Ant­
werp 1998.
13 'it'd2 f6
I like this little nudging move, al­
though 13 .. .f5 14 gxf5 gxf5 15 j_h6
comes to the same thing.
14 j_e3 f5 15 gxf5 gxf5 16 j_h6
This may look dangerous, but Black
has enough resources to hold the bal­
ance. Note that strategically Black does
not mind the exchange of dark­
squared bishops as Jong as his king re­
mains safe.
16 ... j_xh6
Or 16 ... l:i:g8 17 j_xg7+ l:i:xg7 18 !Ixg7
@xg7 19 'it'g 5+ 'it' g6 20 'it'e7+ 'it'f7 with
equality according to Zvjaginsev.
17 'it'xh6 'it'e7 18 h5 !If6
Instead 18 ...!If7 19 !Ig6 j_d7 20 lLig 5
!Ig7 is also okay, J.lvanov-J.lbarra Jerez,
Mondariz 2000.
19 !Ig6 .l:f.xg6 20 hxg6 'it'g7
Black quickly takes over the initia­
tive in the endgame.
21 '1Wxg7+ @xg7 22 exf5 j_xf5 23 gxh7
lLib4 24 @d2 j_xh7 25 !Igl+ @f6 26
lLig5 j_g6 21 f3 !Ihs
Black had some advantage in the
ending in J.Alvarez-E.Pupo, Holguin
1997.
219
A ttacking C h e s s : The King 's I n dian, V o l u m e 1
02) 1o ll:id2
A flexible move.
10...ll:ifd7 11 a3
Instead 11 g4 ll:idc5 takes play back
into note 'a' to White's 11th move in
Line Dl, while 11 o-o ll:idc5 looks fine
for Black
11...f6! 12 il.h4
After 12 it.e3 Black plays 12 ...f5 im­
mediately, with good play.
12 ...ll:ib6
Again we see this move. Instead
12 ... f5? 13 exf5 gxf5 loses an exchange
after 14 iLh5, but 12 ... il.h6!? is possible.
Then 13 b4 ll:ib6 14 �bl axb4 (or
14...il.d7 when 15 bxa5?! ll:ia4! and
... ll:ia6-c5 is good for Black) 15 axb4 f5
16 o-o it.d7 17 '1!r'c2 ll:ia4 18 ll:ixa4 Jt.xa4
19 '1lr'd3 il.d7 gave Black a good position
in D.Brunsteins-Y.Yarmolyuk, corre­
spondence 2006.
13 b3-
13 b4 is not possible, because
Black's last move protected the a8-rook.
Instead 13 f3 just invites 13 ... il.h6. A
good example of Black's play went 14
Ji.f2 ll:ic5 1 5 ll:ib3 ll:iba4! 16 ll:ixc5 ll:ixc5
220
17 b4 axb4 1 8 axb4 �xal 19 '1!r'xa1 ll:ib3
20 '1!r'b2 ll:id4 21 it.d3 f5 22 ll:ie2 fxe4 23
it.xe4 ll:ixe2 24 '1!r'xe2 '1lr'a4 25 '1!r'b2 it.d7
with the initiative in J.Tihonov­
A.Fedorov, Minsk 2010.
13 ... il.h6 14 l':tbl
Or 14 o-o ll:ic5 15 �a2!? (15 l:tb1)
15 ...f5 16 f3 il.e3+ 17 il.f2 il.xf2+ 18
.l:!.xf2 ll:ibd7 (other possibilities are
18 ... '1lr'e7 and 18 ...il.d7) 19 b4 axb4 20
axb4 �xa2 21 ll:ixa2 fxe4! 22 fxe4 (22
bxc5 e3) 22 ...l:txf2 23 '1¥o>xf2 '1lr'f8+ 24
'1¥o>g1 ll:ia6 with equality in E.Mirosh­
nichen ko-A.Tukhaev, Kharkov 2007.
14...ll:ics
15 b4
After this move Black takes over the
initiative. 15 o-o is more solid and after
1 5 ... il.d7 (instead 1 5 .. .f5 16 b4 axb4 17
axb4 ll:ica4 18 ll:ib5 was awkward for
Black in M.Krasenkow-A.Vovk, Vlissin­
gen 2009) White has tried:
a) 16 '1!r'c2 f5 17 b4 ll:ica4 (17 ... axb4
18 axb4 ll:ica4) 18 ll:ib5 Jt.xb5 19 cxb5
'1lr'f7 was unclear in A.Jorgensen­
P .Hertel, correspondence 2003.
b) 16 b4 axb4 17 axb4 ll:iba4 18
 Petrosian Varia t io n : 7 d5 a5 with o u t 8 ii.gs
ll:ixa4 il..xa4! 19 '1!r'e1 il..xd2 20 '1!r'xd2
ll:ixe4 21 '1!r'e3 fS snared an important
centre pawn for Black in S.Nenciulescu­
D.Matic, correspondence 20os, al­
though White can claim some compen­
sation with his bishop-pair.
15 ...axb4 16 axb4 ll:ica4 17 ll:ibs
After 17 '1!r'c2 ll:ixc3 18 '1!r'xc3 Bl ack
has 18 ...�a2, intending .. .fs with a clear
advantage.
17 ...'1!r'd8! 18 '1lr'b3 '1¥o>g7
Black gets off the diagonal of the
white queen. This is solid enough, but
he could consider 18 ... cs!? or even the
forcing 18 ... il..xd2+!? 19 '1¥o>xd2 c6. After
20 dxc6 bxc6 21 cs+ dS 22 cxb6 cxbs 23
'1lr'xds+ '1lr'xds+ 24 exds ll:ixb6 2s il.. xbs
�a2+ 26 '1¥o>e1 ii.ts Black has a strong
initiative in the ending.
19 �d1 c6 20 ll:ia3 cs 21 bs ll:id7 22
ll:iab1 '1!r'as
Black intends ...'1lr'b4 with the initia­
tive, N.Legky-A.Shirov, French League
199S.
221
 Chapter 12
Petrosia n Variation
1 ds a s a ..tgs
1 d4 lLif6 2 c4 g6 3 lLic3 1lg7 4 e4 d6 S
tLif3 o-o 6 1le2 es 7 ds as 8 Ji.gs
This is White's main move in the
Petrosian Variation. By pinning the f6-
knight he makes it difficult for Black to
get in .. .fs. This variation has been fa­
voured with White by GM Igor Naum­
kin for over 25 years.
8 h6
...
In this case breaking the pin is
Black's best because retreating to e3
can be met by ...lLig4. It is important to
play this move immediately, else White
can play 8 ... lLia6 9 lLJd2 h6 10 1le3
222
when it is difficult for Black to create
counterpl ay. Black has sometimes tried
9...�e8 here, as in Line D of the previ­
ous chapter, but White has not wasted
time on the weakening move h2-h3 in
this line.
9 1lh4 lLia6 10 lLid2
This is the most common move, but
White can also play 10 o-o �e8 11 lLJd2
tran sposing to the main lines. Instead
11 lLiel is sometimes played. After
11 ... gs (worse is 11 ... tLics 12 f3 lLih s 13
lLibs, but Black can keep the game
tense with 11 ... 1i.d7!? 12 lLJd3 lLJh 7 13
f3 fS) 12 1lg3 Black has the well-timed
12 ... tLixe4! (a common motif that is al­
ways worth checking ) 13 tLixe4 fS 14
..\;lhs �e7 15 f3 fxe4 16 fxe4 �xfl+ 17
'1¥o>xf1 tLics 18 �e2 g4! (without this
move Black would just be worse) 19
..\;lxg4 1lxg4 20 �xg4 �g s! when the
e4-pawn falls after 21 �e2 �g6 or 21
�xg s hxgs.
10 �eB
...
Generally 10 ... 1i.d7 will just trans-

pose after 11 o-o (or 11 a3 �e8 12 b3)
11 ...�e8.
White's basic plan is to play o-o, a3,
b3 (usually necessary at some point if
White wishes to avoid his pawns being
fixed by ... a4), l':tbl and only then b4.
However, the move order is flexible for
both sides. Usually White castles here
or on the next couple of moves, but he
can also delay or omit castling in order
to accelerate his queenside play. We
will examine:
A: 11 a3 (without an early 0-0)
B: 11 0-0
Instead 11 b3 lLih7 12 a3 returns
play to Line A, as does 11 l':tbl il.d7 12
b3 lLih7 13 a3, while the lunge 11 g4 is
well met by 11...lLih7!.
A) 11 a3
Here we only examine lines where
White delays castling.
11...il.d7
This develops and protects the a8-
rook. Black must always be aware of
Pe tros ian Va ria t i o n : 7 dS a s 8 Ji.gs
the pin on the a-file: for example,
11 ... lLJh 7?! 12 b4! saves White some
tempi.
12 b3 lLih7
This is a multi-purpose move. Black
prepares either .. .fs or ... hs and ....ii.h6.
Another idea is 12 ... tLixe4!? 13 lLJdxe4
fS 14 f3 fxe4 15 tLixe4 il.fs. If Black can
fi ght for the e4-square with pieces, as
he can here, this idea is usually good
enough for equality.
13 l':tbl
Instead 13 o-o takes play into the
realm of Line B.
13 ... h s
Black prepares t o activate his bishop
223
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
on h6. This idea is usually more effec­
tive than pushing the f-pawn, but here
13 .. .fS!? is quite viable for tactical rea­
sons: 14 f3 (14 exfs il.xfs! is one point,
because 1s ll:ide4? walks into 1S ... il.xe4
16 ll:ixe4 �f4) 14 ... ll:ics (Black does not
want the a6-knight to be sidelined; in­
deed, this looks better than 14...ll:if6 lS
b4 axb4 16 axb4 ll:ihs 17 1lf2 ll:if4 18
..\;lf1 when Black's knights are ineffec­
tive) and now:
a) lS b4 axb4 16 axb4 ll:ia4 17 '1!r'c2
(not 17 ll:ibs? ..\;lxbs 18 cxbs ll:ic3 and
wins) 17...ll:ixc3 18 '1!r'xc3 ll:if6 gives
Black comfortable play: 19 cs?! can be
met by 19 ... gs 20 1lf2 fxe4 or 19 ... il.bs !
when White may regret his decision to
postpone castling for so long.
b) lS '1!r'c2 ll:if6 16 b4 axb4 17 axb4
ll:ia4 18 ll:ibs (after 18 o-o ll:ixc3 19
'1!r'xc3 ll:ihs Black has good counterplay:
for example, 20 l':tfcl ll:if4 21 1lf1 gs 22
1lf2 '1!r'g6 23 cs g4).
Here 18 ... il.xbs 19 cxbs '1lr'd7 20 o-o
ll:ihs 21 l':tfcl ll:if4 22 1lf1 has scored
well for White in practice, but instead
Black has 18 ...ll:ixds!, intending 19 exds
224
il.xbs 2 0 cxbs e4! with dangerous ideas
like ... ll:ic3 and ... gs.
14 f 3 il.h6 1S il.f2
Again White can transpose to Line B
with lS o-o.
1s '1!r'e1
...
Worse is 1s ...fs 16 b4, but 1s ...ll:ics !?
is certainly worth consideration. After
16 '1!r'c2 '1lr'e7 17 b4 axb4 18 axb4 ll:ia4
19 ll:ibs White must goes on an adven­
ture to fight for the initiative. Follow­
ing 19 ... c6 there is:
a) 20 ll:ia7 cxds 21 cxds (better is 21
exds, but Black's position is healthy
here too and even the pawn sacrifice
21 ...e4 looks promising) 21 ...l':txa7! 22
..\;lxa7 l:tc8 gives Black a very strong ini­
tiative.
b) 20 ll:ie7 l':tac8 21 dxc6 (21 '1!r'xa4
l':txe7) 21 ... bxc6 22 ll:ia6 cs 23 l':tal!? (23
bS ll:ib6 is unclear) 23 ... cxb4! ? with a
further divide:
bl) 24 l':txa4 il.xa4 2S '1!r'xa4 gives
Black a pleasant choice between
2s ... il.xd2+ 26 '1¥o>xd2 '1lr'gs+ 27 .ii.e3
'1!r'xg2 and 2s ... '1!r'gs 26 '1!r'xb4 '1!r'xg2 27
�gl '1!r'xh2.

b2) 24 ll:ixb4 �b8 2s ll:ids (after 2s
l:txa4 il.xa4 26 �xa4 Black again has
the 26 ...�g s idea) 2s ...�gs 26 ll:ie3 ll:ics
is pretty level.
16 b4
Instead 16 h4 is very ambitious.
White prevents Black from gaining
space on the kingside with ...h4, but
makes castling kingside less attractive:
16 ...ll:ics 17 �c2 fS 18 b4 axb4 19 axb4
ll:ia4 20 ll:ibs has been played a few
times by Zilberman. After 20 ...c6 21
ll:ic7 �ac8 22 ll:ie6 1lxe6 (22...cxdS !? 2 3
tllxf8 ll:ixf8 24 exds e4 was a bit specu­
lative in Y.Zilberman-J. Becerra Rivero,
Havana 1998) 23 dxe6 bS 24 0-0 �xe6
2S .rf.bdl ll:if6 matters were unclear in
Y.Zilberman-S.Porat, Tel Aviv 2002, al­
though Black was up a pawn.
16...axb4 17 axb4 h4
A thematic move. Black gains some
space and creates various possibilities
an the kingside, such as ... il.f4, ...�g s,
-�tf6-hS and ... h3.
• Wc2
White has also played the pawn sac­
nfice 18 cs!? dxcs and now:
Petro s i a n Varia tion : 7 d5 a5 8 1l g 5
a ) 1 9 bxcs!? ll:ixcs 2 0 ll:ic4 ll:ia4
(20 ...h 3 ! ?} 21 d6 �gs is very messy.
b} 19 bS ll:ib4 20 ll:ic4 and here
rather than 20 ... b6 21 d6 cxd6 22 ll:ixb6
OT 20 ...ll:ia2 21 �C2 ll:ixc3 22 �XC3, in
both cases with compensation for
White, I suggest 20 ... h 3 or 20 ...�fd8!?.
18 ...l:tfbB!?
This prophylactic move looks
stronger than the 18 ... �g S 19 �gl c6
(19 ...b6 is safer) 20 cs! of P.Rossiter­
J.Gallagher, British Championship,
Swansea 1987.
19 1ld3 �gs 2o l:tg1 cs!
A useful idea to know. Black seizes
the initiative right across the board.
21 dxc6 bxc6 22 bs ll:ics 23 g3 hxg3 24
hxg3 lLif8 2S ll:ib3 ll:ixb3 26 .t:!.xb3 ll:ie6
With active pieces and a safer king,
Black was much better in G.Danner­
G.Timoscenko, Budapest 1989.
B) 11 0-0
Castling at some point is White's
most logical way of playing.
11...ll:ih 7
Here too Black can play 11 ... il.d7
225
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
which will usually transpose to the
main line after 12 b3 (or 12 '1¥o>h1 ll:ih7,
while 12 a3 can be met by 12 ... a4 13
ll:ibs il.xbs 14 cxbs ll:ics 15 f3 b6)
12 ...ll:ih7 13 a3.
12 a3
White threatens 13 b4. Instead 12
'1¥o>h1 h S 13 f3 il.h6 14 a3 il.d7 15 l':tbl
(15 b3 transposes to the main line)
runs into 1S ...a4! (if Black does not play
this then White will save a tempo on
b2-b3), and then 16 ll:ibs b6 or 16 b4
axb3 17 ll:ixb3 b6, with a good game for
Black in both cases.
White has also been known to go 12
f3 when 12 ... hs, 12 ...il.d7, and 12 .. .fs
are all reasonable and may well trans­
pose to the main line.
12 il.d7
...
Instead 12 ...hs 13 f3 il.d7 trans­
poses, but here Black has to avoid
13 ... il.h6?! 14 b4!. Premature too is
12 .. .fs 13 exfs il.xfs (13 ... gxfs 14 il.hs is
annoying) 14 g4! (not 14 ll:ide4?? il.xe4
15 ll:ixe4 �f4) 14 ... il.d7 15 ll:ide4 with
an advantage for White in A.Veingold­
G.Kasparov, Moscow 1979.
13 b3
The alternatives are also important:
a) 13 ll:ibs hs 14 f3 il.h6 15 '1¥o>h1 (15
b3 is the main line and 15 l':tbl a4 varia­
tion 'b', but 15 '1!r'c2!? is possible) 1S ... il.e3
16 l':tbl and here 16 ...a4! can be played
because the d2-knight is loose.
b) 13 l':tbl is well met by 13 ... a4.
This advance is usually good after
White has spent a tempo moving his
rook. Moreover, if Black does not play
226
this, then White will play b2-b4 in one
move. Conversely Black should usually
wait for l':tbl before playing this ad­
vance, because after the sequence ... a4;
b4 axb3; ll:ixb3 White's rook is usually
better posted on al, from where it de­
fends the a3-pawn and supports the
advance a3-a4-aS.
Now, after 14 ll:ibs hS (the most
flexible, but 14...il.xbs 15 cxbs ll:ics 16
f3 b6 is fine too) 15 f3 il.h6 (Black
should probably avoid 15 ... il.xbs 16
cxbs ll:ics 17 b6!?) 16 il.f2 (or 16 b4
axb3 17 ll:ixb3 il.e3+) Black has:
bl) 16 ...b6 17 b4 axb3 18 ll:ixb3
il.xbs 19 cxbs ll:ics 20 ll:ixcs dxcs 21 a4
'1lr'e7 22 '1!r'c2 ll:if6 23 l':tal ll:id7 24 il.d3
l':ta7 25 z:ta2 �fa8 26 l':tfal h4 27 '1¥o>h1
'1¥o>g7 28 '1!r'e2 (White intends il.el and
as, so Black sacrifices an exchange to
keep the position closed) 28 ... z:tas 29
il.e1 il.f4 30 il.xas l':txas 31 il.c4 ll:if6 32
d6! (White gives back a pawn to open
lines for his rooks) 32 ...'1!r'xd6 33 l':tdl
'1lr'e7 34 '1!r'd3 l':ta8 35 l':ta3 with a slight
advantage for White in D.Vigorito­
A.Sherzer, Philadelphia 1997.

b2) 16 ... '1!r'e7 is more flexible. Black
does not need rush to exchange on b5.
After 17 b4 (17 '1lr'c2 can be met by
17... h4 or 17 ...b6) 11... axb3 18 ll:ixb3 b6
19 a4 f5 20 ll:id2 (White should proba­
bly change gears and play 20 exf5)
20 ... ll:ic5! Black developed pressure
against the e4-pawn in D.Vigorito­
A.Matikozian, Los Angeles 2003. Note
that 21 ll:ixc7? loses to 21 ... il.xa4.
c) 13 '1¥o>h1 ! ? is a way of reaching the
main lines of 'B2' and 'B3', below, with­
out allowing Black the option of 13 b3
f5 (after 13 '1¥o>h1 f5 White isn't vulner­
able down the long diagonal, so can
play 14 exf5! with advantage). Black
can deviate as well, but it is probably
better to transpose to the main line
with 13 ... h5 14 f3 il.h6 15 b3 (15 �b1
a4! was briefly discussed in the notes
to White's 12th, above).
Instead 13...a4 has been tried too,
but White has made a certain gain in
that he has reached this structure
without spending a tempo on �bl as
explained above. In fact, sometimes
(after l':tb1) the white rook return s to al
Petro s i a n Va ri a t i o n : 7 d5 a5 8 il.g5
to cover the a3-pawn and play a4-a5, in
which case here White could claim to
be two tempi ahead. After 14 ll:ib5
Black has:
cl) 14...il.xb5 15 cxb5 ll:ic5 16 f3 and
although it seems that Black is not so
badly off, White has saved time and
scored well from here.
c2) 14 ... h5 15 f3 il.h6 and now:
c21) 16 ll:ib1!? il.xb5 17 cxb5 ll:ic5 18
il.c4 (with the idea of il.f2 and ll:ic3 to
target the a4-pawn) 18 ... il.e3 19 ll:ic3 f5
(19 ... b6) 20 exf5 gxf5 21 '1!r'c2 '1!r'g6 22
l':tael '1!r'h6 (22 ...il.d4 also runs into 23
b6! cxb6 24ll:ib5 and 22 ...il.h 6 23 il.f2 is
no help) 23 b6! favoured White in
J.Nogueiras-A.Zapata, Matanzas 1994.
c22) 16 b4 is the most logical when
16 ... axb3 17 ll:ixb3 has scored tremen­
dously for White, even though Black's
position is not so terrible. That said,
11 .. .f5 (11 ... b6 18 a4!) 18 exf5 il.xf5
(18 ... gxf5 19 f4! also looks good for
White) 19 il.d3 was still pleasant for
White in l.Yanvarjov-0. Loskutov, Mos­
cow 1994.
Returning to 13 b3:
22 7
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
13 h s
...
Black has a sharp alternative in
13 .. .fS!?.
Then :
a) 14 f3 is played rather infre­
quently, but it may be best:
al) 14...ll:if6 lS il.f2 (after lS l':tbl
ll:ihs 16 �e1 ll:if4 17 il.fl gs 18 il.f2 g4
Black has counterplay) lS ...ll:ihs 16 g3
and it is not s o easy for Black t o create
play on the kingside.
a2) 14...b6 lS �bl f4 16 b4 axb4 17
axb4 gs 18 il.f2 hS transposes to varia­
tion 'a3'.
a3) 14... f4 lS l':tbl (White could also
play lS il.f2 or even lS b4!? axb4 16
axb4 ll:ixb4 17 �3 ll:ia6 18 �xb7 �8
19 l':tfb1 with the initiative) lS... gs 16
il.f2 hS 17 b4 axb4 18 axb4 b6 (18...�g6
19 cs g4 20 c6 gives White the advan­
tage) 19 cs! bxcs 20 bxcs ll:ixcs 21
il.xcs dxcs 22 d6 (22 ll:ic4 and 22 ll:ibs
are good alternatives) 22 ...cxd6 23 ll:ic4
and White had the initiative in J.Kraai­
D.Gross, German League 1998.
b) 14 exfs and here:
bl) 14. ..gxfs is very sharp after lS
228
il.hs �c8 (Kasparov's exchange sacri-
fice) 16 il.e7 l':te8 17 il.xe8 �xe8 18 il.h4
e4. White has tried both 19 l':tcl and 19
�c2 here with no clear conclusions.
b2) 14...il.xfs looks quite sound, in­
tending lS g4 e4 (1s .. .il.d7 16 ll:ide4 b6
is also feasible) 16 l':tcl e3!.
White has:
b21) 17 fxe3 �xe3+ 18 il.f2 �g s 19
h3 (no better are 19 '1¥o>h1 il.d7 20 ll:ide4
�e7 or 19 h4 �f4 20 gxfs il.es 21 lLif3
�g4+ 22 '1¥o>h1 �h3+ with a draw)
19 ... il.d7 2o ll:if3 �f4 is level.
b22) 17 gxfs exd2 18 �xd2 ll:ics 19
�dl (worse is 19 il.dl �xfs 20 �e2?
ll:igs 21 �xe8+ �xe8 22 ll:ibs ll:ixb3
with a clear advantage, P.Enders­
V.Babula, Berlin 1996) 19 ...�xfs 20 b4
(after 20 il.g4 �f4 21 �el Black can
play 21...�f8 or sacrifice his queen with
21 ...�xel+ 22 �xel �xg4+ 23 il.g3
ll:igs 24 �e2 hS) 20 ... axb4 21 axb4 ll:ie4
22 ll:ibs l':tf7 23 il.g4 lLihg S with ap­
proximate equality in l.Naumkin­
D.lsonzo, Lido Estensi 2003.
While 13 ...fs is playable, I prefer
13 ... h s to which we now return. Do also

note that White can sidestep 13 .. .fs
with 13 '1¥o>h l !? as discussed above.
14 f3 1lh6
Black has activated his bishop,
which may come to e3. White can pre­
vent the intrusion, dodge the check or
continue his queen side play. We have:
81: 1s i.f2
82: 15 �h1
83: 1s .:b1
Line Bl is independent, whereas
Line B2 can easily transpose into B3.
81) 15 1lf2
Petrosian Varia t i o n : 7 d5 a 5 8 1lg5
This keeps Black's bishop out of e3,
but it allows Black to improve the posi­
tion of his queen.
15 '1lr'e7!
...
From here the queen helps to hold
up White's cs-advan ce, while it may
become active via the gs-square. Typi­
cal ideas for Black include
...h4,...1lf4, ...ll:if6-hS and ... '1¥o>g 7 when a
piece may hop into g3 and ... l':th8 will
be a possibility too.
16 '1!r'c2
By far the most common move, but
there are others:
a) 16 �bl '1!r'g s looks strong, but af­
ter 17 l':tb2 .lii.h3 18 .lii. g3 h4 (Black can
force a draw with 18 ... '1lr'e3+ 19 1lf2
'1lr'g s, but must avoid 19 ... '1lr'xC3? 20
l':tc2) 19 f4 (not 19 gxh 3 hxg3) 19 ... exf4
20 ll:if3 '1lr'f6 21 es dxes 22 il..xh4 g s 23
gxh3 gxh4 24 .lii.d3 matters are not so
clear. Black could also just play the
move 16 ... h4.
b) 16 �a2 protects the d2-knig ht.
Black should just play 16 ...h4.
c) 16 l':tel h4 17 il..fl 1lf4 18 �bl
'1!r'g s 19 �b2 ll:if6 intending ... ll:ih s gives
Black a strong attack.
d) 16 '1¥o>h 1 seems to be asking for
trouble because a knight may come to
g 3 with check. Following 16 ...h4 17
�gl?! (even after 17 �bl ll:ics 18 b4
axb4 19 axb4 ll:ia4 20 ll:ixa4 �xa4 21 cs
h3 22 g3 z:ta2 23 ll:ib3 fS Black had the
initiative in V.Antoshin-B.Kantsler, Ni­
kolaev 1981) 17 ....lii . gs 18 '1!r'e1 '1¥o>g7 19
�bl ll:ics 20 b4 axb4 21 axb4 ll:ia4 22
ll:ib3 ll:ixC3 23 '1lr'xC3 lLif6 24 �al ll:ih s 2s
229
A ttacking C h e s s : Th e K i n g 's I n d i a n , Vo l u m e 1

�a8 �xa8 26 l':tal l':th8 27 '1¥o>g1 il.f4 28

il.e3 '1!r'g s 29 �2 fS Black had enduring
pressure in P.Kiss-V.Kotronias, Kavala
200S.
16...h4 17 �fdl

Now 2 1...ll:ihs was given as clearly

better for Black by Strikovic in Infor­
mant 45, based on 22 cs dxcs 23 bxcs
ll:ixcs 24 il.xcs? il.e3+ 2s il.xe3 '1!r'xe3+
26 '1¥o>h1 ll:ig3+ 27 hxg3 hxg3 and 0-1 in
11 ... il.f4 L.Santa Torres-A.Strikovic, New York
Black can also lead with the queen. 1988. However, White later improved
One successful example of this idea with 24 ll:ic4! in L.Santa Torres-M.Al
was 11 ...'1!r'g s 18 il.f1 ll:if6 19 l':tab1 ll:ihs Modiahki, Moscow Olympiad 1994,
20 b4 axb4 21 axb4 fS 22 cs h3 23 g3 f4 when he had good compensation for
24 g4 il.xg4 2S fxg4 '1!r'xg4+ 26 '1¥o>h1 the pawn. Instead I think that Black
ll:ig3+! 27 hxg3 fxg3 28 il.e2 g2+ 29 should play:
'it>gl h2+ 0-1, G.Flear-R.Douven, Charl­ 21...il.e3!
ton 1983. Black has a good position.
18 �ab1
18 ll:ifl '1!r'g s 19 il.d3 ll:if6 20 �ab1 82) 15 'it>hl
was drawn here in B.Malich­
W.Schmidt, Decin 1976, but Black could
certainly continue with 20 ...ll:ics,
20...'i¥o>g7 OT 20 ...ltJh S.
18 ...'1!r'gs 19 il.f1 ll:if6
Black has a couple of alternatives
here that should be investigated.
19 ... il.e3 fights for the dark squares and
19 ...ll:ics avoids Black's knight being
potentially sidelined on a6.
20 b4 axb4 21 axb4

230

This is a tricky prophylactic move.
Often play will transpose to Line B3,
but here we will only look at independ­
ent lines where White generally avoids
or delays l':t bl.
1s il.e3
...
Black has a couple of alternatives:
a) 1s ...'1!r'b8 intends to bring the
queen to a7. Then 16 '1!r'c2 il.e3 (or
16 ...'1!r'a7 17 l':tael!) tran sposes to note
'f to his 16th move, below, but White
could also play 16 il.f2 when 16 ...'1!r'd8
17 '1lr'c2 gives him an extra tempo over
Line Bl, with Black's queen on d8 in­
stead of the more flexible e1-square.
b) 1s ... ll:ics is worth considering. Af­
ter 16 l:tb1 (but not 16 b4? il.xd2) 16 .. .fs
17 '1!r'c2 ll:if6 18 b4 axb4 19 axb4 ll:ia4
White has:
bl) 20 exfs ll:ixc3 21 '1!r'xc3 il.xfs 22
..i..d3 l':ta2 23 ll:ie4 .�xe4! 24 fxe4 ll:ig4
gave Black the initiative in l.Naumkin­
A.Khalifman, USSR 1984.
b2) 20 ll:ibs was played in
l.Naumkin-J.Trapl, Namestovo 1987.
Black's best is 20...il.xbs 21 cxbs il.xd2
22 '1!r'xd2 fxe4 23 il.xf6 �xf6 24 fxe4
which Palliser considers slightly better
for White because of the open c-file,
but after 24 ... '1lr'e7 Black is probably
close to equality. The e4-pawn i s also a
potential weakn ess and White's bishop
is no better than Black's knight.
16 '1!r'c2
This is critical. White does not
bother with l':tabl and instead prepares
for war across the board. Instead 16
:b1 transposes to Line B3, while 16
Petro s i a n Va ria tio n : 7 d5 a5 8 il.95
z:ta2 '1¥o>g7 17 '1!r'b1 il.cs 18 ll:id1 c6 19
ll:ib2 should be met by 19 ... il.d4 20 ll:id3
ll:ics with equality, rather than
19...cxds?! 20 cxds ll:ic7 21 ll:id3 il.d4
(M. Narciso Dublan-F.Nijboer, Barbera
del Valles 2007) when 22 ll:ic4 gives
White a large advantage after either
22 ... gs 23 il.e1 '1!r'e7 24 ll:ixas or
22...ll:ibs 23 �cl ll:ic3 24 l':txc3 il.xc3 2S
ll:ixd6 '1!r'b8 26 il.e7.
16 fs
•••
This move is very natural, but Black
must keep in mind that White can also
play on the kingside, especially if
Black's dark-squared bishop goes wan­
dering off. The position is very rich and
there are several possibilities, of which
'f' is the most interesting:
a) 16 ... c6 looks premature: 17 il.f2
(even 17 cs!? is dangerous: for exam­
ple, 17 ...il.xcs 18 ll:ic4 intending ll:ia4
with the initiative) 17 ...Lf2 18 �xf2
'1!r'd8 19 l:tffl ll:if6 20 '1!r'b2 h4 21 f4!
(with the dark-squared bishops ex­
changed and Black's knights unable to
reach the es-square, this is a strong
break) 21 ... h3 22 g3 ll:ig4 23 fS ll:ie3 24
231
A t tacking C h e s s : The King 's I n dia n , Vo l u m e 1
�f3 '1!r'b6 25 lLifl ll:ig4 26 '1!r'd2! '1¥o>g7
(26 ...ll:if2+ 27 l':txf2 '1lr'xf2 28 ll:ie3 wins)
27 '1!r'gs ll:if2+ 28 �xf2 '1lr'xf2 29 ll:ie3
il.e8 30 f6+ 1-0 was M.lllescas Cordoba­
B.Gelfand, Linares 1994; 30 ...'1lr'xf6 31
ll:ifs+ wins.
b) 16 ... gs is inflexible, but solid: 17
il.f2 il.xf2 18 �xf2 h4 (worse is 18.. .fS?!
19 exfs .lii.xfs 20 il.d3! with a big ad­
vantage in G.Vescovi-A.Bachmann, Sao
Paulo 2006) 19 l':tg1!? '1lr'e7 20 g3 '1¥o>h8
21 ll:if1 fs 22 ll:ie3 f4 23 ll:ifs '1lr'f6 24
gxh4 gxh4 25 l':tfg2 with just an edge
for White in Liang Chong-Ye Jiang­
chuan, HeiBei 2001.
c) 16 ... ll:ics 17 l':tabl fS 18 b4 axb4 19
axb4 ll:ia4 20 ll:ibs (Palliser suggests 20
ll:id1) 20...il.xbs 21 cxbs f4 22 ll:ic4 il.d4
with a strategically unclear position in
S.Savchenko-V.Grigoriev, Alushta 1994.
d) 16 ... '1¥o>g7 17 �ael '1!r'b8 18 il.d3
il.cs 19 ll:idb1 gs 20 il.f2 h4 21 ll:id1
bS?! (this just weakens Black even fur­
ther on the light-squares) 22 ll:ie3 '1¥o>h8
23 cxbs il.xbs 24 il.xbs '1lr'xbs 25 ll:ic3
'1lr'b7 26 a4 �g8 27 ll:ibs ll:if6 28 ll:ic4
was much better for White in R.Palliser­
G.Jones, British Championship, Torquay
2009.
e) 16 ... il.cs 17 il.f2 il.xf2 18 �xf2 fS
(18...'1lr'e7 is safer) 19 exfs gxfs 20 l':tel
gave White some initiative in R.Palliser­
L.Trent, British League 2002.
f) 16 ... '1!r'b8!? and now:
fl) 17 il.f2 seems to justify Black's
play: 17 ... '1lr'a7 18 il.xe3 '1!r'xe3 19 f4
(Black also reached a good position af­
ter 19 �ael '1!r'h6 20 ll:id1 h4 21 '1!r'b2
232
ll:ics in D.Cummings-P.Cramling, Pula
1997) 19 ... exf4 20 l':tael '1!r'cs 21 '1!r'c1
'1lr'd4 22 �4 �ae8 23 l':tffl '1lr'g7 with a
good position for Black in A.Shirov­
J.Polgar, Madrid 1994.
f2) 17 l':tael! is Kramnik's plan.
White intends ll:id1, '1!r'c1 and il.d3-c2
when all of his pieces will be aimed at
the centre and kingside.
Black must take some care:
f21) 17 ... '1lr'a7 18 ll:id1 il.cs 19 '1!r'c1
l':tae8 20 il.d3 c6 21 ll:ib1 '1!r'b6 (better is
21 ...'1¥o>g7 22 ll:ibc3 with just an edge for
White according to Kramnik) 22 il.c2
'1¥o>g7 23 ll:ibc3 '1lr'a7 24 il.g3 ! was
V.Kramnik-J.Nunn, German League
1994. White intends 25 f4! with the
initiative.
f22) 11 ... il.cs 18 ll:idb1 '1lr'a7 19 ll:id1
c6 20 il.d3 cxds 21 exds (21 cxds l':tac8
looks okay for Black) 21 .. .fs 22 '1!r'd2
'1!r'b6 23 il.c2 (after 23 '1!r'h6 il.e8 24 il.c2
'1lr'C7 Black is ready to evict the white
queen with ...'1lr'g7) 23 ... '1¥o>g7 24 ll:if2
il.d4 25 ll:id3 ll:ics and Black had good
play in T.Polak-L.Gofshtein, Vienna
1996.
 Petrosian Va riatio n : 7 dS as 8 ii.gs

17 exfs gxfs 2 0 '1¥o>h8 21 '1!r'b2 tt:\cs

•••

.-�������-
Not 17 ... il..xfs? 18 il..d3 when Black's
bngside is starting to creak.

18 ii.fl
Or 18 �ael '1!r'g6 19 il..d3 and here:
a) 19 ...il..c s 20 tt:ldb1 '1¥o>h8 21 il..f2
:gs 22 �gl �g7 and Black had coun­
terplay in R.Tuominen-M.Hartikainen,
Finland 1997.
b) 19 ...il..h6 20 g4! ? '1¥o>h8 21 gxfS 22 b4 axb4 23 axb4 tt:\a4 24 tt:\xa4 z:txa4
25 il..d1 �as 26 f4
-
-

according to Komarov.
That may be a little bit extreme, but
Black has good play in any case.
18 il..xf2
•••

After 18 ...il..h 6!?, as played in Y.Pelle­

tier-G.Mohr, Buekfuerdo 1995, switch­
ing back to the queenside with 19 �abl
has been suggested by Palliser. 26 exf4
•••

19 z:txf2 tt:\f6 20 �gl! Opting for piece play, but 26 ... e4 27

Here we see another plan from tt:\f1 '1!r'g6 28 tt:\e3 '1¥o>h7 also looks okay
Kramnik. White again plays on the for Black.
kingside. 27 �e2

233
A ttacking Chess: Th e King 's I n di a n , Vo l u m e 1

27 l':txf4 �es ends White's initiative.

Black's position looks a little loose, but
there are not many pieces on the board
and he has enough activity to hold the
balance.

1 S....lii.e3+ 16 '1¥o>h1
16 .lii.f2 is also possible, though it is
less dangerous.

21 ...�g6
Black could also keep e7 covered by
playing 27 ...�f7. After 28 �d4 bS! 29
.lii . b3 l':tae8 30 l':txe8 l':txe8 31 �xf4 �e7
Black has counterplay. Moreover, 32
�h6+ gets nowhere after 32...'1¥o>g8 33
�g6+?! �g7 34 �xg7 'it>xg7 and Black
was even better in the endgame in
S.Cade-B. Benko, correspondence 2004.
28 l:te7 �f7 29 �xf7 �xf7 30 �d4 After 16 ... .lii.xf2+ (16 ... .lii.cs is sound,
Not 30 .lii.xh s �g7 31 .lii.f3 ll:ig4 e.g. 17 .lii.xcs ll:ixcs and now both 18
when Black has the initiative. �c2 �e7 19 b4 axb4 20 axb4 ll:ia4 and
30...�g7 31 �xf4 18 b4 axb4 19 axb4 ll:ia4, as in V.Anto­
This was V.Kramnik-B.Gelfand, shin-M.Chiburdanidze, Baku 1980, look
Linares 1994. Instead of 31 ... '1¥o>h 7?! fine for Black) 17 �xf2 Black has:
Black should have played 31 ... bS! with a) 11. . .ll:ics 18 �c2 �e7 19 b4 axb4
counterplay. 20 axb4 ll:ia4 21 cs (after the immedi­
ate 21 ll:ibs, Black can play 21 ... .lii.xbs 22
83) 15 �bl cxbs ll:ib6) 21 .. .fs?! (a lapse; 21 ...ll:ixc3
White gets on with his queenside 22 �xc3 fS was better) 22 ll:ibs ! and
play. This does not exclude White from Black's a4-knight was in trouble,
operation s on the kingside, however. F.Vallejo Pons-F.Jenni, Aviles 2000.

234

b) 17 ... 'fke7 is solid when the posi­
tion resembles a Bogo-lndian. With
some pieces exchanged, Black is less
worried about his a6-knight being side­
lined, because White's rook will be tied
to the defence of the b4-pawn. After 18
b4 axb4 19 axb4 practice has seen:
bl) 19 ... h4 20 'fkcl (the more enter­
prising 20 cs!? dxcs 21 bS lLib4 22 lLic4
with compensation has been played a
couple of times with success by Porper)
20... l':tfb8 (more solid is 20 ... cs with
equality) 21 cs! dxcs 22 bS lLib4 23
�c4, intending d6 or f4, gave White
compensation for the pawn in F.Vallejo
Pons-0.Cvitan, European Team Cham­
pionship, Plovdiv 2003.
b2) 19 ...cs prevents White's sacrifi­
cial idea: 20 dxc6 (20 bxcs tLixcs 21
-':Jb3 b6 was equal in 1.Khenkin­
D.Bokan, Moscow 1989) 20...bxc6 21
-'!Ja4 (21 bs lLics 22 lLib3 lLixb3 23 .l:txb3
-':Jgs 24 il.fl lLie6 heading to d4 or cs
can only favour Black) 21 ...l':tfb8 22 cs
occurred in B.Zlotnik-A.Kuzmin, Buda­
pest 1989. Now 22 ... il.e6 is level.
16 ...il.c s
Petro s i a n Va ria tio n : 7 d5 a 5 8 il.g5
Black uses the bishop t o prevent b3-
b4 and hits the a3-pawn at the same
time. Others:
a) 16 ... tLics looks inconsistent with
the bishop on e3. After 17 'fkc2 (17 lLJbs
'fkb8!? 18 b4 axb4 19 axb4 lLJa4 20 l':tb3
is also possible) 17 ...fs 18 b4 axb4 19
axb4 lLia4 20 lLibs il.xbs 21 cxbs f4 22
lLib3 'fkd7 23 tLias! gs 24 il.f2 il.xf2 2s
l':txf2 lLib6 26 lLixb7 White had won a
pawn in A.Yusupov-P.Scheeren, Eind­
hoven 1986.
b) 16 ... fS 17 exfs (after 17 b4 axb4
18 axb4 f4 Black has good play on the
kingside) 17...gxfs 18 f4! exf4 19 il.xhs
'fies is very unclear. Black has played
this several times with mixed results,
but I do not like Black's ragged pawn
structure and open king position.
17 'fkc1
11 ...fs
This is the most direct and looks sat­
isfactory. The position is very rich and
both sides may initiate play on either
side of the board. While the text move
is natural, there are other moves worth
considering:
23 5
A tt a c k i n g C h e s s : Th e King 's I n dian, Vo l u m e 1
a) With 17 .. .f6!? Black finds a way to
initiate play on the kingside without
opening things up too much: 18 ll:ia2
gs 19 il.e1 (19 il.f2) 19 ... g4 20 b4 axb4
21 axb4 il.e3 was unclear in l.Naumkin­
V.Tsarev, Moscow 1988.
b) 17 ...'1¥o>h8 is a normal waiting
move, but here it is slow and well met
by 18 ll:ia2!. Now:
b1) 18 ... gs 19 il.e1 (19 il.f2!?) 19 ... fs
20 b4 axb4 21 axb4 il.e3 22 exfs il.xfs
23 z:tb3 il.d4 24 ll:ic3 ll:if6 2s ll:ibs il.b6
26 il.f2 was agreed drawn in a fairly
level position, l.Naumkin-T.Paehtz, Bu­
dapest 1991.
b2) 17 ...fs 19 b4 axb4 20 axb4 il.e3
21 ll:ic3 cs 22 dxc6 bxc6 23 '1!r'd1!? il.d4
24 ll:ia2 ll:ic7 2 s ll:ic1 ll:ie6 26 exfs gxfs
27 ll:idb3 looked better for White in
l.Naumkin-J.Fedorowicz, London 1990.
Black's position is rather loose.
c) 17...c6 increases the tension. Now
both 18 dxc6 bxc6 19 ll:ia4 il.d4 and 18
'1!r'b2 il.d4 19 b4 axb4 20 axb4 cs look
fine for Black, while 18 il.f2 il.xf2 19
l::txf2 gives Black a pleasant choice be­
tween 19 ...'1!r'd8, 19 ...'1lr'e7 and 19 ...'1!r'b8!?.
236
That leaves:
cl) 18 ll:ia4 il.d4 19 il.f2 il.xf2 20
l':txf2 '1!r'd8! (again the position looks like
a Boga-Indian) 21 dxc6 il.xc6 22 ll:ic3
ll:ics 23 b4 axb4 24 axb4 ll:ie6 2s bS
il.d7 26 ll:ib3 '1!r'b6 27 l':tfl ll:if6 gave
Black an edge in A.Yusupov-B. Daml­
janovic, Saint John 1988.
c2) 18 ll:ia2 is the best try: 18 ... cxds
19 b4! (Black gets counterplay after
both 19 cxds bS! and 19 exds fS!)
19...axb4 20 axb4 il.e3 21 exdS! (better
than 21 cxds when Black gets good
play with either 21 ... �c8 or 21 ... '1!r'b8)
when Black's position looks quite good.
However, actually Black must be
careful: 21 ... il.h6?! weakens cs and af­
ter 22 ll:ic3 fS?! 23 '1!r'c2 (23 cs! looks
even stronger) 23 ...l':tc8 24 l':tfel g s 2S
il.f2 g4 26 cs! White had seized the ini­
tiative by targeting the es-pawn in
l.Naumkin-Ye Jiangchuan, Belgrade
1988. A better try would have been
21 .. .fs!?.
18 exfs gxfs 19 il.f2
After 19 f4 e4 Black looks fine, while
19 ll:ia2 '1!r'g6 20 b4 axb4 21 axb4 il.e3
 Petro s i a n Va ria t i o n : 7 d5 a 5 8 il.g5

22 il.f2 il.xf2 23 l':txf2 has been seen a

few times. Then the untried 23 ... lLif6!
looks flexible and best.
19 lLif6
...

Also satisfactory is 19 ... �g6 20 il.xc S

tLixcs (Black has also tried 20 ... dxcs?!,
but I could not do that to the a6-
knight) 21 b4 axb4 22 axb4 lLia4 with
play similar to the main line in
Z. Franco-C.Checa, Seville 1994.
20 il.xcs tLixcs
Compared to Kramnik-Gelfand in 21 b4 axb4 22 axb4 tLJa4
Line B2, Black is even better off because Black certainly had no problems in
White's pieces are passively placed. R.Pogorelov-Z.Plenkovic, Budva 2009.

237
 Chapter 1 3
Excha nge Variation
7 dxes dxes 8 1fxd8 !txd8
1 d4 ll:if6 2 c4 g6 3 ll:ic3 1lg7 4 e4 d6 5
ll:if3 0-0 6 Ji.el es 7 dxes dxes 8 '1!r'xd8
l':txd8
The 'dreaded' Exchange Variation.
This notorious drawing line has gone
so far as to discourage some players
from even playing 6 ... es. While the Ex­
change Variation does have obvious
drawing tendencies, Black should cer­
tainly not give up hope of playing for a
win. As Gallagher has mentioned be­
fore, this line is primarily played by a
few different types of players. I will
classify these opponents as follows:
238
1. Beggars
These players just want to try to make
a draw. Usually the player of the white
pieces is lower rated or perhaps the
tournament situation dictates that
White should try to draw. Unless a
quick draw suits the black player as
well, I absolutely agree with Gallagher
that these players should not be
granted a draw until every resource
has been exhausted. It is my experience
that m any beggars will crack at the
first sign of trouble. At the very least
we should make the game as miserable
an experience for them as possible.
2. Psychologists
Some players will play the Exchange
Variation anticipating that Black will
feel uncomfortable in the endgame.
There is some logic to this; most King's
Indian players are striving for a sh arp
game. Nevertheless, if Black is well pre­
pared this approach will lose its sting.
It is important in general, of course, to
 Exch a ng e Va r i a t i o n : 7 dxe5 dxe5 8 'jkxd8 'gxd8
be a well-rounded player capable of
continuing to play even in equal end­
games.
3. S pecialists
Some players just love to pl ay dry,
technical chess. These players believe in
White's chances in the Exchange Varia­
tion and are happy to try to prove that
he has a little something. These players
are the most dangerous, because
unlike beggars and psych ologists, they
are likely to be well prepared. Special­
ists will have studied the detail s of the
Exchange Variation and will be ready
to take advantage of lackadaisical play
by Black.
9 1i. gs
It is well known that g ll:ixes can be
met by 9 ...ll:ixe4 when Black is at least
equal. White can also force things with
9 ll:ids. After 9 ... ll:ixds 10 cxds c6 11
l.c4 cxds 12 il.xds ll:id7 13 Ji.gs 'ge8 we
have tran sposed to Line A2, below. If
Black wants to avoid this, he can play
11 ... bs!? 12 il.b3 il.b7 13 il.gs and then
either 13 ...'gd7 or 13 ... 'gc8.
The text reaches the basic position
for the Exchange Variation. The pin
gives White a certain pressure against
both f6 and c?. Indeed, both 10 ll:ids
and 10 hf6 followed by 11 ll:ids are
threatened. However, if Black can neu­
tralize White's slight initiative, he can
feel confident, because White's 9th
move left him with a weak d4-square,
whereas Black can still guard the dS­
square with ... c6 at some point.
From a theoretical perspective,
Black really only needs to study Line A,
but I will examine three systems in de­
tail. One reason for this is variety -
some players may face an opponent
that frequently uses the Exchange
Variation, so it is nice to have more
than one weapon. Another reason is
circumstantial - Black may want to
either play solidly or take some risks,
depending on the opponent and the
situation.
We will explore:
A: 9 .�e8
..
B: 9. ll:ibd7
..
C: 9... c6
Th e first of these lines, 9 ... 'ge8, is a
solid equalizer that was favoured by
Fischer. This is the line recommended
by Gallagher as well. White has great
difficulties obtaining even a shred of
an advantage against 9 ... l:te8, but be­
cause a great deal of simplification of­
ten occurs, it can also be difficult for
Black to try to win. Line B is more com-
239
A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1
bative, but White also has greater
chances of getting an edge. Neverthe­
less, White's advantage is not large and
this is a good choice as Black for those
that want to try to gradually outplay
their opponents. Line C is a gambit con­
tinuation. This option is quite popular
and it is in decent theoretical shape.
9 ... c6 is the most combative option, but
Bl ack does sacrifice a pawn, so he does
risk more than in Lines A and B.
There are two other moves.
a) 9 ... tlla6 is probably not very good.
It is worth examining though, if only to
compare it to a line we will see later in
the Makogonov Variation (in Volume
2). Play continues 10 tllds (10 tllxes
l:te8) 10 ... l:td6 11 �xf6 �xf6 12 tllxf6+
l:txf6 13 tllx es l:te6 14 f4 when Black
will have trouble getting his pawn back
without making any concessions.
The unsatisfactory choice is:
al) 14.. .f6 15 tll g 4 hs 16 tllf2 tllcs 17
o-o (or 17 �f3 tllxe4 18 �xe4 fS 19 �d2
fxe4 20 �e3 with a nice advantage)
17 ...tllx e4 18 tllx e4 l:txe4 19 �d3 and
White is clearly better.
240
a2) 14....rf.e8 1 5 o-o-o (15 �f3 f6 16
tll d 3 �fs 17 tllf2 tll cs is level) 1s ...f6 16
tlld 7! �xd7 17 .rtxd7 .rtac8 18 �f3 tllcs
19 l:tddl! with the idea of 19 ...tllx e4 20
l:thel.
a3) 14...tllcs 15 o-o-o! f6 16 l:td8+
(16 tllg4 tllxe4 17 �f3 is a little better
for White) 16 ... �g 7 and now White has
several ways to get an advantage, in­
cluding 17 tll d 3 (also good are 17 b4
tll x e4 18 �g4 and 17 tllf3 l:txe4 18
tlld4) 17 ...l:txe4 18 tllxcs l:txe2 19 tlld3
when he has the upper hand.
b) 9 ... l:tf8! ? is an idea of Lanka that
has been pioneered by Bologan and
Shirov.
White has:
bl) 10 tll d s tllxds 11 cxds c6 12 �c4
bS 13 �b3 �b7 14 l:tcl as 15 a3 (after
15 a4 bxa4 16 �xa4 cxdS! the main
point of Black's play becomes apparent
- if we compare this to the note to
Black's 12th move in Line A2, we can
appreciate that Black's rook is not
pinned) 1S ... a4 16 �a2 l:tc8 17 �e3 b4
18 axb4 a3! gives Black good play as
shown in several games.
 Exc h a n g e Varia tio n : 7 dxes dxes 8 'ilixd8 1:l.xd8
b2) 10 �xf6 �xf6 11 liJds .�d8 12
liJxes leads nowhere after 12 ...1:l.e8 13
l:.d1 c6 14 liJC3 �as 1s lLlf3 l:txe4.
b3) 10 o-o-o liJc6 (10 ... liJbd7 is Line
B) 11 h3 �e6 12 �e3 .rf.ad8 13 a3 h6
(13 ... liJas ! ? 14 liJd2 liJc6 could also be
tried) 14 b4 as lS bS liJd4! 16 liJxd4
exd4 17 �xd4 l:txd4 18 l:txd4 liJg4 19
�xg4 �xd4 20 �xe6 fxe6 21 liJd1 �cs
22 a4 l:td8 with some compensation for
the pawn in Wang Yue-V. Bologan,
Moscow 2006.
b4) 10 liJxes is critical according to
Bologan. After 10 ...liJxe4 11 liJxe4 �xes
12 o-o-o liJc6 13 f4 �d4 (or 13 ... �g7 14
.!l:if6+ �g7 1s h4!) 14 liJf6+ �g7 1s lLlds
f6 16 �h4 �fs, instead of the 17 �d3
of R.Ravisekhar-V.Bologan, Calcutta
1992, Bologan gives 17 l:thel when
White's good centralization gives him
slightly the better prospects.
We can conclude that 9 ... liJa6 is just
dubious, but 9 ... l:tf8 is playable if Black
is unhappy with the three (!) recom­
mendations given here.
A) 9 l:te8
...
This is the classical continuation.
Black simply protects his es-pawn
while breaking the pin on his king's
knight. Now 10 �xf6?! .�xf6 11 liJds
can be met with 11 ... �d8! when Black
will play ...c6 and have everything he
wants. 9 ... l:te8 has always been consid­
ered a solid equalizer, but correspond­
ingly it has a drawish reputation. Even
so, I think that with careful study, Bl ack
will be well prepared to seize the initia­
tive if White gets careless.
There are two main lines, with the
second being much more popular.
A1: '10 0-0-0
A2: 10 ll'id5
Al) 10 0-0-0
White just continues his develop­
ment, hoping th at his space advantage
will mean more than Black's long-term
control of d4. Generally some pieces
stay on the board in this variation and
there is scope for independent play.
10 liJa6
...
This is a natural way for Black to de-
241
A ttacking C hess: Th e King 's I n d i a n , Vo l u m e 1
velop. The knight covers the c7-square
and may go to cs, from where it pres­
sures the e4-pawn and can continue to
e6, with an eye towards the weakened
d4-square. Instead 10... h6 is also possi­
ble, but it is not really necessary for
Black to chase away the white bishop,
so I prefer to keep this move in reserve.
11 tlle1
Instead 11 tllx es should be met by
11 ...tllc s ! (11 .. .l:txes 12 .l:td8+ tlle8 13 f4
l:te6 14 �g4 l:tb8 lS es is probably
somewhat better for White, but here
12 ... l:te8 13 �xf6 l:txd8 14 �xd8 �h3 is
probably okay) 12 �xf6 �xf6 13 f4 c6
when Black wins back the pawn, while
11 tlld2 c6 12 tllb3 (12 f3 tllcs 13 �e3
tlle6 is also fine for Black) 12 ... tlle7 13 f3
tlle6 14 �e3 �f8 lS a3 b6 16 .l:!.d2 �a6
17 �bl was agreed drawn in
V.Akobian-A.Yermolinsky, Agoura Hills
2004.
11 ... c6
Black has also tried both 11 ...�e6
and 11 ... h6, while another move order
for Black is 11 ... tllcs 12 f3 tlle6 (or the
immediate 12 ...c6) 13 �e3 c6.
242
12 tllc2
Instead 12 f3 tllcs 13 tllc2 leads to
the same thing, but White could con­
sider controlling cs with 12 tlld3 ! ? here.
For this reason 11 ... tll cs might be pre­
ferred.
12 ... tllcs 13 f3
This is a typical position. White can
double rooks on the d-file and try to
expand on the queenside. Black will
often play moves like ... tlle 6, ...�f8,
...tllfd7, and perhaps ... as and ... b6. Ide-
ally Black would like to get in ...�cs too
to activate his dark-squared bishop.
13 ...tlle6
By avoiding ... h6 Black is able to play
this move with tempo. Black could also
maintain the knight on cs for the time
being and play 13 ... as 14 �e3 (or 14 b3
�e6 with the idea of ... tllfd7) 14 ...�f8,
which is level.
14 �e3 tllf4
Black heads in a different direction
and probes the kingside. He has also
tried 14.. .il.f8 with the idea of ... �cs. If
lS b4 as 16 a3 axb4 17 axb4 he could
try 17 ... cs or 17 ... tllf4. However, I would
 Exch a n g e Varia tio n : 7 dxes dxes 8 'ilixd8 '!J.xd8
prefer the prophylactic 14...as! ? when
15 tlla4 tlld7 intending ...�f8 looks fine
for Black.
15 �fl hs 16 h4 �f8 17 b4 tlle6 18 a3
b6 19 �b2 �b7
Chances were about equal here in
Lesiege- Smirin, Biel lnterzonal 1993.
A2) 10 tlld s
This forcing move is more popular.
It is not necessarily better, but it forces
simplification and generally leads to a
symmetrical pawn structure.
10... tll x ds 11 cxd s c6
If Black refrains from this move, his
backward c-pawn will become weak.
12 �C4
White wants to bring his relatively
passive bishop to the nice dS-square.
Other moves are also possible:
a) 12 dxc6 tllxc6 just helps Black de­
velop.
b) 12 d6?! f6 13 �e3 �e6 leaves the
d6-pawn weak. Black will play ... '!J.d8
and ...�f8 to round it up.
c) 12 '!J.dl is an inferior way to sup­
port the dS-pawn. After 12 ... cxds 13
exds e4 14 tlld2 Black can safely play
14...�xb2, while 13 '!J.xds �e6 13 �bs
�d7 14 l:txds �xbs 15 '!J.xbs b6 16 �e2
tll a6 is fine for him too.
d) The most interesting try is the
gambit continuation 12 o-o!?.
Black has:
dl) 12 ...cxds 13 exds e4 (after
13 ...�g4 14 h3 �xf3 15 gxf3 the
bishop-pair was more important than
White's fractured pawns in J.Jansson­
G.Gnichtel, European Club Cup, Fuegen
2006, but 13 ...tlld7 14 l:tfdl, as in
J.Jansson-H.Harestad, Oslo 2006, and
then 14 ... e4! ? is an idea) 14 tlld2 l:tes
(after 14...�xb2 15 '!J.abl �g7 16 '!J.fc1
tlld7 17 d6 tllb6 White has obvious
compensation for the pawn) 15 �e3
'!J.xds 16 tllxe4 tllc6 17 �c4 '!J.d8 18
'!J.adl �fS 19 tlld 6 gave White the ini­
tiative in J.Jansson-Thomassen, Norwe­
gian League 2008.
d2) 12 ... h6 13 �e3 cxds 14 exds fs
15 '!J.fd1 tlld7 16 .l:!.acl f4 17 �cs tllxcs?!
18 '!J.xcs �g4 19 '!J.a was better for
White in J.Jansson-K.Tryg stad, Euro­
pean Club Cup, Fuegen 2006, and 19
243
A ttacking C h e s s : The King 's I n d i a n , Vo l u m e 1
d6 ! looks even stronger. Instead Bolo­
ga n gives 17 ... e4 1s tll d4 tllxcs 19 l:txcs
<Ms with equality.
12 ... cxds
We will stick with this classical con­
tinuation. Black can also play 12 ...bs 13
�b3 �b7 14 lkl as lS a4 bxa4?! 16
�xa4, but here, compared to the 9 .. J:US
line, Black has problems because his c6-
pawn is pinned to the eS-rook. After
16 ...l:tcs 17 d6 f6 1s �b3+ <Ms 19 �e3
tlld 7 20 h4 cs 21 �e6 l:tdS 22 tlld2
White was much better in F.Vallejo
Pons-A.Morozevich, Monaco (blindfold)
2007.
13 �xds tlld 71
Black has also played 13 ...tlla6 and
13 ...tllc6, but White can still hope for an
advantage in those lines.
With the text, Black intends ...tllf6 or
... h6 followed by ... tllf6 when the pres­
sure on the e4-pawn will force White to
part with one of his bishops.
14 tlld2
After 14 �e2 h6 lS �e3 tllf6 or 14
o-o h6 1 s �h4 g s 16 �g3 tllf6 14 o-o-o
h6 lS �e3 tllf6 Black is able to execute
244
h i s main idea without any hindrance.
White can also try 14 l:tcl, which is
not so good, although Black has to be
precise: 14 ... h6! (14 ... tllf6? lS �xf6
�xf6 16 l:to is much better for White
and 14 ...tllb6 lS �b3 �e6 16 �xe6
l:txe6 17 l:to is not want Black wants
either) lS �e3 (lS �h4 g s 16 �g3 tllf6
is also good for Black) 1s ...tllf6 16 �b3
tllxe4 17 l:te7 �e6 lS �xe6 l:txe6 19
l:txb7 l:ta6 20 a3 tlld6! when Black's
better development and active pieces
give him an edge.
14...tllc s
15 0-0-0
other moves do not trouble Black at
all:
a) lS o-o �e6 16 �xe6 tll xe6
(16 ...l:txe6 is also good and has been
played with success by Hebden) 17 �e3
tlld4 and Black has the initiative. In­
deed, White has trouble stabilizing the
position: for example, lS �xd4 exd4 19
l:tacl l:tacS 20 f4 d3! 21 es f6 and
Black's bishop is very strong.
b) lS �e3 tll d 3+ 16 �e2 tllf4+ 17
�xf4 exf4 lS l:tacl �xb2 19 l:tc7 �e6!
 Exc h a n g e Varia tio n : 7 dx es dxes 8 'ilixd8 '!J.xd8
20 �xe6 '!J.xe6 21 '!J.xb7 (21 '!J.b1 �es 22
'!J.cxb7 '!J.a6 is good too for Black)
21 ...�c3 22 liJb1 '!J.xe4+ 23 'itf3 '!J.b4 24
.rtxb4 �xb4 2S '!J.cl '!J.d8 26 �xf4 �d2+
27 liJxd2 '!J.xd2 and Black won a pawn
in J.Vanheste-J.Gallagher, Metz 1991.
c) lS �e2 �e6 (it may be better to
just play 1S ...liJe6 16 �e3 liJf4+ which
is variation 'b') 16 '!J.hcl '!J.ec8 17 �e3
�xds 18 exds liJd7 19 liJe4 �f8 20 �gs
fS 21 liJf6+ liJxf6 was equal in
M.Rutkowski-B.Socko, Warsaw 2010.
d) 1S lLlc4 �f8 and here:
dl) 16 o-o-o �e6 17 �bl '!J.ac8 put s
pressure on c4 and e4.
d2) 16 o-o �e6 17 �xe6 (17 '!J.fdl is
probably a better chance to maintain
equality, intending 17 ...liJxe4 18 �xe4
�xc4 19 �xb7 '!J.ab8 20 �c6)
17 ... '!J.xe6!. At first this looks equal, but
Black actually has some initiative. He
can kick White's knight with ...bs and
his own pieces will all become quite
active. Indeed, after 18 f3 bS 19 lLle3 h6
20 �h4 liJd3 21 liJds '!J.c8 Black was
doing well in A.Acebal-J.Gallagher, Can­
das 1992.
d3) 16 '!J.d1 has been analysed by
Grivas: 16 ... �e6 17 f3 '!J.ac8 18 b3 (18
liJe3 liJa4! gives Black counterplay as he
points out) 18 ...�xds (or 18 ...bs 19 lLle3
�g7, whereas Grivas only quotes a
game with the stran ge 19 ... �d7) 19
.l:!.xds bS 20 liJd6 (20 liJe3? liJe6 is good
for Black) 20 ... �xd6 21 l:.xd6 liJe6 22
�e3 '!J.c2 23 '!J.d2 and now, rather than
Grivas's 23 ... '!J.ec8 24 �e2 when he
claims an edge, it looks to me like Black
can play the simple 23 ... '!J.cl+ 24 '!J.d1
'!J.c2.
Returning to lS 0-0-0:
1s liJe6
...
Black must avoid the greedy
1S ... liJd3+ 16 �bl liJxf2? because 17
'!J.dfl wins for White. However, 1S ...�e6
is a good alternative. After 16 �xe6
liJxe6 (16 ... '!J.xe6 is similar to Acebal­
Gallagher above, but in this case
White's king is more active, which al­
lows him to more easily defend and to
attack on the queenside: for example,
17 �bl �f8 18 �e3 '!J.c8 19 '!J.c1 '!J.ec6
20 f3 and Black has to be careful about
any simplification because White's
245
A ttacking C he s s : The King 's I n d i a n , Vol u m e 1
king can easily become active on the
queenside) 17 �e3 there is:
a) 17 ...�f8 is solid enough when
White has:
al) 18 tll b 3 as 19 l:tds (Black is cer­
tainly okay after 19 �bl a4 20 tllcl �cs
21 �xcs tll x cs) 19 ...a4 20 tll d 2 .l:tac8+
(both 20 ... tllf4 and 20... f6 are sufficient
as well) 21 �bl �cs 22 l:txes �xe3 23
fxe3 tll c s is equal.
a2) 18 �bl �cs 19 �xcs (or 19 tllc4
�xe3 20 tllxe3 �8 21 tllg4 tllcs 22 f3
h s 23 tllf6 .l:!.e6 24 tlld s .l:!.d6 2s l:tc1
V2-V2 B.lvkov-Z.Polgar, Wijk aan Zee
1986) 19 ... tllxcs 20 f3 tlld3 21 .l:thfl
l:tac8 22 tllc4 .l:txc4 23 .l:txd3 V2-V2
A.Onischuk-V.Bologan, Poikovsky 2009.
b) 17 ...tllf4 is more enterprising: 18
g3 (18 �xf4 exf4 19 f3 .l:tac8+ 20 �bl fS
gives Black enough counterplay)
18...tllg 2!
19 tllf1 (19 �gs h6) 19....l:te6 20 .:tg1
(otherwise, 20 �d2? l:tf6! 20 l:td7 l:tf6
gives Black good play, while 20 �bl
l:tf6 21 .l:td2 l:tf3 22 l:te2 tllxe3 23 tllxe3
l:td8 24 .!:tel �h6 with equality is given
by Bologan) 20 ... tllx e3 21 tllxe3 l:tf6 22
246
l:tgfl �h6 2 3 �bl �xe3 24 fxe3 l:txfl
2S l:txfl l:tc8 was a little better for
Black, albeit still drawish in
Ki.Georgiev-E.Sutovsky, Gibraltar 2006.
16 �e3 tllf4 17 �xf4 exf4
Black has not had any problems
from this position.
18 f3
Instead 18 tllc4 �e6 19 tlld6 .l:te7
does not get White anywhere. other
tries are:
a) 18 tllb3 �e6 19 �xb7 (better is 19
f3, transposing to the main line)
19 ....l:tab8 20 �ds �xds (20... as! gives
Black good play) 21 l:txds (or 21 exds
f3 ! 22 gxf3 l:te2) 21 ... l:txe4 is level.
b) 18 �bl �e6 19 �xe6 .l:txe6 20 f3
fS ! 21 exfs (after 21 l:tcl .l:td8 22 l:tc2
Black can play 22 ... �d4 23 exfs gxfs or
22 ... .l:ted6 23 tllc4 l:tdl+ 24 l:txdl l:txdl+
2S .!:tel l:txcl+ 26 �xcl fxe4 27 fxe4 bS)
21 ....:te2 (21 ... gxfs looks okay too - the
doubled pawns control some impor­
tant squares) 22 fxg6 hxg6 23 g3 (23
.l:thgl l:td8) 23 ... bS! with good compen­
sation.
18 ...�e6 19 tllb 3
 Exc h a n g e Va ria t i o n : 7 dxes dxes 8 'ilixd8 '!J.xd8
19 �xe6 '!J.xe6 is level as well.
19... �xds 20 '!J.xd s
20 ...fsl
Thematic. Black must open lines for
his rooks.
21 exfs l:tac8+
21 ....rf.e2 22 .rtd2 '!J.xd2 23 tllxd2 gxf5
is also good enough.
22 �bl '!J.e2 23 '!J.d2
Pia Cramling has preferred to make
draws after 23 f6 Lf6 24 '!J.d2 '!J.ce8
{1/2-1/2 P.Cramling-J.Gallagher, Biel
1991) or here 24 ... '!J.e7 M-V2
P.Cramling-E. Grivas, European Team
Championship, Debrecen 1992).
23...l:txd2 24 tllxd2 gxfs 25 '!J.e1
Or 25 .!:tel '!J.e8 26 .rf.C7 .rf.e2, while 25
.!Llb3 .rf.e8 26 tllcl l:td8 27 �c2 '!J.c8+ 28
it>bl '!J.d8 29 �c2 '!J.c8+ 30 �bl and
1h-V2 was J.Nogueiras-D.Anagnosto­
poulos, Linares 1996.
I know all of this seems rather bor­
ing, but Black should know how to
generate enough activity to hold the
balance.
2s ...�d41?
Instead 25 ...'!J.c6 26 .l:te7 '!J.g6 27 '!J.e2
�7 28 tllc4 .rf.e6 and 25 ...�7 26 .:tel
'!J.e8 27 '!J.C7+ .l:te7 28 '!J.xe7+ �xe7
should draw.
26 '!J.e2
The idea is that after 26 .l:te7 �e3 27
tllb3 .l:tc6! 28 .rtxb7 '!J.g6 Black has coun­
terplay.
26 ... �e3 2 7 tllb3 b6 28 '!J.c2 '!J.d8
Black has good counterplay in this
position.
B) 9...tll b d7
This combative move aims to keep
pieces on the board without giving up
material. It has been played by Rad­
jabov, Nakamura and a young Kas­
parov. If White does not act quickly,
Black will consolidate his position and
aim to exploit his dark-square control,
so White must play purposefully.
10 0-0-0
The most obvious and best. Others:
a) 10 tlld5 is premature: 10 ... c6 11
tlle 7+ �8 12 tll x c8 '!J.dxc8 is very com­
fortable for Black. White has spent sev­
eral tempi exchanging off the c8-bishop
and Black's pieces coordinate well:
24 7
A ttacking C he s s : Th e Kin g 's In dian, Vo l u m e 1
al) After 13 tlld2 tllcs 14 f3 tlle6
(14 ... as) l S �e3 tlld7 16 o-o-o �e8 17
g3 �f8 18 tllb3 as Black was quite com­
fortable in N.Brunner-M.Kazhgaleyev,
Vandoeuvre 200S.
a2) 13 o-o-o tll c s 14 �xf6 �xf6 lS
�d3 as 16 l:.hel l:te8 17 �fl �d8! was
S.Danailov-G.Kasparov, Dortmund
1980. Despite the opposite-coloured
bishops, Black is much better because
of his grip on the dark squares.
b) 10 .l:tdl l:tf8 (an important tactical
necessity because 10 ... l:te8? loses to 11
tllb s!) and here:
bl) 11 0-0 c6 is fine for Black, who
will play ... l:te8, ... as and ... �f8. The play
is similar to that in Line Al, except that
here White's king is not able to support
his queenside play.
b2) 11 tllb s is a shot in the dark:
11...c6 (or 11...tllxe4!? 12 tllxe7 tllx gs 13
tllxa8 tlle 6! with compensation in
C.Gorobinskiy-M.Golubev, Ukraine
1999) 12 tlld6 tllcs 13 tllxc8 (13 �xf6
�xf6 was quite comfortable for Black
in another C.Gorobinskiy-M.Golubev,
Ukraine 1999, clash) 13 ...l:taxc8 14
�xf6 �xf6 was better for Black in
R.Asylguzhin-A.Kuzmin, Dubai 2003.
The play is similar to Danailov­
Kasparov above.
b3) 11 tlld s should be compared to
Line B2, below. After 11 ... c6 (not
11 ... tllxe4 12 �e3 !) 12 tlle 7+ �h8 13
tlld2 (or 13 �e3 l:te8 - Black could also
try 13 ...tllb6 or 13 ... b6 - 14 tllxc8 l:taxc8
when the position is similar to Line B2
except that White's rooks are not con-
248
nected, which should favour Black
somewhat) 13 ... tllcs 14 tllxc8 l:tfxc8 l S
f3 as 16 tllb3 tlle6 17 �e3 �f8 Black
was fine in V.Perekhodkin-E.Pakhomov,
Dagomys 2010.
10 l:.f8
...
Again 10 ...l:te8? fails to 11 tllb s!.
Now White has two main tries:
.. 8i1;J.1i)ct2
1b: 11..�lcfs
White has also tried 11 tll e l, but
this does little to challenge Black's set­
up. After 11 ...tllcs (or 11 ...c6) 12 f3 tlle6
13 �e3 cs it's equal.
Another idea is 11 b4 as 12 a3 axb4
13 axb4 h6! (Golubev points out that
13 ... l:ta3 14 �b2 l:txc3 fails to l S tlld2!)
14 �h4 (both 14 �e3 tll g 4 and 14 �d2
tllb6! l S cs tlla4 are pleasant for Black)
14...l:ta3 ! ? (also tempting are 14 ... cs
and 14...tllh s) lS �b2 l:txc3 16 �xc3
(not 16 tlld2 gs ! 17 �g3 l:txg3)
16 ...tllx e4+, as given by Golubev, which
gives Black decent compensation for
the exchange.
 Exc h a ng e Va riati o n : 7 dxes dxes 8 'ilixd8 '!J.xd8
91.J 11 tll d2
This has been played rather often,
but it does not seem dangerous to me
if Black takes care.
11. •.c6
Black intends ... tllcs. The immediate
11...tllc s? was dealt a serious blow by
12 tll d s! (12 �e3 tll e 6 13 tll d s c6 leads
nowhere) 12 ... tllcxe4 13 tllxe4 tllxe4 14
ie3 ! when Black has great difficulties:
for example, 14... c6 (or 14...b6 lS �f3
fS 16 tllxe7 '!J.b8 17 �xe4 fxe4, as in
V.Kartachov-V.Evelev, Moscow 1998,
and now 18 tllbs would give White the
upper hand) lS tlle 7+ �h8 16 tllxc8
:fxc8 17 '!J.d7 and White won back the
pawn with a big advantage in L.S chan­
dorff-M.Golubev, German League 2002.
12 b4
White seizes space while he can. After
12 �e3 '!J.e8 (12 ... l:td8 !?) 13 cs �f8 14
b4 (14 tllb3 b6 14 tlla4 bS lS cxb6 axb6
16 tllxb6 tllxb6 17 �xb6 '!J.xa2 is also
pleasant for Black) 14... as lS a3 axb4
16 axb4 b6 Black seizes the initiative.
Black is fine too here after 13 f3 �f8 14
tllb3 tllh s ! 1s g3 tllg 7 (or 1s ... b6) with
the idea of ... tll e6, while after 13 h4 I
would suggest 13 ... tllf8!? heading for
the e6-square.
12 ...'!J.e8
It is worse to play 12 ... as 13 a3 l:te8.
Indeed, 14 �b2 tllf8 (14 ... b6 lS bS �b7
may be a better try) lS cs tlle6 16 �xf6
.�xf6 17 tllc4 axb4 18 axb4 tlld4 19
tll d 6 '!J.f8 20 �c4 '!J.b8 21 tlla 4 �e7?
(better is 21 ...�d8, but Black is
cramped in any case) 22 tllb6 was good
for White in T.Engqvist-T.Ernst, Han­
inge 1997. Here Golubev has pointed
out the resource 22 ...�xd6 23 cxd6 '!J.d8
24 f4! when Black is in big trouble.
13 �b2
Instead 13 cs looks premature after
13 ... as 14 a3 axb4 l S axb4 b6!, M.Haag­
A.Kochyev, Balatonbereny 1988, while
Black has counterplay after 13 tllb3
�f8 14 a3 as (14 ...b6 and 14...h6 also
look good) lS cs (not lS tllx as cs!)
1s...axb4 16 axb4 and now either
16 ... h6 or 16 ... �g 7, in both cases plan-
ning ...b6.
u...tllf 81? 14 cs tll e 6 15 �xf6 �xf6 16
tllc4 tlld4 17 tlld6 l:.f8
249
A ttacking C h e s s : The King 's I n dian, Vo l u m e 1
Black intends ... b6 (Golubev) with
counterplay - this is the reason for
leaving out ... as. Already I prefer Black.
e2) 11 tlld s
This is critical.
11 ...c6
Panczyk and llczuk suggest
11 ...tllx ds 12 cxds fs, but after 13 exfs!
I think White is better prepared for the
opening of the position.
Capturing the pawn with 11 ...tllxe4
runs into a familiar theme: 12 �e3!
(instead 12 tlle 7+ �h8 13 tllxc8 tlld cS!
14 tlle 7 tllxg s lS tllxg s �f6 is certainly
fine for Black and 12 tllxe7 tllxg s 13
tllxa8 tll e 6 is similar to one of the
Gorobin skiy-Golubev games in the
notes to White's 10th, above) 12 ... tllb6
(not 12 ... tlldcS?! 13 tlld2 ! when Black
has problems - Golubev) 13 tllxe7 l:tb8,
which leaves White with several good
continuations, such as 14 tlld 2.
12 tlle7+ �h8
13 ..lri.e 3
White prevents ...tll cs. After 13
tllxc8 l:taxc8 Black is fine, while 13 l:td6
250
tllxe4 14 tllxc8 allows the nice trick
14...tlld cS! when White cannot hold on
to everything.
u...l:te8
Another idea is 13 ... b6!?. Then 14
tllxc8 .rtaxc8 lS g4 tllcs gives Black
counterplay, while 14 l:td6 ..lri.b7 lS tllg s
h6 1 6 l:thd1 i s unclear. Instead 14 tllxc6
tllxe4 is untried and may be critical,
but the thematic move is 14 g4. Now
14 ... �b7!? (14 ... tll cs 1s tllx es lllfxe4 16
tll sxc6 looks insufficient) lS gs l:tae8
16 gxf6 (16 lllfs gxfs 17 exfs ..Iii.cs 18
gxf6 tllxf6 is murky) 16 ...tllxf6 17 tlld s
cxds 18 exds tllg4 gives Black counter­
play.
14 tllxc8 l:tax c8
Black can also play 14 ... l:texc8 as lS
tllg s �g8 16 ..lri.g4 .rf.d8 1 7 �xd7 tllxd7 is
fi n e for him. Instead lS g4 looks more
dangerous, as below.
After the text, White has the bishop­
pair, but Black has good squares for all
his pieces. White must try to seize the
initiative or Black will have good
chances to take over.
15 g41
 Exch a n g e Va ria t i o n : 7 dxes dxe5 8 'ilixd8 '!J.xd8
The most dangerous try. White ex­
ploits the shaky knight on d7 to disrupt
Black's coordination. Advancing the g­
pawn also gains space and gives the
e2-bishop some scope. Other moves are
less impressive:
a) lS tll g s �g8 (Black has also
played 1s ... l:te7 16 f3 and then either
the flexible 16 ...�f8 or the rigid 16 ... cs)
16 �g4 'fJ..e 7 17 �xd7 tllxd7 is fine, as in
the note to Black's 14th move, above.
b) 1s cs �f8 16 b4 (16 tllgs �g7 11
�c4 �xcs! is fine too) 16...b6 17 tll g s
�g8 18 �c4 'fJ.. e 7 19 cxb6 axb6 20 f3 bS
21 ..lib3 h6 22 tllh 3 'fJ.. e e8 (22 ... cs!) 23
tllf2 �xb4 24 �xh6 cs 2s .lids c4 (or
2s ... �c3 with the idea of ... �d4) was
very comfortable for Black in
S.Cvetkovic-G.Kuzmin, Linz 1990.
c) 1s tll d 2 �f8 16 tll b 3 b6 (16 ... as is
interesting, but then Black probably
should have recaptured on c8 with the
other rook) 17 g4! tllb8 (17 ... 'fJ.. ed8 looks
better) 18 �f3 cs 19 gs tllg8 20 �g4
'fJ..e7 21 h4 hS 22 �h 3 and White had
the initiative in N.Ostojic-M.Markovic,
Cetinje 1992.
1s...tllfs
Black cannot, of course, capture ei­
ther pawn.
16 tlld 2
White protects both e4 and g4 with
this move.
16 ...tlle 6 11 gs tlld 7 18 tllb 3
18 �g4 lk7 19 tllb3 transposes.
18 ...'fJ..c 7
After 18 ...'fJ..cd8 19 �g4 White will
take on e6, when Black will have to ei-
ther compromise his pawn structure or
put him self into a pin along the d-file.
19 �g4 �f8
Now Black is ready to free himself
with ... tlldcs.
2o �xe6 'fJ.. x e6 21 'fJ..d 3
Black has experienced some diffi­
culties from this position.
21 ...tllb6
Other moves are possible as well:
a) 21 ... cs 22 tlld2?! fS! 23 f3 f4 24
..lif2 h6 2s gxh6 gs 26 tll b 1 'fJ..x h6 gave
Black counterplay in T.Roussel Rooz­
mon-H.Nakamura, Philadelphia 2006,
but as Mikhalevski points out, 22 l:thdl!
tllb6 23 tll a s is very good for White.
b) 21 ...l:.d6 is probably the safest.
White has:
bl) 22 'fJ..h dl 'fJ.. x d3 23 'fJ..x d3 �e7 24
f4 b6 (24 ... �g7 2S �c2 f6 is a better try
to hold according to Mikhalevski) 2S fS
f6 26 h4 and White maintained an
edge because of his space advantage in
l.Piven-1. Drozdov, Budapest 1994.
b2) 23 'fJ..d l �e7 24 h4 (Mikhalevski
suggests 24 f4) 24 ... tllb6 2S tlld2 tll d7
(2 S ... �g7 with the idea of ...f6 looks
251
A ttacking C h e s s : The King 's I n dian, Vo l u m e 1
better) 26 tllf3 and White kept an edge
in A.Mirzoev-M.Erdogdu, Istanbul 2010.
22 tllc sl l:td6
This walks into some clever tactics.
After 22 ...�xcs 23 �xcs White's bishop
is strong and Black cannot play
23 ...tllxc4? because 24 l:td8+ �g7 2S
�f8+ mates, but 23 ...h6!?, with the idea
of 24 gxh6 tllxc4, is probably the best
try.
23 l:txd6 �xd6 24 l:td1l
24 b3 was enough for an edge, but
this is even stronger.
24 ...tll xc4 25 b3 tll x e3 26 l:.xd6 tllg4
This was Z.Gyimesi-1. S mirin, Maalot­
Tarshiha 2008. Now the cleanest line is
27 l:td2! tll xh2 (27 ...b6 28 tll a 6! followed
by 29 h3 traps the knight) 28 .l:td8+ �g7
29 tlld7!, winning the exchange.
C) 9 ... c6
This enterprising gambit became
very popular in the late 1980s. Black
makes the move he wants to make and
leaves the es-pawn to its fate. At first
9 ... c6 scored tremendously - White
players did not seem ready to grab a
252
pawn and then defend. Nowadays, this
line does not surprise anyone, but it
remains a viable choice for those will­
ing to takes some risks.
10 tll xes
White really has to take the pawn,
otherwise there is no chance to fight
for any advantage at all. After 10 o-o,
10 ...l:te8 is safer than 10 ... tlla6 11 tllxes
l:te8 12 f4, but Black could also try
10...h6, 10 ...t2Jbd7 OT 10 ... �g4.
10... l:te8
Not 10 ... tll a 6? 11 tllxf7 (11 tllxc6
may be even better) 11 ...�xf7 12 es
when Black will remain down a pawn
for nothing.
11 0-0-0
White tactically defends the eS­
knight and accelerates his develop­
ment. This certainly seems to be the
most logical move, but he has tried
other things too:
a) 11 f4 tllh s ! 12 �xhs gxhs (Black
threatens 13 .. .f6, winning a piece) 13
�h4 tlld 7! 14 �g3 tll xes 1s fxes �xes
16 �xes l:txes was level in E.Maljutin-
1.Glek, Moscow 1989.
 Exc ha n g e Va ria tion: 7 dxes dxes 8 'ilixd8 '!J.xd8

b) 11 �f4 tll a 6 (11...tllxe4 12 tllxe4 Grabbing the pawn is also poor:

�xes 13 �xes .l:txes 14 l:td1! is better 11 ...tll x e4 12 tllx e4 �xes 13 '!J.hel (not
for White) 12 o-o-o tllcs and Black will 13 tllf6+? �xf6 14 �xf6 tlld 7!) and
win back the pawn: for example, 13 f3 White has the initiative. However,
(or 13 .l:thel tllfxe4, while 13 �f3 runs 11 ...h6!? 12 �f4 tlla6 is an alternative.
into 13 ... g S !) 13 ... tllh s 14 �e3 tllxe4 lS
tllxe4 (after lS fxe4 �xes 16 �xhs
gxhs 17 �d4 �g4 18 �xes, as in
T.Johansen-S.Gabrielsen, Stavanger
1997, and then 18 ... �xdl 19 �f6 �g4
20 es Panczyk and llczuk claim that
White has compensation, but after
20...�fs 21 '!J.fl �g6 Black is much bet­
ter) lS ... �xes 16 �d4 �fs was
S.Bouaziz-J.Nunn, Szirak lnterzonal
1987. Black has recovered the sacrificed
pawn and has the initiative to boot. After 11 ... tlla6, the es-knight really
c) 11 '!J.dl is similar to 11 0-0-0, but is hanging. White has three options:
White's development is worse: 11...tlla 6
12 lllf3 (after 12 �6, 12 ...tllds ! ? is pos­ C1; U ti}fJ
sible, while 12 f4 h6 13 �h4 gs 14 �g3 C2:U f4
tllcs leaves White's king uncomfortable CJ: U lld6
on el, as shown by lS �f3 tllfxe4! 16
tllxe4 tllxe4 17 �xe4 gxf4 18 �xf4 Cl) 12 tllf3
�xes) 12 ... tll cs 13 tlld2 ( 13 es tllg4
does not help) 13 ... h6 14 �e3 (after 14
�xf6 �xf6 lS f3 as White's king can­
not help support the queenside)
14 ... tllfxe4 lS tll dxe4 tll x e4 16 tllxe4
'!J.xe4 17 '!J.d8+ �h 7 with the idea of
...�f6 compares favourably for Black to
Line Cl, below, because White's hl­
rook is not in play. There White's king is
on cl, although here it is still equal.
11 tlla 6
...

Taking the knight is bad: 11 ...'!J.xe s? This simple move has not scored
12 '!J.d8+ '!J.e8 13 �xf6 '!J.xd8 14 �xd8 well, but that is because White strug­
leaves White a pawn up for nothing. gled to consolidate in the early days of

253
A t tacking C h e s s : Th e King 's Indian, Vo l u m e 1
the gambit. In fact, the move is not
harmless and Golubev considers that it
causes Black some problems. Often
Black will secure the bishop-pair and
good squares for his pieces, but
whether this is sufficient compensa­
tion for a pawn is open to question.
Perhaps it is a bit of a matter of taste.
12 tllc s!
...
I believe this is best, as an examina­
tion of the alternatives reveals:
a) 12 ... tllg4 13 �h4 (13 l:thfl tllc s
wins back the pawn pretty cleanly)
13 ... tllcs when both 14 h3 tllf6 lS l:thel
and Golubev's suggestion 14 tlld 2 �xc3
lS bxc3 tllxe4 16 tllxe4 .rtxe4 17 l:td8+
�g7 look insufficient for Black.
b) 12 ... h6 13 �xf6 �xf6 14 tlld4 (14
es �xes lS tllx es .rtxes 16 .rf.d8+ �g7
with the idea of ... tlle7 is probably okay
for Black) 14...tllcs lS f3 when Black has
some compensation for the pawn, but
White is well centralized and I imagine
only the first player can really try to
win this position.
c) 12 ... �g4 13 �xf6! (after 13 h3
tllxe4! 1 4 tllxe4 l:txe4 13 es tlld 7 1 4 h3
�xf3 lS �xf3 tllx es 16 �e2 tllc s
Black's easy development and dark­
square control compensated for
White's bishop-pair in A.Torrecillas
Martinez-G. Hernandez, Benasque
1997) 13 ...�xf6 14 tlld4! l:tad8 (after
14 ... �xd4 Golubev gives the line lS
l:txd4 �xe2 16 tllxe2 cs 17 .rf.d7 l:txe4 18
tllc3 .rtxc4 19 l:txb7 tllb4 20 �bl! l:td8
21 a3! tlld s 22 tllxds l:txds 23 .:tel! l:th4
24 h3 l:tgs 2s g3 l:txh3 26 l:txa7 l:th2 27
254
l:tc2 when the passed a-pawn gives
White an enduring advantage) lS
�xg4 �xd4 16 f3 tllb4 17 �bl ! �xc3
(better is 17 ...�f8, but White still has
an extra pawn) 18 l:txd8! (18 bxc3 tll d3
19 �c2 tlle s gives Black compensation)
18 ...l:txd8 19 bxc3 tlla6 (not 19 ...tlld 3?
20 l:tdl with a deadly pin) 20 �c2 and
White has good chances of realizing his
extra pawn.
13 tlld 2
After 13 �xf6 �xf6 the pressure on
c3 and e4 should allow Black to win
back his pawn without any real conces­
sions.
13 h6
...
Another idea is 13 ...tllg4 14 �xg4
�xg4 lS f3 �e6 16 �c2 and now
16 ... tll a6 !? (Golubev gave 16 ... bs 17
�e3 tll a6 18 cxbs cxbs 19 �bl) 17 a3
(17 �bl tllb4) 17 ... bs 18 cxbs cxbs 19
tllx bs .l:tec8+ 20 tllc3 tllcs with the idea
of ...tll a4 could be investigated.
14 �e3
Perhaps better is 14 �xf6 �xf6 and
then:
a) lS f3 as (Black could try 1S ... �xc3
 Exch a n g e Varia ti o n : 7 dxe 5 dxe5 8 'ilixd8 '!J.xd8
16 bxc3 fS) 16 tllb3 tll e 6 17 tlla4 and
Black struggled to find enough for the
pawn in L.Voloshin-J.Helbich, Litomysl
1995.
b) 15 �c2 as (Bl ack should avoid
15 ... �xc3 16 �xc3 tllxe4+ 17 tllxe4
'!J.xe4 18 '!J.d8+ �g7 when both 19 �d3
and 19 �f3 !? '!J.xc4+ 20 �d2 are better
for White) 16 '!J.hel tlle6 (maybe not the
best; 16 ... a4! ? could be tried) 17 g3
tlld4+ 18 �bl was M. Sorokin-D.Pereyra
Arcija, Buenos Aires 1993. Black has
some compensation here, but again it
is not clear that it is enough.
14... tllfxe4 15 tlldxe4 tllxe4 16 tllxe4
'!J.xe4 17 '!J.d8+ �h7
This looks uncomfortable at first,
but Black intends .. if6, ...l:te8, and
.. ie6 when the position will be com­
pletely equal.
C2) 12 f4
Perhaps this move is the most obvi­
ous. White simply protects the ad­
vanced knight. The downside to this
move is that the e4-pawn may become
weak. Also Black has a tactical possibil-
ity which has evolved into his main
approach .
12 h6
...
After 12 ... tllcs 13 �f3 several moves
have been tried, but without much sue-
cess:
a) 13 ... as looks too slow after either
14 '!J.d6 OT 14 �h4.
b) 13 ...�e6 14 �h4 tllfd7 15 tllxd7
tllxd7 16 �c2 just leaves White a pawn
up for nothing.
c) 13 ... tllfxe4 14 tllxe4 tllxe4 15
�xe4 f6 16 �xf6 �xf6 17 '!J.hel when
White is a pawn ahead and well cen­
tralized.
d) 13 ... tlle6 14 Lf6 Lf6 15 tlld3
looks insufficient - White e4- and f4-
pawn duo is more of a strength than a
weakness.
e) 13 ...h6 14 �xf6 (14 �h4 trans­
poses to th e notes to Black's 13th in our
main line) 14 ...�xf6 15 tlld3 and White
keeps a useful extra pawn because
1s ...�xc3 can be met with 16 tllx cs!.
f) 13...tllh s 14 �xh s (14 tlld3 ?!
tllxd3+ 15 '!J.xd3 h6 is already better for
Black) 14...gxhs (threatening .. .f6) 15
255
A ttacking C h e s s : The King 's I n dia n , Vo l u m e 1
�h4 �h6 16 �c2 �xf4 17 tllxf7! tllxe4
18 l:thfl tll xc3 19 �xc3 �g4 20 l:td3 is a
rather peculiar position, but White still
maintains the better chances.
13 �h4
After 13 �xf6 �xf6 White's es- and
c3-knights allow Black to easily recover
the pawn.
u gsl
...
The alternative is 13 ...tllc s 14 �f3
gs lS �f2 tlle6 (after 1S ... gxf4 16 �xcs
l:txes 17 �d6 l:te8 18 �xf4 tll d7, as in
B.Finegold-G.Barbero, Wijk aan Zee
1991, Panczyk and llczuk claim that
Black has compensation, but I doubt
that it is enough) 16 fS! (after the 16
tllxf7 �xf7 17 es tllxf4 18 exf6 �xf6 19
tlle4 of J.Bellon Lopez-J.Becerra Rivero,
Santa Clara 1997, Black should play
19 ... �es 20 �d4 �fs, while after Bellon
Lopez's suggested 20 g3 Black has
20 ... g4!) 16 ...tllf4 (16 ...tllf8 17 �g3 is no
good for Black) and now:
a) 17 tll d 3 tll x d3+ 18 l:txd3 g4 with a
further divide:
al) 19 .lii.e2 tll xe4 20 tll xe4 l:txe4? 21
l:td8+ was winning for White in
256
G.Camacho-L.Valdes, Cuba 1998, but
20... �xfs is fine for Black.
a2) 19 �dl tllxe4 20 tllxe4 �xfs! 21
tll cs b6 22 .:te3 (or 22 l:tel bxcs 23
.l:.xe8+ l:.xe8 24 l:td6? �es 0-1 J.Bonin­
K.Pohl, Philadelphia 2000) 22 ...bxcs 23
.:txe8+ l:txe8 24 �xcs hS with some ini­
tiative for Black in J.Bonin-C.Riordan,
US League 2006.
b) 17 g3 ! is a problem, however, as
17 ...tll h3 18 tlld3 leaves Black strug­
gling.
14 �g3
This is the most common move.
White lends support to the h 2-b8 di­
agonal. There are a couple of alterna­
tives too:
a) 14 �f2 gxf4 1S tllf3 and then:
al) 1S ...tll x e4 16 tll x e4 �fs 17 tllfd2
�xe4 18 tllxe4 l:txe4 19 �d3 l:tes is
level. White has some compensation
for the pawn, but no more than that.
a2) 1S ... tll g4! ? 16 .lii.d4 tll e 3! ? (alter­
natives are 16 ... .lii.e6 and 16 ... �xd4 17
l:txd4 tllc s) 17 l:td2 �g4 18 h3 �xf3 19
gxf3 �xd4 20 l:txd4 l:tad8 looked prom­
ising for Black in J.Bellon Lopez-
 Exch a n g e Va riati o n : 7 dxe5 dxe5 8 'ilixdB '!J.xdB
S.Kindermann, Biel 1991, because of
White's bad bishop.
a3) 15 ... �g4 16 h3 with a further
split:
a31) 16 ... tllxe4 17 tllxe4 '!J.xe4
(17 ...�fS! ? is also possible) and here,
instead of 18 hxg4? .l:!.xe2 which was
clearly better for Black in H.Hoeksema­
J.Nunn, Groningen 1988, White should
settle for 18 �d3 �xf3 19 gxf3 '!J.es
which is level.
a32) 16 ...�hs 17 es �xf3 18 gxf3
.!l:ihs 19 '!J.hg1 �f8 20 tlle4 �xes 21
�d3 was P.Cramling-M.Wahls, Ham­
burg 1991, when Cramling gives
21 ... tllg 3 22 tllx g3 fxg3 23 �xg3 tll cs 24
i.h7 �xg3 25 l:txg3 as equal.
b) 14 fxg s hxgs 15 �g3 tllcs and
here:
bl) 16 �d3 tllh s 17 lllf3 �e6 18 �c2
.!l:ixd3 19 .:txd3 �xc4 20 '!J.d7 g4 21 tlld2
�a6 was better for Black in D.Starke­
P.Bakalar, Prague 1990.
b2) 16 l:thfl when Black has:
b21) 16 ...�e6 17 �f3 (17 lllf3 tllfxe4
18 tllxe4 tllxe4 is pleasant for Black)
17 ...g4 18 �e2 tllcxe4 19 tllxe4 tllxe4 20
tllxg4 fS 21 tllf2 tll xg3 22 hxg3 �f7 23
�d3 '!J.e3 with compensation for the
pawn in R.Thomas-J.Hebert, correspon­
dence 1997.
b22) 16 ...tllcxe4 is more forcing.
Black must be a little careful, but he
seems to be fine with exact play after
17 tllxe4 tllxe4 18 �hs tllx g3 (18 ... �e6
19 �xf7+ �xf7 20 tllxf7 tllx g3 21 hxg3
'!J.e2 looks a bit risky) 19 �xf7+ �h7 20
�g6+ �h6 21 hxg3 �xes (but not
21 ...'!J.xes 22 �c2! with a big advantage
for White) 22 �xe8 �g4 23 �f7 �xdl
24 '!J.xdl �xg3 with a drawn ending.
14...tllc s 15 � f3 g4 16 �f2
This leads to complicated play, but
ultimately leads to approximate equal­
ity. Instead 16 �e2 tllcxe4 17 tllxe4
tllxe4 18 �xg4 tllx g3 19 hxg3 �xes 20
�xc8 �xb2+ 21 �xb2 '!J.axc8 22 '!J.xh6
l:te2+ 23 �b3 '!J.xg2 is a drawish double­
rook ending.
16...gxf3
16 ...�f8 has also been tried, but 17
.l:!.hgl! is a clever way for White to keep
the tension .
17 �xcs fxg2 18 l:thg1 tll h s
257
A ttacking C h e s s : The King 's I n dian, Vo l u m e 1
19 �e3
White has also tried 19 tlld 3, but
Black gets sufficient counterplay here
as well:
a) 19 ... �xc3 20 bxc3 l:txe4 21 .l:txg2+
�h7 22 �d4 �h3 23 l:tf2 �f5 24 tllc5
l:te7 was level in B.lvkov-C.Werner, Wijk
aan Zee 2000.
b) 19 ... b6 20 �d6 �xc3 21 bxc3
l:txe4 22 .rtxg2+ �h7 23 tll e5?! �h 3 24
l:tf2 f6 25 tllxc6 l:txc4 26 tll e7 .l:txc3+ 27
�b2 .l:tc4 and Black had a good extra
pawn in F.Bindrich-G.Kamsky, Caleta
2010.
19... �h3 20 l:td3
White intends 21 �d2, uncovering
an attack on the h3-bishop.
20....l:tad8
The most precise move order. In­
stead 20 ...�xe5 21 fxe5 l:tad8 22 l:txd8
l:txd8 23 tlldl �h7 would transpose,
but gives White the extra option of 22
l:td6!?.
21 .l::!. xd8 .l:txd8 22 tlld1 �xes 23 fxes
�h7 24 tllf2 �e6
The ending still looks complicated,
but if neither side is able to generate
258
any initiative, the opposite-coloured
bishops will give the game rather
drawish tendencies.
2S b3
This allows Black to protect the far­
flung g2-pawn, but 25 .l:txg2 �xc4 is
not too dangerous. After 26 tllg4
Golubev gives 26 ...l:tg8 with the idea of
... l:tg6 and .. .f6.
2s ...J:tgs 26 �d2
White wants to walk his king to f3
to round up the g2-pawn.
26... bsl
Black is in time to create counter­
play on the queenside.
27 cxbs cxbs 28 �e2 .l:tc8 29 �d2 .l:tg8
3o �e2 as
Black could repeat, of course, but
this does not change the assessment of
the position.
31 �f3 a4 32 bxa4 bxa4 33 a3 �c4
34 .li es
Of course, 34 .l:txg2? allows
34... �e2+, while 34 tllg4 �fl (34 ... �e2+
35 �xe2 l:txg4 36 �f3 l:th4 37 l:txg2
l:th3+ 38 �e2 tll g7 is pretty level) 35
tllxh 6 l:tc8 looks very unclear.
 Exch a n g e Va ria tion : 7 dxes dx e5 8 'ilixd8 '!J.xd8
34...�fl 35 tllg4 l:tb8 36 tlle3
White could play 36 �b4, because
36...'!J.xb4 37 axb4 a3 38 tlle3 a2 39 tllc2
stops the a-pawn. Still, after 39 ... tllg 7
one gets the feeling that any chances
that there are lie with Black.
36...'!J. b3
Now any neutral move such as 37 �d6
should draw, but after 37 �f2? tllf4 38
.!Lixfl gxfl'ili+ 39 �xfl '!J.b1+ 40 �2
.!Llh3+ Black won in H.Terrie-F. Felecan,
Kona 1998.
C3) 12 '!J.d6
This counterattacking move was
brought to prominence by Tal, and was
later employed in a well-known game
between Korchnoi and Kasparov. It is
quite challenging and I will cover the
two main lines.
C31: 12 ..i.e6
••
C32: 12 tll h s
••.
Black has a few other options, of
which only 'c' is really important:
a) 12 ...'!J.xes 13 �xf6 �xf6 14 '!J.xf6
tllcs lS '!J.d6! (even in a quiet variation
Tal finds the tactics; instead lS f3 �e6
leaves the f6-rook in trouble) 1s ... tllxe4
16 '!J.d8+ �g 7 17 tllxe4 '!J.xe4 18 �f3
'!J.xc4+ 19 �d2 was much better for
White in M.Tal-1. S mirin, Podolsk 1990.
Black always has to watch out for these
back-rank pins.
b) 12 ...tll x e4 hands White the
bishop-pair: 13 tllx e4 �xes 14 tllf6+
�xf6 1s �xf6 tllcs 16 �f3 (16 �d3
tllxd3+ 17 '!J.xd3 �fs is equal) 16 ...�fs
17 '!J.hdl and White is obviously better.
c) 12 ...�f8 is rare: 13 '!J.xf6 '!J.xes 14
�f4 (Golubev's suggestion 14 �e3 may
be even stronger) 14...'!J.e8 lS �d6 �g 7
16 �xf8+!? (16 es �xd6 17 '!J.xd6 '!J.xes
18 '!J.d8 tlle7 looks okay for Black)
16 ... �xf6 17 �d6 with good compensa­
tion for the exchange.
C31) 12 ...�e6
This is Black's main move and it has
been more popular than all his other
moves combined. Of course it is play­
able, but I do not see a clear way for
Black to equalize.
259
A ttacking C h ess: Th e King 's I n dian, Vo l u m e 1
13 f4
Instead 13 f3 leads nowhere after
13 ... h6 14 ..ih4 gs 1s ..ig3 ll'lh s 16 ll'ld7
l:tad8 17 l:thd1 ll'lxg3 18 hxg3 l:txd7 19
l:txd7 ..ixd7 20 l:txd7 ll'lcs 21 l:td2 (after
21 :d1 ..ixc3 22 bxc3 fS Black is slightly
better) 2 1... ..ies when Black wins back
the pawn, 1.Umanskaya-A.Shchekachev,
St Petersburg 1994.
Another less-critical continuation is
13 ll'lf3 ll'lcs 14 ll'ld2 ll'lfd7 (trying to
corral the d6-rook, but both alterna­
tives 14.....if8 and 14...ll'lg4 lS ..ih4
..ic8!? look like better tries) lS ll'ldbl!?
fS! when Black has some initiative for
the pawn.
13 ...ll'lcs
This is the most popular move, but
13 ... h6 was Kasparov's choice. This al­
ways means something, but it still does
not look like a lot of fun for Black.
White has:
a) 14 ..ixf6 ..ixf6 lS l:thdl (after l S
ll'lf3 ..ixc3 1 6 bxc3 ll'lcs White must
avoid 17 l:td4 ..ig4 18 h3 ll'lxe4 19 hxg4
ll'lg3 when Black wins, and play 17 ll'ld2
when 17.....ifS 18 es l:tad8 gives Black
2 60
compensation according to Kasparov)
lS ... ..ixes 16 fxe S lfac8 (instead
16 ... 'iW8 17 cs! ll'lxcs 18 b4 is good for
White, but 16 ...l:te7 17 l:td8+ l:te8 is
fine) 17 'it>d2 l:te7 18 'iii> e3 l:tce8 19 l:td8
'iii>f8 was V.Korchnoi-G.Kasparov, Til­
burg 1991. White remains up a pawn,
although it will be very difficult to con­
vert.
b) 14 ..ih4 gs 1s fxgs hxgs 16 ..ixgs
ll'lcs 17 ..ixf6 ..ixf6 18 ll'ld3 ..ie7 19
ll'lxcs (after 19 l:td4 ..if6 20 es ll'lxd3+
21 l:txd3 ..ixes Black has compensation,
while after 19 es ..ixd6 20 exd6 it is
White who has compensation for the
sacrificed material) 19 ... ..ixd6 20 ll'lxb7
is complicated, but it looks a little bet­
ter for White.
14 ..if3
Or 14 ..ixf6 ..ixf6 lS ll'ld3 (after lS
J:thdl ..ixeS 16 fxeS aS O T 16.. lte7 Black
is fine) 1s ... ll'lxd3+ 16 ..ixd3 (16 l:txd3 ?!
..ixc4) 16.....ie7 17 cs ..ixd6 18 cxd6 f6
19 es fxes 20 fxes l:tf8 when the posi­
tion is unclear, but I do not think Black
is any worse.
14 ...ll'lhs
Here 14...h6 lS ..ih4 gs does not
work well because after 16 ..if2 g4? 17
..ixcs gxf3 18 gxf3 Black is just down a
couple of pawns. Instead 14...ll'lfxe4 lS
ll'lxe4 ll'lxe4 16 ..ixe4 f6 17 ..ixf6 ..ixf6
18 ..id3 also looks in sufficient
15 ..ixhs gxhs 16 ll'ld3!
This is better than 16 ..ih4 when
Black can simply play 16 ... ..ixes 17 fxes
..ixc4 or try 16 .. .f6 17 ll'lf3 l:tad8 with
counterplay.
 Exc h a n g e Va riat io n : 7 dxes dxes 8 -.xd8 :xd8
16 .....ixc3
Instead 16 ... ll'lxd3+ 17 l:txd3 ..ixc4
18 l:tg3 'iW8 19 es is much better for
White according to Golubev.
17 ll'lxcs ..ib4 18 ll'lxb7 ..ic8 19 a3!
White is better. Black should proba­
bly play 19 ...l:txe4, but White has sev­
eral moves to keep some initiative: for
example, 20 axb4 (other possibilities
are 20 l:txc6 ..ixb7 21 l:te7, 20 l:td8+ 'iii>g 7
21 ll'ld6 ..ixd6 22 :xd6 l:txc4+ 23 'it>d2
and 20 cs l:tc4+ 21 'iii>b 1 ..ixcs 22 ll'lxcs
.:xcs 23 l:thdl) 20...l:txc4+ 21 'iii> b 1 ..ixb7
22 l:thdl with a slight advantage.
c32) 12 ... ll'lhs
White's es-knight is now attacked
twice and .. .f6 is threatened, so the
knight has to move.
13 ll'lg4
This has been played several times
but may not be best. Alternatively:
a) 13 ll'lf3 ll'lcs 14 ..ie3 (14 ll'ld2 ..ies
lS :d8 .rtxd8 16 ..ixd8 ll'lf4 only gives
White trouble, while 14 es ll'le6 with
the idea of ...ll'lef4 is given by Golubev)
14.....ixc3 lS ..ixcs ..ig7 ! 16 ..ie3 (16 es
ll'lf4 17 ..ifl ..if8) 16.....if8 17 l:tddl is
given by Golubev, although after
17 ...ll'lf6 18 ll'ld2 ll'lxe4 19 ll'lxe4 l:txe4
20 .rf.d8 'iii>g 7 or simply 17 ...l:txe4 this is
probably not much for White.
b) 13 ll'ld3 ! ? contains the a6-knight
and may well be best: 13 .....if8
(13 .....ixc3 14 bxc3 l:txe4 lS l:td8+ 'iii>g 7
16 l:tel must favour White) 14 l:td8
(similar is 14 l:td4 ll'lg7 1s :d8 l:txd8 16
..ixd8 ..ie6 17 ..if6, C.Lakdawala­
J.Peters, Irvine 1998) 14...l:txd8 lS ..ixd8
..ie6 (1s .....ih6+?! 16 'iii>c2 ..ie6 17 ..ih4
..ixc4 18 ..ixhs gxhs 19 f3 was better
for White in J.Gustafsson-A.Matth aei,
Hamburg 1993) 16 ..ig s ..ixc4 17 l:td1
ll'lg7 18 f3 ll'le6 19 ..ie3 looks somewhat
better for White because of his strong
central position.
13 ... ll'lcs
After 13 .....ixc3 14 bxc3 (14 ll'lh6+!?)
14...ll'lcs (Black does better to play
14... ..ixg4, transposing to the next line)
lS ll'lh6+ 'iii>g 7 16 f3! ll'le6 17 ..id2!
White keeps an advantage according to
Golubev. A better move order for Black
is 13 ... ..ixg4 14 ..ixg4 ..ixc3 lS bxc3
261
A ttacking C h e s s : The King 's I n d i a n , Vo l u m e 1
l:txe4 with the idea of 16 l:td4 l:tes.
14 ll'lh6+
Instead 14 ..ie3 ..ixc3 lS ..ixcs ..ixg4
16 ..ixg4 l:txe4! is a clever trick given by
Golubev. After 17 ..ixh s Black can play
17 .....ies! with equal chances.
14...@fs 15 ..ixhs
Now lS f3 runs into 1S ...ll'le6! with
the idea of ...ll'lhf4. Probably best is lS
..ie3! ll'lxe4 16 ll'lxe4 l:txe4 17 ..ixh S
gx hS 18 l:td8+ �e7 19 l:txc8! .l:i.xc8
(worse is 19 ...l:txc4+ 20 @bl l:txc8 2 1
ll'lfs+ @d7 2 2 ll'lxg7 l:tg8 2 3 ll'lxhs l:txg2
24 ll'lg3) 20 ll'lfs+ @f8 21 ll'ld6 l:tce8 22
ll'lxe4 l:txe4 23 b3 as given by Golubev.
White has the better pawn structure,
but the position is drawish.
1s ... ..ixc3! 16 ..if3
Not 16 bxc3 ll'lxe4.
16 ... ..ies
This is queried by Golubev, who in­
stead suggests 16 ... ..ib4 17 l:tddl ll'lxe4
18 ..ixe4 l:txe4 19 l:td8+ l:te8 20 .l:i.xe8+
@xe8 or 16 ... ..ias!?.
11 .l:i.d2
After 17 l:tddl ..ic7! 18 l:thel fS !
(Black tries to exploit the possibility of
262
a knight landing on the d3-square) 19
es ..ixes White starts to experience
some difficulties because of the knight
stuck on h6.
17 .....ic7!
This move is an idea of the Ameri­
can IM Joe Fang. Black intends ... ll'le6,
hitting the gs-bishop and eyeing the
f4-square. The bishop retreat main­
tains control of f4 and uncovers an at­
tack on e4-pawn. All of White's at­
tempts to defend it make some kind of
tactical concession, as his line-up of
pieces on the cl-h 6 diagonal will prove
to be vulnerable.
In stead, both 17 ... ..ig7 18 l:thdl ll'le6
19 ..ie3 cs 20 l:tds b6 21 es! and
17 ... ..ie6 18 b3 (or 18 ..ie3 b6 19 b3)
18... as 19 ..ie3 b6 (A.Onischuk-
M.Golubev, Leningrad 1989) 20 @bl!
(Golubev) allow White to fight for an
edge.
18 .l:i.e2
In stead, 18 l:tel walks into 18 ... ..ias
and 18 ..ie3 ll'lxe4 19 l:tel @g7 20 ..id4+
@xh6 21 .l:i.xe4 l:txe4 22 ..ixe4 ..if4 23
..ie3 ..ixe3 24 fxe3 ..ie6 is a little better
 Exc h a n g e Va ria t i o n : 7 dxes dxes 8 -.xd8 'IJ.xd8

for Black. and maintains the initiative.

18 ...'iii>g 71 19 @bl 19 . ..ll'le6 20 h4
Black has all sorts of tricks: for ex­ If the bishop retreats, Black simply
ample, 19 h4 f6 20 ..id2 ..if4! 21 ll'lg4 plays 20 ...ll'lf4.
.ixg4 22 ..ixg4 l:txe4 and 19 g3 f6 20 20 ...ll'lxgs 21 hxgs ..if4
.i..d2 ll'ld3+ 21 'iii> c 2 ll'les. In both cases Black wins back the pawn and
Black wins back the sacrificed pawn stands well.

263
 Part Ill
The Samisch: variation
1 d4 t'i)f6' z c4 g6 3 liJc3 i.g7 4 e4 d'6
5 f3 o-o 6 ..te3 tiJc6 7 ttJgez a6 8 1fd2 :bs

1 d4 ll'lf6 2 c4 g6 3 ll'lc3 .i.g7 4 e4 d6 5 ming-up of the kingside pieces, the d4-

f3 square is sometimes weak
s ...o-o 6 .i.e3
The alternatives 6 .i.gs and 6 ll'lge2
will be discussed in Chapter 20. After 6
.i.e3 Black has a wide choice, as we
noted in the Introduction. The most
popular move today is definitely 6 .. cs,
.

but I had no hesitation in going for:

6...ll'lc6

The Samisch Variation is White's

other big main line with 4 e4. White
creates a strong centre and maintains
a great deal of flexibility. There are a
couple of downsides to White's scheme
of development, however. The move 5
f3 takes away the most natural square
for White's king knight and developing
it to the e2-square will block in the fl- This is the Panno Variation. It is
bishop. Because of this slight gum- flexible and ambitious. Black takes aim

2 64

at the d4-square with his knight, aim­
ing for active piece play rather than the
protracted manoeuvring that is com­
mon with the more classical 6 ... es and
6... cs. Black holds ... es in reserve and
after White overprotects the d4-square
with ll'lge2, Black will look to the
queenside with ...ftb8, ...a6 and ... bs.
This plan takes into account that White
may castle long and from a positional
standpoint it highlights the fact that
the c4-pawn is lacking defence.
The Panno is not very popular
nowadays, but I think this is mostly due
to the success of 6 ... cs at high levels
and consequently the decline of the
Samisch in general. 6 ...ll'lc6 has re­
mained popular in correspondence
play and has been played frequently by
Nunn and Gufeld, while Kaspan;>v has
been found on both sides of the board.
I believe it is due for a fresh look.
7 ll'lge 2
White sometimes plays 7 ..ifd2 first,
although it makes little difference
unless Black intends to play the
Th e S d m is ch Va riation
Westerinen Variation of Chapter 20, in
which case 7...l:te8 must be played, be­
cause 7 .. ltb8? 8 ds ll'les 9 i..x a7 wins a
pawn as the c4-pawn is still protected.
This and other early deviations will be
discussed in Chapter 20.
1 a6 8 ..ifd2 ftb8
...
Black is ready to play ... bs and so
White must decide on a plan. He may
play the aggressive 9 h4, which is con­
sidered in the next three chapters, pre­
pare to develop his kingside with g ll'lcl
(Chapter 17), initiate kingside play
without 9 h4 (Chapter 18), or play on
the queenside (Chapter 19).
265
 Chapter 14
Panno Variation
g h4 without 9 hs ..•
1 d4 ll'lf6 2 c 4 g 6 3 ll'lc3 .i.g7 4 e 4 d 6 5
f3 o-o 6 .i.e3 ll'lc6 7 ll'lge2 a6 8 'ilfd2 .l:i.b8
9 h4
This is the most aggressive move, al­
though White may still play position­
ally in certain lines. The move's basic
intention is obvious - White intends to
attack the black king. The play very of­
ten is similar to the Sicilian Dragon.
Usually Black halts the advance of the
white h-pawn at once with 9 ... h s, al­
though there was a time when it was
thought that this move was too weak­
ening. He may also strike in the centre
266
with 9 ...e 5 or initiate queenside play
with 9 . .bs.
.
Although 9 ...h s is considered the
main line nowadays and will be the
subject of our next chapter, 9 .. es and
.
9.. bs have not been refuted and they
.
lead to complex play. I imagine that
these two lines could prove to be un­
pleasant to face for some white play­
ers, especially those who intend to play
the positional 9 h4 h S 9 ll'lcl lines of
Chapter 16.
A: 9 es
...
B: 9 bs
...
A) 9 es
...
Black follows the logic that a flank
attack should be met with play in the
centre. However, because the d4-
square is well protected, it turns out
that Black actually cannot play in the
centre after all. Indeed, if 9 ... es is to
prove viable, it will be by combining
defence and counterattack.
 Pa n n o Va ria tio n : 9 h4 with o u t 9 . . . h 5
10 ds ll'las 1 1 ll'lg3
Geller's old book only considered 11
b3, but this move just weakens the
queenside for no reason and Black has
good play after 11...c5, intending ...b5.
The text move activates White's king­
side and threatens 12 b4.
11... cs 12 hs
After 12 .l:i.bl I think Black should
play Gallagher's recommendation
12 ...h5!, preventing White from having
his way all over the board. Instead
12 ...b6 13 h5 ..id7 14 b4 cxb4 15 l:txb4
"fic7 16 h6 ..ih8 17 'ilfb2 b5 18 cxb5
.:.fc8 19 bxa6! was winning for White
in V.Kramnik-V.lsupov, Kuibyshev 1990.
12 .....id7 13 ..ih6
It is hard to believe that 13 .l:i.cl can
trouble Black after 13 ...b5, while 13
o-o-o b5 14 ..ih6 ..ixh6 15 �xh6 trans­
poses to the main line.
13 .....ixh6
This is a common idea that is also
frequently seen in the Dragon. Black
invites the white queen to h6 hoping
that White's queenside will prove to be
more vulnerable from its absence
there. It is also possible to play 13 ... b5!?
as after 14 ..ixg7 'iii>xg7 if White wants
to invade with his queen, he must play
15 hxg6 fxg6 first, which could prove to
be premature. After 16 'ilfh6+ 'iii> g 8
Black can defend himself laterally
along the second rank.
14 'ilfxh6 bs 15 o-o-o
White can also play 15 ..ie2 when
15 ...'ilfe7 (this is the most popular
move, but I do not like it; the queen
should hang around because White has
ll'lf5 ideas in the air, but d8 is a good,
flexible square! ) 16 o-o-o ll'lxc4 (Black
could also try 16 ...'iii>h 8!? or 16 ... l:tb7) 17
..ixc4 bxc4 18 ll'lfl allows him a typical
267
Attacking C h ess: Th e King 's Indian, Vo l u m e 1
build-up with ll'le3, :d2 and g4. More
critical is the immediate 15 ... ll'lxc4 16
..ixc4 bxc4 17 0-0-0 l:tb4 when Black's
queen can still go to either side of the
board.
1s ...ll'lxc4
Instead 1S ... 'ilfe7 16 l:td2 .l:i.b7 was
seen in Y.Kraidman-H.Westerinen,
Gausdal 1983, but this looks too defen­
sive to me.
16 ..ixc4 bxc4 17 ll'lfl ..Was!
Now that there is no ll'lfs to worry
about, the queen takes up an aggres­
sive post on the queenside.
18 .lld 2
After 18 g4? l:txb2! (also strong is
18 .....ia4! 19 l:td2 l:txb2! 20 ll'lxa4 l:tb7,
winning) White resigned in
5
U. Lenhardt- .Crowdy, correspondence
1985. Following 19 'iii>xb2 l:tb8+ 20 @c2
..ia4+ 21 ll'lxa4 ..Wxa4+ 22 @d2 ll'lxg4!
23 fxg4 c3+ White's position falls apart.
18 ...l:tb4
Black has good counterplay. For ex­
ample, the natural 19 ..Wgs can be met
by 19 ...l:tfb8 (19 ... ll'lxhs 20 l:txhs .l:i.fb8
268
2 1 ..We3 gxhs 2 2 ll'lg3 i s less clear) when
Black's attack is the stronger. For ex­
ample, 20 'ilfxf6? .l:i.xb2 and White will
not last long.
B) 9...bS
Black continues with his plan.
10 hs es
This seems similar to the previous
line, but here Black has played ... bS to
attack c4 first, so the move ... cs will not
be necessary (although it is still possi­
ble). There are a couple of interesting
alternatives too:
a) 10 ... bxc4 and then:
al) 11 g4 ..ixg4!? (Black could inves­
tigate other moves) 12 fxg4 ll'lxg4 13
o-o-o ll'lxe3 14 �xe3 e6 15 hxg6 hxg6
(in stead 1S ...fxg6 16 'ifh3 ..Wgs+ 17 'iii>b1
l:txb2+ 18 'iii>xb2 .rf.b8+ 19 @al ll'lb4 20
..Wxe6+ ! 'iii>h 8 21 a4 ll'lc2+ 22 @a2 ll'lb4+
23 'iii>b l ll'ld3+ 24 'iii>al is winning for
White) was G. Kasparov-B. S passky, Nik­
sic 1983, and now 16 ll'lgl! looks good
for White.
a2) 11 hxg6 fxg6 12 ll'lf4 e6 (after
12 ... ll'las 13 o-o-o c6 14 g4 White has
 Pa n n o Varia t i o n : 9 h 4 wit h o u t 9 . . . h 5
the initiative, as shown i n some of
Murey's games) 13 ..ixc4 dS 14 ..ib3
l:txb3! lS axb3 dxe4 16 o-o-o exf3 17
gxf3 is unclear after either 17 ...0ie7 or
11 ... l'iJas.
a3) 11 ..ih6 has been considered a bit
of a refutation of Black's play. I am not
so sure about this, though: 11 ...0ib4! 12
0ig3 ..ixh6 13 'ilfxh6 0ic2+ 14 'iii>d l l'iJxal
lS hxg6 (after 1s l'iJds, instead of 1 s ... e6
16 hxg6 which was winning for White in
M.Ceteras-D.Elliott, Mamaia 1991, be­
cause of 16...exds 17 l'iJhs, Black can
play 1s ...gS! abruptly ending the white
attack) 1 s.. .fxg6 16 l'iJdS ..ie6 (but not
16....l:i.f7? 17 0ixf6+ exf6 18 ..ixc4, win­
ning) and here:
a31) 17 0,f4 ..if7 18 l'iJfhS was
V.Rajlich-K.Zalkind, Budapest 2000.
Here 18 ... gxhs 19 l'iJfs 0ie8 20 l:th3 ..ig6
21 ..ixc4+ e6! 22 ..ixe6+ �U7 23 l:tg3
'ilff6 would successfully defend.
a32) 17 0ixf6+ l:txf6 18 'ilfxh7+ 'iii>f8
19 'ilfh8+ ..ig8 20 ..ixc4 e6 2 1 l:th7 �U7
22 ..ixe6 lhh7 23 'ilfxg8+ 'iii>e 7 24 'ilfxg6
'it'h8 2S l'iJfS+ 'iii>f8 is unclear. Black is up
a lot of material, but it is hard to see
how he can untangle.
b) 10...l'iJas 11 0ig3 (11 0,f4?! pre­
vents White from playing ..ih6 and
makes little sense) 11...0ixc4 12 ..ixc4
bxc4 and now:
bl) 13 o-o-o c6! 14 ..ih6 ..Was 1s
..ixg7 'iii>x g7 16 hxg6 fxg6 17 'it'h6+ 'iii> g 8
18 es dxes 19 0ige4 was J.Peters­
A.Matikozian, Los Angeles 2004, and
here 19 ...:f7! looks good for Black.
b2) 13 ..ih6 ..ixh6 14 'ilfxh6 and
here:
b21) 14...g s 1s ..Wxgs+ (1s es!? may
also be good) 1 s ... @h8 16 o-o-o c6 17
:d2 l:tg8 18 'ilff4 ..Was 19 0if1 ..ie6 20
0ie3 gave White some advantage in
P.Van der Sterren-M.Bosboom, Brussels
1993.
b22) 14...l:txb2 !? is a sharp try: lS es
(or lS hxg6 fxg6 16 es dxes 17 0-0-0
l:txa2!?) 1s ... dxes 16 o-o-o (not 16 dxes
'ilfd4!) 16 ...exd4 is unclear. If 17 hxg6
(or 17 0ice4 .ifs) 17 .. .fxg6 18 0ice4 .l:i.f7
19 'iii>xb2, Black could try 19 ... cs,
19 .....ifds or 19 ...0ixe4 20 0ixe4 .ifs, in
all cases with a total mess.
We now return to 10...es:
269
A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1
11 d5
Invariably played, but 11 0-0-0 must
also be possible. Indeed, we'll briefly
discuss it in the notes to White's 11th
move in Line A of Chapter 18.
11 ll'las
...
12 ll'lg3 bxc4
It can be difficult to decide whether
to take with the pawn or the knight.
Taking with the pawn keeps the offside
knight, but this knight is more likely to
participate in an attack than White's
light-squared bishop, which does not
have much of a role. Taking with the
knight can still make some sense,
though, especially when White has al­
ready spent a tempo moving his bishop
or when the black queen can go to the
as-square. Here neither of those factors
come into consideration, so it is not
surprising that taking with the pawn is
correct.
Instead after 12 ...ll'lxc4 13 ..ixc4 bxc4
14 o-o-o ll'ld7 lS ..We2 (lS hxg6 fxg6)
1s ...ll'lb6 (some correspondence play has
shown that 1s ...fs 16 hxg6 f4 17 ll'lhs! is
promising for White), both 16 .l:i.h2, from
2 70
J.Tarjan-L.Nezhni, USA 1982, and Galla­
gher's suggestion 16 ll'lfl, with the idea
of ll'ld2xc4, favour White.
13 0-0-0
13 J:tb4!?
..
This is a good, active move. More­
over, the alternatives may leave the as­
knight looking vulnerable:
a) 13 ...ll'ld7 14 hxg6 fxg6 lS ll'lbl!
l:tbs (1s ...ll'lb7 16 ..ixc4 just looks good
for White) 16 b4 cxb3 17 ..ixbs cs 18
dxc6 (18 ..ie2 is also strong) 18 ... axbs
was J.Timman-G.Kasparov, Bugojno
1982. Here 19 cxd7 ll'lc4 20 ..igS ! would
be decisive.
b) 13 ... ..id7 14 ..ih6 (White has an
extra option in 14 ll'lbl or 14 hxg6 fxg6
lS ll'lbl, although the latter is not so
clear after 1s ....rf.bS: for example, 16 b4?
�8! and White was in some trouble,
V.Rajlich-E.Kahn, Budapest 2000)
14... ..ixh6 lS �xh6 .l:i.b4 16 l:td2 'ilfe7
(but not the 16 ... �b8? of A.Harakis­
E.Kahn, Budapest 2000, because of 17
ll'lfs! ..ixfs 18 exfs with a winning at­
tack) 17 ..ie2 leads back to the main
line.
 Pa n n o Va r i a t io n : 9 h4 with o u t 9 . . . h 5
14 ..ih6 ..ixh6 1 5 �xh6
15 .....We7
The alternative is 1s ...'iii>h 8. Black
does not commit his queen and creates
the possibility of chasing the white
queen away with ... ll'lg8. After 16 hxg6
fxg6 17 ..ie2 (not 17 'ilfxg6? l:tg8)
17 .....id7 18 ll'lfl Black has sometimes
played the funny 18 ...'iii> g 8!?. This looks
like a silly waste of time, but Black is
happy to resolve the tension on the
kingside so that he may defend later­
ally along the second rank Thus he
moves the king back so that the g6-
pawn is no longer hanging. After 19
ll'le3 ..ifb8 20 :d2 ..ifb6 the position was
unclear in H.Rauch-J.Stephan, corre­
spondence 1996. Here Black could also
consider 18 ... l:tg8, although using the
rook to protect the g6-pawn looks
rather inefficient. After 19 ll'le3 (19 g4
is given as good for White by Djurhuus,
but this is still not so clear) 19 ...'ilff8 20
�h2 ll'lhs ! 21 g3 �h6 22 'ilff2 �gs 23
:hgl, as in L.Psakhis-R.Djurhuus,
Gausdal 1994, both 23 ... :f8 and
23 ...:gb8 give Black decent play.
16 ..ie2 ..id7
Instead 16 ... ll'ld7 17 l:tdfl! is given
by Kasparov. White intends to play f3-
f4 with the initiative.
17 ll'lfl
This is a typical regrouping, espe­
cially when it is clear that the knight
will not be hopping into the fS-square.
If White starts instead with 17 :d2,
Black can try 17 ...c6!? (instead 17 ...l:tfb8
18 ll'lfl heads back to the main line) 18
ll'lfl when, instead of 18...'iii>h 8? 19 hxg6
fxg6 20 �xg6 cxds 21 ll'le3 ! with a big
advantage for White in J.Lautier­
P.Svidler, Internet 2004, Black should
have played the logical 18 ... l:tfb8.
11 .. AfbB 18 .lld2
White has solidified his castled po­
sition and is ready to build up on the
kingside. Black has:
a) 18 ... cs?! (this does not do much
for the black cause) 19 ..id1! (after 19
g4 ..ia4, with the idea of ... �7. Black
can hope to create some counterplay)
19 ... ll'le8 20 hxg6 fxg6 2 1 g4 'ilf g7 22 gs !
and White has the initiative even i n the
ending. Now 22 ...�xh6 23 :xh6 ll'lg7
2 71
A ttacking C he s s : The King 's I n dian, Vo l u m e 1
24 f4! prevented ...lllh s and after
24...exf4 2S l:tdh2 White was much bet­
ter in G.Kasparov-V.Loginov, Manila
Olympiad 1992. Thus Black should
probably leave the cs-square open for a
knight.
b) After 18 ... ..ie8 Kasparov gave 19
g4 llld7 20 hxg6 fxg6 21 ll:le3 as much
2 72
better for White. Matters are not so
clear, however, following 21 ... lllc s 22
gs ..id7 (to stop lllg 4) when White can­
not regroup so easily because both 23
..id1? llld 3+ and 23 ..if1? llla b3+ are
good for Black.
c) Another possibility is 18 ... ..ia4 19
g4 llld7 20 hxg6 fxg6 21 ll:le3 lllc s.
 Chapter 15
Panrio Vari:ation
9 h4 hS 10 a.o-o
1 d4 ll'lf6 2 c4 g6 3 ll'lc3 .i.g7 4 e4 d6 5
f3 o-o 6 .i.e3 ll'lc6 7 ll'lge2 a6 8 1lfd2 l:t b8
9 h4 hs 10 o-o-o
This is the main line. White contin­
ues playing for an attack, although he
may prefer to completely change gears
with 10 ll'lcl, which is covered in the
next chapter.
Before we examine 10 0-0-0, we
should look at a couple of other moves
which sometimes transpose to the
main lines:
a) 10 ll'lds bS 11 cxbs (alternatively,
11 ll'lxf6+ .i.xf6 12 cxbs axbs 13 l:tc1
transposes, while 11 0-0-0 is Line A,
below) 11 ...axbs 12 l:tcl .i.d7 makes a
funny impression . Instead of develop­
ing his pieces, White tries to play all
over the board.
Play is likely to continue 13 ll'lxf6+
.i.xf6 (or 13 ...exf6!? when Black has
counterplay after 14 g4 hxg4 lS fxg4
l:te8 or 14 ds ll'le7 1s ll'ld4 fS; in the lat­
ter variation, another option is 14... ll'les
1s ll'ld4 fS !? 16 exfs .i.xfs 17 ll'lxfs gxfs
with an unclear position) 14 g4 hxg4
lS fxg4 (after lS hS es 16 dS gxf3 17
dxc6 fxe2 18 'ilfxe2 .i.e6 19 'ilfg2 gs
2 73
A ttacking C he s s : The King 's I n d i a n, Vo l u m e 1
Black was better in S.Nikolic­
M.Cvorovic, Kladovo 1991) 1s ...es 16 dS
ll'ld4 17 ll'lxd4 (Black has compensation
after 17 ..ixd4 exd4 18 gs ..ies 19 ll'lxd4
l:ta8!?) 17 ... exd4 18 ..ig s (following 18
..ixd4 both 18 ... .l:i.e8 and 18 ... ..ixg4 look
good) 18 ... .l:i.e8 19 �f4.
Perhaps slightly surprisingly con­
sidering White's 11th and 12th moves,
we have come to a theoretical divide:
al) 19 .....ixgs 20 hxgs �e7 21 ..ig2
and here Black can play either 21 .....Wes
22 �xes (22 o-o 'ilfxf4 23 nxf4 l:tb7 was
given by Gallagher) 22 ...l:txes 23 l:txc7
..ixg4 24 l:tfl l:tf8, with an unclear end­
ing, or 21 ...l:tbc8!? with the idea of 22
'it'h2 ..Wes.
a2) 19.....ies 20 �f3 and now 20.. .f6
2 1 ..id2 (R.Gunawan-Ye Jiangchuan,
Singapore 1987) 21 ... �c8!? 22 ..ie2 d3
gives Black counterplay. Black can also
consider the immediate 20...�c8!? 21
..ie2 d3! when White must find 22 hS
(worse are 22 ..ixd3 ..ixg4, 22 �xd3
..ixg4 and 22 ..id1 ..ixb2) 22 ... ..ixg4!
(but not 22 ...dxe2? 23 hxg6 fxg6 24
'ifh3) 23 'ilfxg4 'ilfxg4 24 ..ixg4 ..ixb2 2S
2 74
..if3 ..ixcl 26 ..ixcl b4 with an unclear
ending.
b) 10 ..ih6 is fairly common and
looks more natural than variation 'a'.
Here Black can to transpose to some
other lines or explore independent
paths:
bl) 10... es 11 ..ixg7 (instead 11 dS
..ixh6 12 �xh6 ll'ld4 transposes to
variation 'b3' below, but that is dan­
gerous for Black, so he should consider
12 ... ll'las 13 ll'lcl cs, as pointed out by
Golubev) 11 ...@xg7 12 dS! ? (12 o-o-o bS
takes play into Line (32, below)
12 ... ll'le7 (or 12 ... ll'las! ? 13 ll'lg3 cs) 13
ll'lg3 c6 14 dxc6 ll'lxc6 lS o-o-o ..ie6 16
@b1 ll'le8 (Black could also consider
16 ... ll'ld4!? 17 ll'lge2 ll'lc6 18 �xd6 'ilfxd6
19 .l:i.xd6 ..ixc4) 17 ll'lds bS was
B.Spassky-R.Fischer, 8th matchgame,
Sveti Stefan 1992. White may try to
claim some an edge here, but it is not
much and in fact Black went on to win.
b2) 10 ... bs is the simplest. Now 11
o-o-o is Line C and 11 ..ixg7 @xg7 12
o-o-o es is Line C32, which are both
satisfactory for Black, as we shall see.

Instead 11 g 4 is not really sound after
11 ... hxg4 12 hS gxf3 and then:
b21) 13 lllg 3 ..ixh6 14 'ilfxh6 f2+! lS
�dl lllg 4 16 ..ifd2 es is winning for
Black (Boleslavsky).
b22) 13 hxg6 fxe2 (or even 13 .. .fxg6)
14 ..ixe2 (14 ..ixg7 exfl�+ lS @xfl
fxg6 should also win for Black) 14.. .fxg6
lS ..ixg7 @xg7 16 �h6+ 'i;f7 looks un­
sound too.
b23) 13 ..ixg7 @xg7 14 lllg 3 f2+! ?
(14...es l S hxg6 fxg6 1 6 'it'h6+ @f7 is
also good) lS �xf2 es was much better
for Black in E.Cooke-Y.Zimmerman, Bu­
dapest 2000.
b3) 10 ... ..ixh6 11 �xh6 es (instead
11 ... bs 12 o-o-o es is Line Cl, which is a
tittle risky for Black) 12 dS!? (here
White can steer things towards the
aforementioned Line Cl with 12 o-o-o
bS) 12 ...ll:ld4 (this does not work out
well, so Black should consider 12 ...llle 7
or 12 ...llla s) 13 o-o-o cs ( 13 ...lllx e2+ 14
.i..xe2 with the idea of g2-g4 looks good
for White) 14 dxc6 bxc6 (no better are
14... lllxc6 l S cs or 14 ... lllx e2+ l S ..ixe2
bxc6 16 'ilfd2!, as given by Golubev) lS
P a n n o Va r iatio n : 9 h4 h5 1 0 o - o - o
lllxd4 exd4 16 l:txd4 l:txb2.
If Black does not play this, he will
have nothing to show for the pawn.
Moreover, the temporary sacrifice looks
very appealing, but it does not work
out well against exact play:
b31) 17 @xb2 'ilfb6+ 18 lllb s axbs
gave Black good play in V.Khomyakov­
M.Golubev, Ostrava 1992.
b32) 17 es .ifs 18 ..id3 :xg2 19
..ixfs �as ! 20 ..ife3 l:te8 21 l:td2 'ilfa3+
22 @d1 :xes 23 ..ie4 l:txd2+ 24 'ilfxd2
lllxe4 2S fxe4 l:txe4!? 26 lllxe4 �f3+
was quickly drawn in R.Meessen­
M.A hn, European Club Cup, Eupen
2006.
b33) 17 cs! (Golubev first pointed
out the strength of this move) 17 ...:b8
(no better are 17 ....l:i.b7 18 l:txd6! �as
19 @c2 or 17 ..ltf2 18 ..ife3 ..Was 19 l:ta4!
..Wxcs 20 ..Wxcs dxcs 21 llld 1) 18 l:txd6
�as 19 'iii>c2 ..ie6 (19 ...'ilfa3 20 'ilfc1
should also win) 20 .l:i.xe6 fxe6 (after
20 .....ifa3 21 l:td6! :b2+ 22 'it>d3 ..Wxcs,
23 'ilff4! is probably clearer than 23 :d4
'ilfa3) 21 ..ic4 'ilf7 22 es 'ilfb4 23 exf6
..Wb2+ 24 @d3 l:tbd8+ 2 s .ids! wins.
2 75
A ttacking C he s s : Th e King 's I n di a n , Vo l u m e 1
We can now return to the main sub­
ject of this chapter:
10 0-0-0
Play becomes very sharp, but Black
can hold his own in the complications.
10...bs
Black initiates his counterplay. Note
that there are certain similarities with
the Soltis Variation of the Sicilian
Dragon (1 e4 cs 2 ll'lf3 d6 3 d4 cxd4 4
ll'lxd4 ll'lf6 S ll'lc3 g6 6 ..ie3 ..ig7 7 f3
ll'lc6 8 ..ifd2 o-o 9 ..ic4 ..id7 10 o-o-o lk8
11 ..ib3 ll'les 12 h4 h s). Here White
must decide how to proceed:
A: ulCds
• :u: lhf4,
:C: 11.i.h6
A) 11 ll'lds
With this move White removes the
f6-knight from the defence of the black
king. However, this costs time and
Black can create counterplay.
11 ... bxc4 12 ll'lxf6+
Consistent. Beating other moves
used to be a specialty of John Nunn:
2 76
a) 12 g4 ll'lxds (worse is 12 ...hxg4?!
13 ll'lxf6+ ..ixf6, which is note 'b' to
Black's 13th move in our main line) 13
exds ll'lb4 14 ll'lc3 c6 lS ..ixc4 (instead
1s dxc6 ..ie6 16 gxh s ..Was 17 a3 ll'lds
18 h6 ..if6 looks good for Black, while
here 18 ll'lxds 'ilfxds 19 hxg6 c3! is
given by Nunn) 1s...cxds 16 ..ib3 (or 16
ll'lxds ..ie6 17 ll'lxb4 ..ixc4 18 ll'lc6 'ilfc7
19 ll'lxb8 ..ixa2+ 20 ..ifc3 'ilfxb8 when
Black is much better according to
Nunn) 16 .....ifb6 (also good is 16 ...hxg4
17 h S gxf3 with the idea of ....ifs as in
A.Ewald-M.Fischer, German League
2oos) 17 :hgl (Nunn gives 17 :dgl fS,
with the idea of 18 gxh s f4, and here
17 ... hxg4! ? is also possible) 17 ... ..ie6 18
gxh s .ifs 19 l:tgs ..ih6 20 l:tdg1 ..ixgs
21 l:txgs e6 22 hxg6 fxg6 23 h s l:tb7 24
..Wg2 l:tg7 2s h6 l:tb7 26 l:txfs exfs 27
'ilfxg6+ 'iii>h 8 28 ..Wg2 f4 29 ..ig1 l:tg8 0-1
was A.Kuligowski-J.Nunn, Wijk aan Zee
1983.
b) 12 ..ih6 ll'lxds 13 exds ll'lb4 14
ll'lc3 c6 (14 ... ..ixh6!? lS 'ilfxh6 c6 is a
good alternative) looks good for Black
as well.

White has tried:
bl) lS ..ixc4 cxds 16 ll'lxds ..ixh6 17
�xh6 ll'lxds 18 ..ixds .ifs 19 ..ie4 ..ixe4
20 fxe4 'ilfc7+ 21 'iii>b l l:txb2+ 0-1 was
H.Hurme-J.Nunn, Helsinki 1983.
b2) lS dxc6 was suggested by Nunn,
but after 1S .....ixh6! (worse is lS ... ds 16
g4) 16 �xh6 ..ifb6 (with the idea of
...ll'la2) 17 �d2 ..Wxc6 Black has the up­
per hand.
b3) lS g4 �as ! (1s .....ixh6!?) 16
..ixc4 (both 16 ..ixg7 ll'lxa2+ and 16
'iii>bl ll'lxds should be winning for Black)
16 ... cxds 17 ..ib3 ..ie6 18 ..ixg7 'iii> x g7
19 'iii>bl (19 gxh S l:tfc8! 20 'iii>bl ll'lxa2!
19 l:tdel l:tfc8 20 'it>d1 ll'lc6 21 gxh s l:tb4
was much better for Black in S.Kitte­
M.Zulfugarli, Szeged 1994) 19 ...l:th8
(19 ... hxg4 and ....ifs is winning as well)
20 l:tdgl ll'lc6! (with the idea of ... l:tb4
and ...�6) 21 ..ia4 l2Ja7 22 l:th 2 l:tb4 23
..ib3 ll'lc6 24 ll'le2 ..ifb6 2s gxhs .ifs+ 26
'iii>a l l:txb3 27 axb3 ll'lb4 28 ll'lc3 ll'lc2+
29 'iii>a2 l:tb8 0-1 was M.Petursson­
J .Nunn, Lucerne Olympiad 1982.
12 ..ixf6
...
Here 12 ... exf6 looks too risky: for ex-
P a n n o Va ria t i o n : 9 h4 h 5 1 0 o - o - o
ample, 1 3 g 4 hxg4 (or 13. . .ll'lb4 1 4 ll'lc3
cs lS dxcs ..ie6 16 a3 ll'ld3+ 17 ..ixd3
cxd3 18 gxh s) 14 hS gxf3 lS h6 fxe2 16
..ixe2 and White has a strong attack.
13 g4 ll'lb4
A thematic move. The knight takes
aim at both d3 and a2, while the rear c­
pawn is freed to advance, which will
both strike at the centre and open a
path for the black queen. other moves
are ri sky but are not necessarily bad:
a) 13 ... es 14 gxh s ll'lb4 (after
14... exd4 lS ll'lxd4 ll'les 16 hxg6 cs 17
gxf7+ 'iii> h 7 18 ll'lfs? ..ixfs 19 exfs :xb2!
Black won in J.Herbert-V.Walsh, corre­
spondence 1989, but 18 f4! looks very
strong) lS ll'lc3 exd4 16 ..ixd4 ..ie6 17
hxg6 fxg6 18 'ilfh6 ..ixd4 19 'ilfxg6+ ..ig 7
20 ..Wxe6+ 'iii>h7 21 ..ixc4 gave White the
upper hand in J.Sammour Hasbun­
J.Stopa, Pawtucket 2008 .
b) 13 ... hxg4 14 hS is risky but may
be playable:
bl) 14... es lS hxg6 fxg6 16 ll'lc3
ll'lxd4? 17 ..ixd4 exd4 18 ..ixc4+ l:tf7 19
..ixf7+ 'iii>xf7 20 ll'lds ..ig7 21 fxg4 d3 22
l:th 7 �g8 23 ..Wgs 1-0 was Nguyen Anh
Dung-Wang Rui Budapest 2000.
b2) 14 ... gs!? lS ..ixg s es 16 ..ixf6
'ilfxf6 with an unclear position has been
tried with some success.
b3) 14 ... gxf3 ! ? (the 'grab everything'
policy is sometimes seen in similar
structures in the Dragon) lS ll'lc3 gS!
16 ..ixg s ..ixgs 17 'ilfxg s+ 'iii>h 7 was un­
clear in K.Cedikova-D.Vismara, Estensi
2001, although Black did go on to win.
14 ll'lc3
2 77
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
14 cs
...
Instead 14.....ie6 ! ? may be an even
better move order: lS gx hS cs 16 hxg6
�as 17 a3 (after 17 gxf7+ 'iii>xf7 18 a3
Soloviov gives 18 ... cxd4 19 ..ixd4 ll'ld3+
20 ..ixd3 ..ixd4 with counterplay, while
18 .. J:tfc8! ? 19 es?! cxd4 20 ..ixd4 dxes
21 ..ie3? ll'ld3+ 22 ..ixd3 cxd3 0-1 was
V.Quintana Femandez-D.Roiz Baztan,
Asturias 1996) 17 .. .fxg6.
Here White has several options:
a) 18 axb4 cxb4 19 ll'lbl c3 gives
Black the initiative.
b) 18 ..Wg2 'iii>f7 19 .l:i.gl .l:i.g8 20 axb4
cxb4 2 1 ll'lbl c3! 22 dS cxb2+ 23 'iii>d2
..ic3+ 24 'iii>e 2 it.d7 with good counter-
278
play according to Soloviov.
c) 18 f4 ll'lc6! 19 �g2 (after 19 dS
l:txb2! 20 'iii> x b2 :b8+ 21 'iii> c2 'ilfxa3 22
l:tb1 ll'lb4+ Black is winning and 19 es?
ll'lxd4 20 ..ixd4 cxd4 2 1 1Wxd4 dxes 22
..Wg1 .ifs 23 ..ixc4+ 'iii> g 7 24 l:.ds 'ilfc7 2 s
J:tcs �b7 gave him a winning attack in
J.Tarjan-R.Gunawan, Indonesia 1983),
and now Black must avoid 19 ...'iii>f7? 20
fS! and play either 19 .....if7 (unclear
according to Soloviov) or 19 ....l:i.f7! ? with
good counterplay.
d) 18 l:tg1 'iii>f 7! 19 l:tgs?! (19 axb4
cxb4 20 ll'lbl c3 19 ..Wg2 is variation 'b')
19 .....ixg s 20 hxg s cxd4 2 1 ..ixd4 ll'lc6
22 ..ie3 l:tb3 23 'iii>c2 .l:i.fb8 24 l:tb1 ll'les
2S ..ie2 �c7 26 f4 ll'ld3 27 ..ixd3 cxd3+
28 �xd3 �b7 29 ..icl �6 30 ..ife3
.l:i.xc3+! 31 'it>xc3 �c6+ 32 'it>d2 .l:i.h8! was
winning for Black in V.Churkin­
S.Soloviov, St Petersburg 2007.
e) 18 h S ! ? (very ambitious) 18 ... cxd4
19 ..ixd4 ..igs! (better than 19 ...it.xd4
20 �xd4 ll'lc6 21 �e3 �b6 22 �xb6
l:txb6 23 ..ie2 with a pull for White in
the ending) 20 hxg6! (critical and best;
White had to avoid 20 ..ie3? .l:i.xf3 ! 21
..ixg s .l:i.xc3+! when he loses and 20 f4
..ixf4 21 ..ie3 ..ies 22 ..id4 ..ixd4 23
'ilfxd4 �g s+ is also in Black's favour)
20 .....ixd2+ 21 l:txd2 .l:i.f6 22 axb4 ..Wal+
23 ll'lbl c3! and now 24 bxc3?! ..ia2 2S
..id3? es! left White in huge trouble in
L.Gerzhoy-S.Ganguly, Chalkis 2010, but
24 l:tdh2! cxb2+ 2S 'it>d2 ! 'ilfxbl 26 .l:i.h8+
'iii> g 7 27 l:tlh7+ 'iii> x g6 28 .l:i.h6+ would
have left Black unable to escape the
perpetual.

15 ..ixc4
This move is almost always played,
but it may not be the best:
a) 1s dxcs? is bad after 1s .....ixC3!
16 bxc3 ..Was ( 16...ll'ld3+ 17 ..ixd3 ..Was
is also good) 17 cxb4 'ilfa3+: for exam­
ple, 18 'iii> b l l:txb4+ 19 'iii>a l c3 20 'ilfc1
.:lb2 21 ..ic4 ..ie6 22 l:tds ..ixds 23
i.xds l:tfb8 with a winning attack.
b) lS a3?! ll'lc6 (Black can also play
1s ...ll'ld3+ 16 ..ixd3 cxd4 17 ..ixd4 ..ixd4
18 ..ixc4 ..ig7 with counterplay or sim­
ply 1s ... cxd4 16 ..ixd4 ..ixd4 17 'ilfxd4
-Zlc6 with a good Dragon position) 16
dxcs ll'las 17 'ilf c2 hxg4 looks good for
Black.
c) 1s gxh s ..Was (Black is probably
better off with lS ... ..ie6, transposing to
the notes to his 14th move, above) 16
..ixc4 and here:
cl) After 16 ... ..ie6?! 17 ds ll'lxa2+ (or
17 .....id7 when, rather than 18 hxg6
..ixc3 19 'ilfxc3 ll'lxa2+ 20 ..ixa2 'ilfxa2
21 ..ih6 ..Wal+ 22 'iii> c2 ..ia4+ 23 'iii>d3
'ilfxb2, White should play 18 a3!) 18
..ixa2 l:txb2 19 'ilfxb2 ..ixc3 20 'ilfb3
White had a winning position in
Pan n o Var i a t i o n : 9 h4 h 5 1 0 o - o - o
l.Krasik-D.Naroditsky, U S League 2008.
c2) 16 ...cxd4 is better, although af­
ter 17 ..ixd4 ..ixd4 18 'ilfxd4 'ilfxh S 19
l:tdgl White has some initiative.
1s ...cxd4 16 ..ixd4 'ilfc7
Or 16 ... ..ixd4 17 'ilfxd4 ll'lc6, which
was agreed drawn in the game
A.Severjuk hin-D.Skorchen ko, Kazan
2001. Black could certainly continue
here; after the queen moves, 18 ... ll'les
with the idea of ... hxg4 gives him coun­
terplay.
17 ..ib3 ..ixd4
Instead 17 ... ..ie6 18 ..ixf6 ..ixb3 19
axb3 exf6 20 gxh s (20 'iii>b 1!?) 20 .....Was
21 'iii> bl 'ilfxhs was fine for Black in
L.Oll-B.Gelfand, USSR 1984.
18 �xd4 ..ie6 19 gxhs
Alternatively, 19 ..ixe6 fxe6 (Black
threatens ... ll'lxa2+) 20 'iii> b l ll'lc6 21
�c4 'ilfb6 22 'ilfxe6+ 'iii> g 7 23 l:th2 ll'ld4
24 �ds ll'lxf3 2s l:tg2 ll'lxh4 26 l:tc2 l:tf2
27 'ilfg s 1/2-1/2 was M.Tochacek­
R.Maliangkay, correspondence 2002;
27 ...l:txc2 28 'ilfxe7+ results in perpetual
check.
19 .....ixb3 20 axb3
2 79
A ttacking C h e s s : Th e King 's I n dia n, Vo l u m e 1
20 :ifas
..
A common theme - the black queen
is able to recapture on h S by threaten­
ing ...'if al+. Instead 20 ...ttJa2+ is possi­
ble: 21 'iii>c2 l:.b4 22 'if d2 l:.fb8 23 hxg6
l:.xb3 24 gxf7+ 'iii>f8 2S l:.bl 'ifc4 26
'ifh6+ 'iii>xf7 27 'ifhs+ 'iii>f8 28 'iVh8+ (not
28 'ifh6+? 'iii>e 8 and the king escaped in
O.Haugseth-H.Theting, correspondence
1990) 28 ...'iii>f7 29 'ifhs+ with a draw in
W.Schmidt-A.Sznapik, Prague 198S.
21 'iii>b1 'ifxhs
Black was at least equal in l.Nowak­
A.Sznapik, Polanica Zdroj 198S.
B) 11 t[)f4
280
This is a strange-looking move. In
order to open the diagonal for his fl­
bishop, White prevents himself from
playing ii.h6 and lessens his control
over the d4-square.
11 bxc4
...
Even though this allows White to
immediately activate his bishop, prac­
tice has shown it to be the best move.
Instead 11 ... es 12 dxes ttJxes 13 cs (13
cxbs may also be good) 11 ...b4 12 ttJcds
ltJh7! ? 13 g4 e6 14 gxhs exds 1s cxds
gives White good compensation for the
piece. Black's main alternative to
11 ...bxc4 is 11 ...ii.d7 and then:
a) 12 cxbs axbs 13 ttJxbs (after 13
ii.xbs both 13 ...es 14 dxes ttJxes and
13 ... ttJxd4 14 ii.xd7 ttJxd7 lS ii.xd4 e s
16 ii.e3 exf4 17 ii.xf4 ttJes, as in
F.Hanssen-J.Simmelink, correspon­
dence 1987, give Black compensation
for the pawn) 13 ... es 14 dxes ttJxes 1 s
ttJc3 'ifc8 and Black had good play in
J.Timman-J.Nunn, Tilburg 1982.
b) 12 g4 ttJxd4 (12 ... hxg4!?) 13 ii.xd4
es 14 ii.xes (instead both 14 ii.a7 l:.b7
1 s ttJxhs gxhs 16 ii.e3 and 14 ttJxhs
gxh s are unclear; however, in the latter
variation, 14... exd4?! l S ttJxg7 dxc3 16
'ifxc3 'iii>x g7 17 g s was better for White
in M.Petursson-S.Los, Bela Crkva 1983)
14...dxes 1s g s b4 16 ttJcds ttJxds 17
ttJxds ii.e6 18 'if e3 ii.xds 19 l:.xds looks
a little better for White, although after
19 ... 'ife7 20 ii.h3 (20 'ifcs !?) 20 ... l:.fd8 21
l:.hd1 l:.xds 22 l:.xds l:.d8 23 'if cs ii.f8
24 'ifxe7 ii.xe7 2 s l:.xes ii.d6 26 l:.ds c6!
27 l:.d3 ii.f4+ 28 'iii>c2 l:.xd3 29 'iii>xd3 cs

30 b3 the game was drawn in H.Ree­
J.Nunn, Wijk aan Zee 1983.
12 ii.xc4
The pin of the f7-pawn on the a2-g8
diagonal will sometimes come into
play now. White should not get carried
away with 12 g4?!, though: 12 ... es 13
dxes ttJxes 14 gxh s ttJxf3 1s 'ifg2
ttJxe4! 16 ii.xc4 ttJxc3 17 bxc3 'ife8 18
'ifxf3 ii.g4 19 'if g3 'ifa4 20 'ifxg4 'ifxc4
0-1 was M.Noble-H.Garcia Rojas, corre­
spondence 2004.
12 ...es
13 dxes
The tempting 13 ttJxg6?! backfires
after 13 ... exd4 14 ii.xd4 l:.b4! lS ttJxf8
l:.xc4 16 ii.xf6 'ifxf6 when Black is bet­
ter.
13 ...ttJxes 14 ii.b3 'ife8!
This is a key move. Black removes
his queen from the d-file and defends
the es-knight, thereby making ... cs a
possibility because a subsequent 'ifxd6
will not exchange queens. Moreover,
from the e8 the queen may emerge on
the a4-e8 diagonal and the advance
... aS-a4 is supported. Black has scored
Pa n n o Variatio n : 9 h 4 h5 1 0 o - o - o
well from this position.
15 'iii>b l
This is a natural move, but it does
not work out well. White has also tried
1s ttJfdS ttJxds 16 ttJxds ii.e6 and th en:
a) 17 ttJxc7? fails to 17...'ifc6+ 18
'ifc2 (18 'ifc3 ii.xb3 19 'ifxc6 ttJxc6 20
axb3 l:.fc8 is great for Black) 18...ttJc4
with a winning attack.
b) 17 'ifas? ii.xds (17 ...'ifc6+ 18 'iii>b 1
'ifb7 is also very strong) 18 'ifxds (18
l:.xds 'ifc6+ has the idea 19 'iii>bl ltJc4,
winning) 18 ... cs! shows all of Black's
ideas in action: 19 l:.d2 c4 20 ii.c2 l:.bs
(20 ...'if c8! is even better) 21 'ifxd6 l:.xb2
22 'ifxa6? (22 'iii>xb2 c3+ 23 'iii>c l cxd2+
24 ii.xd2 was the only chance) 22 ...'ifb8
and Black was winning in H.Ree­
J.Mestel, European Team Champion­
ship, Plovdiv 1983.
c) 17 f4 ii.xds 18 exds ltJg4 is very
comfortable for Black.
d) 17 ii.h6 ii.xds 18 exds ii.xh6 19
'ifxh6 as 20 l:.d4 l:.b4 21 'ifd2 l:.xd4 22
'ifxd4 'ifbs 23 l:.dl l:.b8 gave Black
some initiative in B.Belotti-B.Zueger,
Chiasso 1991.
281
A ttacking Chess: The King 's I ndian, Vo l u m e 1
e) 17 ii.d4 ii.xds 18 exds as 19 'ifc2
'ifbs gave Black counterplay in A.Finke­
M.Spaans, correspondence 1987.
f) 17 'iii>b 1 ii.xds 18 'ifxds cs 19 ii.cl
was J.Merriman-G. Lane, London 1993,
and here both 19 ...'ifc8!? (supporting
the push ... c4) and 19 ... as deserve con­
si deration.
1s ...as
16 'ifc2
The main alternative is 16 ii.d4
tt:Jfd7 17 tt:Jfds c6 18 ttJe3 (after 18 ttJf4
ii.a6 19 ii.xes ttJxes 20 'ifxd6 ii.c4 Black
had the initiative in H.Naundorf­
J.Moucka, correspondence 2000)
18 ...ii.a6! (Black eyes the d3-square) 19
ii.xes (19 f4 tt:Jd3 20 ii.xg 7 'iii>xg7 is
much better for Black according to Gel­
fand) 19 ...ii.xes 20 g4 (Gelfand also
gives 20 ii.c4 'ifc8 21 g4 l:.b4! 22 ii.xa6
'ifxa6 with a strong attack) and now:
a) 20 ... ds!? 21 gxh s (instead 21 exds
ttJcs 22 ii.c4? loses to 22 ... ii.xc3 23
'ifxc3 tt:Ja4 24 'iff6 ttJxb2 2S 'iii>al 'ifxe3
because 'ifxg6+ is not possible, but 22
'if c2 could be tried because after 22 ...a4
23 ii.c4 ii.xc3 24 'ifxc3 Black cannot
282
play ...tt:Ja4 anymore) 2 1...ttJcs 2 2 tt:Jg4?
(Gelfand indicates that 22 hxg6 tt:Jxb3
23 axb3 l:.xb3 24 ttJexds cxds 2s 'ifxds
is a better try, although 2s ... 'ifb8 26
gxf7+ 'iii>h 7 should still win for Black)
22 ... ii.d3+ 23 ii.c2 d4 24 tt:Jds ii.xc2+ 2s
'ifxc2 cxds 26 'ifxcs .l:lxb2+ 27 'iii>xb2
�8+ with a winning attack in
V.Dydyshko-B.Gelfand, Minsk 1986.
b) 20...ttJcs! may be even stronger:
21 ii.c4 ii.xc3 (21 ... ds looks good too) 22
'ifxc3 tt:Ja4 23 'iff6 l:.xb2+ 24 'iii>al ii.xc4
2S ttJxc4 l:.c2 26 tt:Jxd6 tt:Jc3 27 'ifxC3!
was given by Gelfand, but Black can
play 26 ...'ife7! ! with the idea of 27
'ifxe7 (27 'ifd4 cs 28 'ifxa4 'ifes+ is
winning for Black as well) 21 ...tt:Jc3,
mating.
16 ...ii.d7 17 ii.d4
17...'iii> h 7!
Black removes his king from the a2-
g8 diagonal. Instead 17...tt:Jc6 18 ii.xf6
ii.xf6 19 ttJxg6 .l:lxb3 20 ttJxf8 .l:lxc3 fails
to 21 tt:Jxd7 ! - Nunn.
18 ttJd3
Instead 18 tt:Jh 3 has been recom­
mended. However, after 18 ... cs 19

ii.xes 'ifxes 20 f4 'ife7 21 ttJgs+ 'iii> g 8 22
'ifd3 a4 23 ii.xa4 ii.xa4 24 ttJxa4 'ifa7
(24....l:lb4 2S ttJc3 .l:lfb8 with a strong
attack may be even better) 2S ttJc3 ltJg4
26 .l:ld2 c4 27 'ifxc4?! .l:lfc8 28 ttJbs l:.xc4
29 ttJxa7 l:.c7 Black won a piece in
T.Emst-Y.Zimmerman, Creon 2oos.
18...ttJc6
Black could also play 18 ... ttJxd3 19
l:.xd3 ii.e6, but the text is more ambi­
tious.
19 ii.xf6
After 19 ii.e3 ltJb4 20 ttJxb4 .l:lxb4
Black clearly has the initiative.
19. �xf6 20 ttJds ii.g7 21 ii.a4
..
Forcing things with 21 ttJxc7 back-
fires after 21 ... 'ifd8 22 ttJds ltJd4 23 'ifc4
ii.bs 24 'ifc7 a4 2s 'ifxd8 .l::i.fxd8 26 ii.c2
l:.dc8 and Black wins.
21...ttJes ! 22 ii.xd7 'ikxd7 23 .l:ld2
The c-pawn is clearly untouchable:
23 'if xc7? 'ifxc7 24 ttJxc7 ttJxd3 2s l:.xd3
l:.xb2+ 26 'iii>c l l:.c8.
23 ....l:lb7 24 ttJe3 ttJxd3 25 'ifxd3 l:.fb8
Black had a very strong attack in the
game R.Vaganian-J.Nunn, Skelleftea
1989.
Pa n n o Varia t i o n : 9 h4 h5 1 0 o - o - o
C) 1 1 ii.h6
White wants to exchange the pow-
erful g7-bishop. Now 11 ... es is the main
line, but other moves are also of inter-
est. We have:
C1: 11 ii.xh6
•.•
C2: 11 bxc4
•••
C3: 11 es
..•
Instead 11 ...ttJas! ? is relatively un­
explored, leading to 12 ii.xg7 'iii>x g7 13
ttJf4 and here:
a) 13 ... c6 14 g4 .l:lh8 (lS 'ifg2 'ifc7 16
gxhs ttJxhs 17 ttJxhs+ .l:lxhs was fine
for Black in H.Hurme-J.Nunn, Helsinki
1981) 1s cxbs axbs 16 gs ltJd7 17 es b4
was unclear in Kliesch-W.Wittmann,
correspondence 1984.
b) 13 ...ttJxc4!? 14 ii.xc4 bxc4 lS 'ife2
c6 16 es?! (16 'ifxc4 looks equal) 16 ...ttJds
17 ttJfxds cxds 18 g4 (or 18 ttJxds ii.e6)
18.. lth8 19 .l::i.d2 e6 (19.. .ii.e6!) 20 f4 hxg4
21 'ifxg4 dxes (21 ...l:.hs!) 22 fxes .l::i.hs 23
l:.f2 ii.d7 24 l:.hf1 .l:lfs 2s hs gave White
compensation in U.Schuster-M.Litovicius,
correspondence 1997.
283
A ttacking C h e s s : Th e King 's Indian, V o l u m e 1
Cl) 11 ii.xh6
...
This is a controversial move. Black
invites the white queen into his own
king's den, but if the king survives, the
white queen could be missed in the
centre and on the queen side.
12 'ifxh6
12 ...es 13 d s
Instead 13 dxes ttJxes is comfort­
able for Black, while 13 g4 could be met
with 13 ...bxc4 or even 13 ... ii.xg4!?.
13 ttJas 14 ltJ g3 bxc4
...
I think Black should consider the un­
tried 14...b4! ? here. After 15 ttJb 1 ttJb7
Black can play ... as, ...ttJcs, ... ii.d7 and
... c6 with counterplay. This plan is not
so common, but it comes up some­
times and is well worth being familiar
with, especially when White's kingside
play is also rather slow.
15 ii.e2
White intends .l:ld2 and ii.d1 to tidy
up the queenside, after which he can
proceed with his kingside campaign
with ttJf1, intending ttJe3 and g4.
1s .l:.b4
..
284
We have seen this idea before in the
10 ... bs lines. Here, however, White
cannot open the h-file so he must aim
for an eventual g4-break.
16 l:.d2
After 16 ttJfl ii.d7 17 ttJe3 'ifb8 18
l:.d2 'ifb6, 19 ltJfs? is an instructive mis­
take, since 19 ... ii.xfs 20 exfs .l:lb8 leaves
Black's attack much the faster: 21 ttJd1
l:.a4 22 'iii>bl ttJb3 23 axb3 'ifxb3 24 fxg6
'ifa2+ 25 'iii>cl 'ifal+ 26 'iii>c 2 l:.ab4 27
ii.fl l:.xb2+ 28 ttJxb2 l:.xb2+ 0-1 was
A.Muir-A.Fedorov, Batumi 1999.
16...ii.d7
Instead after 16 ...cs?! 17 dxc6 ttJxc6
18 a3 l:.b7 19 ii.xc4 ii.e6 White has:
a) 20 ii.xe6 fxe6 and White cannot
play 21 'ifxg6+? l:.g7.
b) 20 ttJds ii.xds? 21 .l:lxds ltJd4
(Black tries to improve on 21 ...ttJxds 22
ttJxhs gxh s 23 l:.h 3, as in A.Schneider­
A.Stummer, Budapest 1993, but this
loses too) 22 .l:lxd4! with a winning po­
sition in F.Volkmann-A.Stummer, Bu­
dapest 1993. However, instead of
20 ... ii.xds, Black can improve with
20...ltJd4!.

c) 20 ttJfs! gxfs 21 exfs ii.xfs (the
point is that 21 ...ii.xc4 fails because of
22 l:.h3 ! ) 22 'ifgs+ 'iii>h8 (22 ...ii.g6? 23
'ifxg6+) 23 'ifxfs and White is clearly
better.
17 ttJf1 'ifb8
It seems that staying near the king
with the alternative 17...'ife7 ! ? may be
a bettertry.
18 g4 'ifb6 19 ii.di l:.b8 20 l:.hh2
Black's position looks good, but is
difficult for him to make progress. The
tempting 20 ...ttJb3+?! does not seem to
work after 21 axb3 'ifas 22 ii.c2 (22
'iii>b l!? Is an interesting alternative)
22 ...'ifal+ 23 ii.bl l:.xb3 24 l:.c2 as
(Black should try 24...hxg4!?) 2S ttJe3
and White took over the attack in
P.Neuman-Ji.Houska, Klatovy 2002.
Perhaps Black must try something like
20...c6 21 ttJe3 'ifd8 here, but if that is
the case then he probably should im­
prove earlier, perhaps on move 14 or
move 17.
C2) 11...bxc4
This is a sharp try.
Pa n n o Va riation: 9 h 4 h5 1 0 0 - 0 - 0
1 2 g4
White can also exchange bishops
immediately with 12 ii.xg7 'iii>x g7 and
here 13 ttJf4 es 14 dxes ttJxes is note 'b'
to White's 14th move in Line (32 be­
low, but he can also try:
a) 13 ds ltJb4 (13 ...ttJas is also possi­
ble, while 13 ...ttJes 14 ltJd4 cs 1s dxc6
'ifas was D.Ruzele-B.Gelfand, USSR
1986; here 16 f4 ttJeg4 17 ii.xc4 is good
for White, but earlier 14...ii.d7 looks
like a better try for Black) 14 ttJf4 c6 lS
ii.xc4 cxds 16 ttJfxds ttJbxds 17 ttJxds
ii.e6 18 'iii>b1 ii.xds 19 ii.xds as looks
okay for Black, but he should not be too
complacent: 20 l:.cl a4 21 l:.hdl 'ifb6
22 'ifd4 'ifas 23 l:.c4 a3 24 b3 l:.fc8 2 s
ii.c6 'ifa6 26 l:.dcl l:.b6 27 'ifc3 l:.cb8? !
(27 ... 'iii> g 8) 28 g4! gave White an attack
in M.Cebalo-Z.Djukic, Yugoslav Team
Championship 1983.
b) 13 g4 es 14 ii.xg7 'iii>x g7 trans­
poses to the note to White's 13th move
in our main line, which is satisfactory
for Black, and here 13 ...hxg4 is also
possible.
c) 13 ttJg3 es 14 dxes (14 ds ttJas)
2 85
A tt a c k i n g C h e s s : Th e K i n g 's I n d i a n, Vo l u m e 1
14... ttJxes 1s f4 (lS 'ifg s?! can be met
with 1s ....l:le8 or even 1s ... 'iii>h 7)
1s ...ltJd3+ (better than 1S ...ttJeg4 16 es)
16 ii.xd3 cxd3 17 'ifxd3 ii.g4 18 .l:ld2
'ifc8 is fairly level.
12 ... es
It is time to strike in the centre.
Others:
a) 12 ...'iii>h 7 wastes a tempo: 13
ii.xg7 'iii>x g7 14 ltJg3 l:.h8 and now both
1s gxh s ttJxh s 16 ttJxh s+ l:.xh s 17 ii.xc4
and lS gs ttJd7 16 ii.xc4 look a little
better for White.
b) 12 ...ttJb4 13 ltJg3 cs 14 ii.xg7
'iii> x g7 1s gxh s (instead 1s ds l:.h8 16
ii.xc4 was V.Salov-A.Khalifman, USSR
Championship, Minsk 1987, and here
Black should try 16 ...hxg4! ? with the
idea of 17 f4 as 18 a3 ii.a6 or 17 a3
gxf3! 18 axb4 cxb4 intending ...'ifc7)
1S ...cxd4 16 'ifxd4 gave White the ini­
tiative in A.Czerwonski-D.Pedzich,
Gdansk 1994.
c) 12 ...hxg4! ? 13 hS is scary but
playable: 13 ... ii.xh6 (not 13 ... gxf3 14
hxg6 fxe2 lS ii.xe2 with a winning at­
tack) 14 'ifxh 6 g s 1s 'ifxg s+ 'iii>h 7 16 es!
286
gxf3 17 exf6 exf6 18 'iff4 fxe2 19 ii.xe2
ds! 20 .l:ldg1 ttJe7 21 'ifxf6 ttJfs 22 'if es
l:.e8 23 'iff4 'ife7 gives Black counter­
play and now 24 ii.f3 ? l:.xb2! was sud­
denly winning for Black in C.Van
Buskirk-J.Peters, Irvine 1997.
13 ds
Black is well placed after 13 ii.xg 7
'iii>x g7 14 ds (or 14 gs ttJd7 with a good
game for Black) 14...ltJb4 lS ltJg3 hxg4
16 ii.xc4 and now 16 ...gxf3! 17 a3 ltJg4
18 axb4 l:.xb4, with the idea of ...'iff6,
gives Black excellent play for the piece.
13 ..ii.xh6
.
Black has several alternatives to ex­
plore as well:
a) 13 ...ltJb4! ? 14 ltJg3 hxg4 1s hs (or
1s ii.xg7 'iii>xg7 16 hs gxf3) 1s ... gxf3
looks risky, but my computer thinks
that Black is much better...
b) 13 ...ltJd4 14 ttJxd4 (14 ii.xg7
ttJxe2+ lS ii.xe2 'lii>xg 7 16 l:.dgl .l:lh8
looks solid enough for Black) 14... ii.xh6
(worse is 14... exd4 lS ii.xg7 'iii>xg7 16
'ifxd4 hxg4 17 f4 with the initiative for
White) lS 'ifxh6 exd4 16 .l:lxd4 cs? (for
better or worse, 16 ...hxg4 had to be

tried) 17 l:.d2 (good enough, but 17
dxc6! with the idea of 11... l:.xb2 18
ii.xc4! is crushing) 17 ...hxg4 18 hs gs
19 'ifxg s+ 'iii>h8 20 ii.xc4 gave White a
big advantage in T.Reilly-5.Rutherford,
Sydney 199S.
c) 13 ... ttJe7!? is an unusual move in
the Panno, but it is not bad here: 14
ltJg 3 hxg4 1s hs gxf3 16 hxg6 fxg6 17
ii.xg7 'iii>x g7 18 'tlt'h6+ �7 19 ii.xc4 ii.g4
20 ii.xa6 l:.h8 21 'ife3 'iii>g 7 was
R. Filutowski- H. Lassen, correspondence
1997. The position is complicated, but
should probably favour Black.
14 'ifxh6 ii.xg4!? 15 fxg4
Worse is lS dxc6 ii.xf3.
1s...ttJxg4 16 'ifd2 ttJf2
The position is unclear.
C3) 11...es
This is the ma in line. Black com­
bines central play with the attacks on
the wings. We h ave a further split:
C31: 12 liJds
(32: 12 hg7
Pa n n o Va riatio n : 9 h4 h5 1 0 o - o - o
White has also tried a couple of
other moves:
a) 12 ds ttJas (12 ... ii.xh6 13 'ifxh 6
takes play back into Line Cl) 13 ltJg3
b4! ? (13 ... ii.xh6 14 'ifxh6 is again Line
Cl) 14 ttJb1 is very similar to note 'a2'
to White's 13th in Line C32 and is likely
to transpose there.
b) 12 ii.gs bxc4 13 ttJds ttJxd4 14
ttJxd4 exd4 1s ttJxf6+ (lS 'ifxd4 ttJxds!
16 ii.xd8 ii.xd4 17 .l:lxd4 l:.xd8 18 exds
c3 19 bxc3 .l:le8 is at least equal for
Black) 1s ... ii.xf6 16 ii.xf6 'ifxf6 17 'ifxd4
'iff4+ 18 'ifd2 'ifxd2+ 19 l:.xd2 ii.e6 20
.l:lc2 l:.b4 and Black was much better in
R. Wunsch-1.Firnhaber, correspondence
1986.
c31) 12 ttJds
This is considered to be poorly
timed because of Black's response.
12 ...ii.xh6!
Instead 12 ...bxc4 13 ttJxf6+ (13 ii.xg7
'iii>x g7 is note 'b' to White's 13th in Line
C32) 13 ...ii.xf6 14 ds (not 14 ii.xf8 exd4!
with good play for Black) 14... ltJb4 lS
ltJc3 l:.e8 16 ii.xc4 as 17 g4! gave White
287
A tt a c k i n g C h e s s : The King 's I n dian, Vo l u m e 1
a strong attack in Lin Ta-Peng Xiaomin,
China 1991.
13 'ifxh6 bxc4 14 g4
After 14 ttJxf6+ 'ifxf6 15 d5 ttJa5
both 16 ttJc3 ii.d7 and 16 ltJg3 'iff4+ 17
'ifxf4 exf4 18 ttJe2 ii.d7 19 ttJxf4 l:.b4
look good for Black.
14 ...ttJxds
Black must avoid 14...ttJxd4? 15
ttJxd4 exd4 16 gxh5 ttJxh 5 17 ii.xc4 c6
18 l:.dg1 cxd5 19 l:.xg6+ fxg6 20 ii.xd5+
.l:lf7 21 'ifxg6+ 'iii>h8 22 ii.xf7 ttJf6 23
l:.gl ltJg4 24 .l:lxg4 'ifc7+ 25 'iii>d 1 1-o, as
in P.Hummel- E.Gufeld, Los Angeles
1998.
However, 14 ... ii.xg4! ? is a good al­
ternative to the text: 15 ttJxf6+ 'ifxf6 16
fxg4 exd4 17 'iff4 'tlt'h8 18 'ifd2 d3 19
gxh5 (19 ttJc3 hxg4) 19....l:lb7 (or
19 ... 'ifxh 5 2o ttJf4 'ifc5) 20 hxg6 fxg6 21
ttJc3 l:.fb8 and Black had the initiative
in K.Wiacek-5.Korosec, correspondence
2002.
15 exds ttJb4
Also possible is 15 ... ttJxd4 16 ttJxd4
exd4 17 ii.xc4 'iff6 18 gxh5 'ifxf3 19
hxg6 'if e3+ 20 'ifxe3 dxe3 21 gxf7+
288
l:.xf7 with equality in A.Moiseenko­
A.Cabrera, Internet 2004.
16 ttJc3 exd4 17 l:.xd4 'iff6 18 'iff4
White has to bail out into an end­
ing, as 18 l:.f4 runs into 18...ttJxa2+!
and 18 l:.d2 'ifxf3 also wins for Black.
Instead 18 'ife3 looks reasonable, but
after 18 ...hxg4 19 fxg4 ii.xg4! 20 ii.xc4
(or 20 l:.xg4 ttJxa2+ 21 'iii>c2 'iff5+ 22
l:.e4 ttJxc3 23 bxc3 l:.fe8 24 ii.g2 'ifd7!?
with good compensation) 20...l:.fe8
Black was doing well in R.Grosso­
R.Berdichesky, correspondence 1994.
18 ...'ifxf4+ 19 l:.xf4 ltJd3+ 20 ii.xd3
cxd3
The endgame was even in E.Higuita­
W.Megier, correspondence 1992.
C32) 12 ii.xg7 'iii>xg7
13 dxes
Instead 13 cxb5 axb 5 14 dxe5 ttJxe5
(or 14... dxe5) transposes to our main
line here. Others:
a) 13 d5 is not well timed and after
13 ...ttJa5 14 ttJg 3 (14 cxb5 axb5 1 5 ttJg3
ii.d7 looks quite pleasant for Black)
there is:

al) 14 ...ttJxc4 lS ii.xc4 bxc4 16 ltJfl
as 17 ttJe3 ii.a6 18 g4 (18 ttJe2 could be
met with 18 ....l:lb4 or even 18 ... ttJd7!? 19
'ifxas ttJcs) 18 ... l:.h8 19 .l:ldg1 ttJd7! and
Black had good prospects in l.Mason­
J.lrvin, correspon dence 1996.
a2) 14...b4! lS ttJb1 c6 (Black has
other good moves too, such as
1S ... ii.d7, with the idea of 16 ii.e2 ii.a4!
17 l:.del ii.b3!, 1 S ....l:lb6 16 'ifgs ttJb7 17
ttJd2 ttJcs and 1S ...ttJd7 16 ii.e2 ttJcs 17
'ife3 'iff6 18 ltJfl 'iff4 19 g3 'ifxe3+ 20
ttJxe3 fs, which was better for Black in
M.Ostreic h-P .Pflichthofer, correspon­
dence 1993) 16 'ifgs cxds 17 exds ttJb7
18 ii.d3 ttJcs and again Black was bet­
ter in A.Gual Pascual-M.lllescas Cor­
doba, Ayamonte 2002.
b) 13 ttJds bxc4 14 ttJxf6 (14 ttJe3 is
safer) 14...'ifxf6 lS dS and here:
bl) 1S ... c3!? 16 'ifxc3 (16 ttJxc3 ltJd4)
was O.Reeh-0.Dannevig, Gausdal 1992.
Here 16 ...ltJb4 17 'iii> b l as would assure
Black of queenside counterplay.
b2) 1S ... ttJa7 16 ttJc3 ttJbs 17 ii.xc4
'ife7 (17 ... ii.d7!?) 18 ttJe2 l:.d8 19 'ife3 cs
20 dxc6 'ifc7 21 .ids ii.d7 22 'iii>bl ii.xc6
P a n n o Va ri a t i o n : 9 h 4 h 5 1 0 o - o - o
23 l:.cl ttJa7 was J.Rowson-J.Nunn, Ox­
ford 1998, and here 24 ttJc3 looks a lit­
tle better for White.
b3) 1S ...ltJb4 16 ttJc3 ltJd3+ (Hazai's
suggestion of 16...c6 is also interesting)
17 ii.xd3 cxd3 18 'ifgs ii.d7 19 .l:lxd3
l:.b7 20 'iii>b1 'ifxgs 21 hxgs f6 22 gxf6+
'iii>xf6 gave Black counterplay in the
endgame in J.Rowson-J.Gallagher, Brit­
ish Championship, Torquay 2002.
13 ttJxes
...
Black can also play 13 ... dxes and
here:
a) 14 'ife3 'ife7 lS ttJds ttJxds 16
cxds ttJas 17 ttJc3 (17 'iii>bl ltJc4 18 'ifc3
might be a better try) 17 ... ttJb7 18 ii.d3
'if cs 19 'ifxcs ttJxcs 20 b4 ttJb7 21 'iii>b2
ii.d7 22 .l:lcl .l:lfd8 23 l:.hdl c6 24 dxc6
ii.xc6 was level in M.Wiedenkeller­
N.Stewart, German League 2001.
b) 14 'ifg s 'ife7 and now:
bl) lS ttJds ttJxds 16 exds (after 16
cxds Black can play either 16 ... 'ifxgs+
17 hxgs ttJas 18 ttJc3 ttJb7 or 16...f6 17
'ifd2 ttJd8 and ...ttJf7) 16.. .f6 17 'ifd2
l:.d8! 18 g4 bxc4! 19 ttJc3 (or 19 gxhs
ltJb4 20 ttJc3 .ifs 21 ii.xc4 'ifcs 22 'ife2
289
A tt a c k i n g C h e s s : Th e King 's I nd i a n , Vo l u m e 1
ii.d3! 23 ii.xd3 ttJxa2+ 24 'iii>c2 .rf.xb2+ 2S
'iii>xb2 ttJxc3 with a winning attack)
19 ...hxg4 20 ii.xc4 ltJd4 21 fxg4 .1'.xg4
22 l:.dfl l:.b4! 23 hs .rf.xc4 24 hxg6
.l:lxc3+ 2S 'ifxc3 ttJe2+ 26 'iii> c2 ttJxc3 27
.l:lh7+ 'iii>x g6 0-1 was J .Mestel-E.Gufeld,
Hastings 1986/87.
b2) lS ttJg3 ! bxc4 16 ii.xc4 .l:lh8?
(better tries are 16 ... 'ifcs 17 .ids .rf.b6
and 16 ... 'iii>h 7!?) 17 ttJfs+ ii.xfs 18 exfs
with a clear advantage for White in
A.Ostrowski-Y.Zimmerma n, Oberwart
1998.
Returning to 13 ...ttJxes:
14 cxbs
Again there are alternatives to ex­
plore:
a) 14 cs? b4 1s ttJds ttJxds 16 'ifxds
l:.bs and Black won a pawn in
P.Marttinen-A.Pankratov, correspon­
dence 2002.
b) 14 ttJf4 bxc4 lS ii.e2 (even worse
are lS 'ifd4?! ttJfd7 16 ii.xc4? cs and lS
ttJfds ttJxds 16 ttJxds ii.e6 17 ttJe3?!
'iff6, as in M.Myc-B.Socko, Sopot 1997)
1s ...ii.d7 (1s .. :.rf.h8 has been played fre­
quently as prophylaxis against the g4-
290
advance, but this i s not really neces­
sary) 16 'iii>bl (after 16 g4? hxg4 17 hS
both 11... gxf3 and 17... gs are good for
Black) 16 ... ii.c6 17 'ifd4 'ifc8 18 ttJcds
ii.xds 19 ttJxds ttJxds 20 exds 'ifts+ 21
'iii>al 'iff6 and Black was better in
B.Popelka-P.Boukal, correspondence
1991.
c) 14 ltJd4 b4 1s ttJds cs 16 ttJb3
ttJxds 17 'ifxds and now Black can play
11 ....l:lb6 18 ttJxcs l:.c6 19 ttJb3 ii.e6 20
'ifd2 ii.xc4 21 ltJd4 l:.c8 with equality,
or 17 ... ii.e6 18 'ifxd6 'ifxd6 19 l:.xd6
ii.xc4 20 ii.xc4 ttJxc4 21 l:.xa6 .l:lfd8 with
compensation for the pawn in
E.Soloz hen kin-W.Sc hmidt, Katowice
1991.
14. ..axbs 15 ttJf4
1s ... b4
Black could also play 1s ... ii.d7 16
ii.e2 and then either 16 ... .l:lh8 or
16 ...l:.a8, but there is no need to hesi­
tate and 1S ...b4 has scored well for him.
16 ttJcds ttJxd s
There is little point in giving White
an extra option with 16 ... cs 17 ttJe3! ?.
17 ttJxd s cs 18 f4

Instead 18 ttJe3?! ii.e6 (18 ... l:.a8! ?)
19 'ifxd6 (19 f4 ltJg4) 19 ... 'ifxd6 20 l:.xd6
ii.xa2 gave Black a pleasant ending in
O.Sande-V.Zagorovsky, correspondence
1981.
However, 18 .�e2 is a sharp alterna­
tive. Black seems to have at least a
draw after 18 ....l:la8 (other possibilities
are the immediate 18... ii.e6 and
18 ... 'ifa5) 19 f4 (19 'iii>bl can be met by
19 ...ii.e6 or 19 ... 'ifa5) 19...ttJg4 20 f5
.'Of2 21 f6+ 'iii>h7 22 .�xh 5 (or 22 ii.C4
.!Llxhl 23 .l:lxhl ii.a6 !) 22 ...ttJxhl 23 'ifg5
(23 .l:lxhl runs into 23 ... .l:lxa2 ! ) 23 ... ltJg 3
(after 23 ... .l:lxa2 24 ii.xg6+ fxg6 25 h5
:.a1+ 26 'iii>c2 'ife8 27 .l:lxal 'ifxe4+ 28
itd2 'ikd4 it's a draw) 24 ii.xg 6+ fxg6 25
0,e7 'ife8 26 h5 ttJxh 5 27 .l:lh1 'iff7?! 28
!!:lxg6 'ifc4+ 29 'iii>d2 'ifd4+ 30 'iii>cl 'ifc4+
31 'iii>d2 'ifd4+ 32 'iii>cl was drawn in
P.Murdzia-P.Hanko, Krakow 2003.
However, going back, 27 ... .l:lxf6 28 'ifxf6
:.xa2 looks to be winning for Black.
18 ...ttJg4
This is considered safer than
18 ... ttJc6, but that move is probably
okay too if Black is careful: 19 f5 ltJe5
Pa n n o Var i a t i o n : 9 h4 h5 1 0 o - o - o
(both 19 ...ltJd4 and 1 9....l:la8 are inter­
esting alternatives) 20 g4 ttJxg4 21 ii.e2
ttJf2 (21 ...ttJf6! ?) 22 f6+ 'iii>h 7 23 ii.xh5
ttJxhl 24 .l:lxhl l:.e8? 25 'ifg 5 l:.e5 26
ttJe7! was winning for White in M.Rivas
Pastor-J.Mestel, Marbella 1982. How­
ever, Black has a clear improvement in
24 ... b3! 25 a3 'ife8! 26 ttJe7 'ifa4! when
he will be able to exchange queens
with ... 'ifc4-c2+.
19 ts
Both 19 ii.d3 and 19 .�c4 can be met
with 19 ...ii.e6.
19 ... gxfs!?
This weakens Black's kingside, but it
prevents the advance of White's pawn
to f6 and does win a pawn. Instead
19 ... .l:le8!? may also be possible.
20 ii.d3
White continues to play for the ini­
tiative. Instead 20 ttJf4 .l:lh8 21 'ifxd6?!
'ifxd6 22 l:.xd6 ttJf2 23 l:.g l fxe4 gave
Black the better endgame in
A.Czerwonski-K.Chorfi, correspondence
1998.
20 .fxe4 21 ii.xe4 .l:lh8 22 .l:ldfl ii.e6
••
23 ttJf6!?
291
A ttacking C h e s s : Th e King 's Indian, Vo l u m e 1

This simplifies to an equal ending, H.Simon-K.Harald, correspondence

but Black's pieces are well placed and it
is difficult for White to find any real
way to attack
23 ...ttJxf6 24 'if gs+ '>W8 25 l:.xf6 ii.xa2
26 'iff4 'ife7 27 l:.xd6 l:.e8 28 .ids
Instead 28 l:.h6 'iii>g7 29 l:.xh8 .l:lxh8
30 'ifg 3+ was agreed drawn in
2000.
28 ... ii.xds 29 .l:lxd s 'ife3+ 30 'ifxe3
l:.xe3 31 �xcs l:.e4
Black had no problems at all in
J .Griffin-K.Harma n, correspondence
1992, and he even managed to eventu­
ally win.

292
 Cha pter 16
Pan no Variation
9 h4 h5 10 �C1
1 d4 ttJf6 2 c4 g6 3 ltJc3 ii.g7 4 e4 d6 5
f3 o-o 6 ii.e3 ltJc6 7 ttJge2 a6 8 'ifd2 .l:lb8
9 h4 hs 10 ttJc1
White seems to 'change his mind'
and switches to positional play. This
method of play has obvious similarities
with the 9 ttJc1 of our next chapter, but
the advance of both h-pawns will often
favour one side or the other.
10 es
...
This is the thematic move - Black
strikes at the newly weakened d4-
square. Still, Black seems to have some
difficulties in this line and he may wish
to consider avoiding what follows.
Smirin has experimented with 10 ... ttJe8
11 ttJb3 fs, but this looks very loosening
to me. I think Black could try waiting
with 10...ii.d7! ?.
Here White has:
a) 11 ii.e2 es looks good, as 12 ds
ltJd4 now hits the e2-bishop and should
be okay for Black, while 12 ttJb3 exd4 13
ttJxd4 bs! gives him counterplay.
White's h4-advance looks out of place.
b) 11 ttJb3 is more testing: 11 ... l:.a8 ! ?
(instead 11. . .as! ? 12 d s ! ttJes 13 ttJxas
cs 14 ttJb3 �6 lS 'ifc2! does not look
293
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

quite sufficient) looks silly but makes

some sense. After 12 ii.e2 as 13 o-o a4
14 ttJc1 es Black has counterplay. Per­
haps 12 o-o-o is more critical when
Black should still play 12 ... as. This idea
could use some tests.
11 ds ttJd4
Black has experimented with
11 ... ltJe7, although not with much suc­
cess.
12 ttJb3
Instead 12 ttJ1e2 does not make We will explore:
much sense when compared to the 9
ttJc1 es 10 ttJb3 lines because the addi­ A: 12 ttJxb3
•..

tion of the moves h4 and ... h s does not B: 12 cs

•••

work well for White here. Black should

play 12 ... cs 13 dxc6 ttJxc6 (instead A) 12 ...ttJxb3 13 axb3 cs
13 ... bxc6 is Line B below). After 14 .l:ldl Black tries to keep the queenside at
ii.e6 1s b3 (or 1s ttJds bS) 1s ...bs (also least somewhat closed. No better is
possible is 1s ...'ifas 16 ttJds and now, 13 ...c6 14 ii.e2 cxds 1s cxds with an
instead of 16 ... 'ifxd2+ which should edge for White.
leave White a bit better in the end­
game, Black could try 16 ... 'ifa3 ! ?) 16
cxbs (16 'ifxd6 'ifxd6 17 .l:lxd6 ltJb4
gives Black reasonable compensation)
16 ... axbs 17 'ifxd6 'ifc8! Black had good
play for the pawn in E.Knol­
R.Wojtkowiak, correspondence 2002.
After 12 ttJb3 Black has an impor­
tant choice. He can simplify the posi­
tion or he can play a thematic pawn
sacrifice. Unfortunately, neither of
these seems completely satisfactory for Black has trouble in this structure
the second player, although he is not after ... cs because the inclusion of the
without his opportunities after the moves 9 h4 and 9 ...h s makes it difficult
sharp pawn sacrifice, 12 ... cs. for him to create play on the kingside:
... ttJh s is no longer possible and ... fs is
(seefollowing diagram) very risky. White, however, can play on

294

either side of the board. His most dan­
gerous plan is to prepare g2-g4.
14 ii.e 2
This flexible move seems best.
White has also tried to initiate play on
the queenside immediately with 14 b4
and here:
a) 14...b6 1s bs axbs 16 cxbs ttJe8
17 ii.d3 (or 17 b4 fs) 17.. .fs 18 'ifc2 'ife 7
(18 .. .f4 19 ii.f2 ii.f6 looks like a better
try) 19 'iii>dl 'iff7 (again, 19 .. .f4!?) 20
'iii>c1 ttJf6?! 21 exfs gxfs 22 ii.xfs ttJxds
23 ii.g6 'ife6 24 ttJxds 'ifxds 2s .l:ld1
'ife6 26 ii.e4 ii.b7 was L.Portisch­
J.Nunn, Szirak lnterzonal 1987, and
now 27 .l:la7 ii.xe4 28 'ifxe4 would give
White a comfortable advantage.
b) 14... cxb4 lS ttJa4 b6 (instead
1s ...bs 16 cxbs axbs 17 'ifxb4 ttJe8 18
.ie2 looks very nice for White and here
I do not trust 16 ... as either) 16 'ifxb4 as
17 'ifd2 ttJd7 18 ii.d3 ttJcs 19 ttJxcs bx cs
20 .l::i.xas .l:lxb2 21 'ifxb2 'ifxas+ 22 ii.d2
1l'd8 was level in F.Lucas-J.Gallagher,
Bargteheide 1988.
14...ii.d7
There is nothing much better:
a) 14 ... b6 1s ttJd1 ! ? (1s o-o-o bs)
1s...as?! 16 ttJf2 gave White a big stra­
tegic advantage in L.Christiansen­
T.Magar, Philadelphia 1990.
b) 14 ...ttJe8 lS g4 fs? ! (a misguided
plan; better choices are 1s ... hxg4 16
fxg4 ttJf6 11 ii.gs bs and 1s ... ttJf6 16
ii.gs hxg4 17 fxg4 bs) 16 exfs gxfs and
here both 17 ii.gs and 17 gxfs ii.xfs 18
ttJe4 look good for White.
15 ttJd1!
Pa n n o Va riation: 9 h4 h5 1 0 ttJc1
This allows White to play b4 if he
wishes, but more importantly i t pre­
pares ttJf2 and g4.
1s ...'ikb6
Instead 1s ... ttJh7 16 g4 'ifc8 16 b4 b6
17 ttJf2 ttJe8?! (this is not very good, but
it is hard to suggest an improvement)
18 g4 fs?! (after 18 ... ttJf6 both 19 ii.gs
and 19 l:.gl prepare the g4-advance) 19
gxfs gxfs 20 exfs ii.xfs 21 .l::i.g 1 gives
White a clear advantage.
16 ttJf2
White sacrifices a pawn for the ini­
tiative. Also possible is 16 b4! ? 'ikxb4 17
'ifxb4 cxb4 18 ii.d2 when White will
quickly regain the pawn with an edge
in the ending.
16 ...'ifxb3 17 ttJd3 l:.fc8 18 o-o ii.es 19
.l:la3 'ifb6 20 b4
White had excellent compensation
for the pawn in L.Christiansen-J.Nunn,
German League 1988.
B) 12 ...cs
This pawn sacrifice is Black's main
resource, but there are still some prob­
lems to solve.
295
A ttack ing C h e s s : Th e King 's I n dian, Vo l u m e 1
13 dxc6 bxc6
Instead 13 ...tt:Jxb3 is well met by 14
c7! 'ifxc7 lS axb3 ii.e6 16 ii.e2. This is
perhaps not as bad for Black as it is
without 9 h4 and 9...hs thrown in, but
White still holds an edge.
14 tt:Jxd4 exd4 15 ii.xd4
For the pawn Black has some open
lines, a slight lead in development and
the possibility of opening the centre
with ... ds. At first it seems that the in­
clusion of the moves 9 h4 and 9 ...h s
should help Black because the white
king does not have an obviously secure
shelter. However, analysis shows that
this is not necessarily the case.
15...:.es
This is the most common move, but
of the alternatives Black should pay
close attention to variation 'd':
a) lS ...cs 16 ii.e3 ii.e6 17 b3 gives
White a clear advantage.
b) 1s ... ds 16 cxds cxds 17 es tt:Jd7
(instead 17 ....l:le8 18 ii.e2 transposes to
Line Bl below) 18 f4 f6 19 e6 l:.e8 20
o-o-o tt:Jb6 21 fs (or 21 g4, as in
M.Massimini Gerbino-K.Sakai, corre-
296
spondence 2003) 2 1... gxfs 2 2 'iff4 ii.xe6
23 ii.xa6 was much better for White in
S.Giannetto-J.Van Unen, correspon­
dence 2003.
c) lS ...'ifas 16 l:.c1 (White could also
try 16 'ifg s !? which forces 16 ...'ifd8)
16 ...l:.e8 (also unsatisfactory is 16 ... ii.e6
17 b3) 17 ii.e2 ds 18 cxds cxds 19 es
ttJd7 20 f4 tt:Jf8 21 o-o tt:Je6 22 ii.f3
tt:Jxd4 23 'ifxd4 .l:lb4 24 'ifd2 'ifb6+ 2S
l:.f2 ii.e6 26 b3 ii.h6 27 g3 .l:ld8 was
l.lbragimov-J.Sammour Hasbun, Phila­
delphia 1992, and here 28 tt:Ja4 intend­
ing tt:Jcs is the most convincing.
d) 1s ... ii.e6! ? is perhaps the best al­
ternative and it may be this move
which resurrects Black's chances with
12 ...cs.
Following 16 ii.e2 (perhaps White
should look into 16 l:.c1! ?) 16 ... ds 17
cxds cxds 18 exds (18 es tt:Jd7 19 f4 f6
20 exf6 tt:Jxf6 gives Black some coun­
terplay) 18 ... tt:Jxds 19 ii.xg7 'iii>xg7 20
l:.dl ttJxc3 21 'ifxc3+ 'iff6 22 'ifxf6+
'iii>xf6 gives Black good chances to hold
the ending. Indeed, after 23 b3 l:.fc8 (or
23 ... as! ? 24 .l:ld2 l:.b4 2s �2 a4 26 bxa4
 Pa n no varia t 1. o n 9 h 4 h5 1 0 ttJc1
·
·

.l:lxa4 27 .l:lal an d V2-V2 was J . Adler­

E Buk1c,· Bled 1992 ) 24 'iii>d2 as 2S .l:lcl
� d8+ 26 'iii>c3 .l:lb4 27 l:.bl l:.c8+ 2 8 'iii>d2
a4 an d V2-Vi, Black had held his
gran dmaster opponent in
.
J.FedorOWlCZ- E . Ronneberg, Chicago
199S.
16 ii.e2
After 16 o-o-o 'ifa S Black at least has
counterp] aY but 16 l:.d1 11 is a reason-
· ·

'
able altemative. Foll 0wing 16 ...d s 17
cxds cxds 18 es ltJd7 (instead 18 ...ttJ h7
19 f4 f6 20 ii.e2 fxe s 21 fxes ii.xe s 22
This is the mo st common and was
o-o was much better for White in
Karpov's c ho1·ce but may not be best.
K.Chorf-F.Balabaev, correspondence '
lndee d, Black even has two options:
2001) 19 f4 ii.h6 20 'iff2 ttJf6 21 g3 ii.g4
22 .l:ld2 'ifc8 23 ii.e2 ttJe4 24 ttJxe4 dxe4
2S o-o ii.f8 26 ii.xg4 'ifxg4 27 l:.c2 l:.ed8
Black h ad some counterplay in V.Pere-
. .
vertkin-N.Patnc1, correspondence 2 oos .
811) 17 ... cxds
16...d s
The most natural and obvious move
is to s1mp
. ly recapture.
18 es ltJd7 1 9 f4 .iL
� h61 •

.
Black must stnke wh 1"le the white
. .
kmg lS still in the centre. Here Wh1·te
has two options:
Worse is 19 .. .f 6 as Black's kingside
would then be very weak . However,
81:>fi7.mls
after the text Black threatens to cap-
ll: t7 e5 . .
ture on es.

297
A ttacking Chess: The King 's In dian, Vo l u m e 1
20 'ife3!
Instead 20 ttJxds ttJxes 21 ii.c3 ii.b7
gives Black good play, while 20 ii.e3
runs into 20 ...ttJxes! 21 fxes ii.xe3 22
'if xe3 d4 when Black is doing well.
20 ..Jlxb2 21 .l:ld1
After 21 o-o ttJxes! 22 ii.xes .l:lxes 23
'ifxes ii.g7 24 'ife3 (not 24 'ifxds ?
ii.d4+) 24... ii.xc3 25 'ifxc3 .l:lxe2 Black
has good compensation for the ex­
change.
2 1...ttJf6
Now 21 ...ttJxe s?! does not work be­
cause of 22 ii.xes l:.xes 23 'ifxes ii.g7
24 l:.xds (24 'ife3 is also good) 24...�6
25 'ife8+ 'iii>h 7 26 ltJe4!.
22 0-0
White should avoid 22 ii.f3 ? ltJg4 23
'ifcl ttJxes and Black also has good play
after 22 'ifcl .l:lxe2+! 23 ttJxe2 ii.g4 24
l:.d3 ii.xe2 25 'iii>xe2 'ifd7!.
22 ...ttJg4 23 'ifg3 il.e6 24 ii.xa6 'ifas 25
ii.d3
2 s...l:.c8!
Black plays as actively as possible.
This move was first pointed out by
Knaak. Instead 2s ...ii.f8? 26 fs gave
298
White a strong attack i n the well­
known game A. Karpov-S.Kindermann,
Baden -Baden 1992.
26 ts
Instead 26 ttJxds ii.xds 27 ii.xb2
�6+ 28 .l:lf2 ii.f8! 29 ii.e2 ttJxf2 30
'ifxf2 'ifxb2 31 l:.xds was R.Alonso­
J .Butterworth, correspondence 2002.
Here 31 ...l:.c2! 32 'iii>f l ii.cs ! would have
been very strong, since 33 .l:lxcs 'ifcl+
34 'ife 1 'ifxf4+ wins for Black.
26 ... gxfs 21 ii.xfs
If White plays 27 ttJe2 l:.xa2 28 ttJf4
Black has resources, such as 28 ... 'ifa4!
29 ii.xfs ? ii.xfs 30 ttJxds .l:lc1!.
21 ....l:lxc3
Thi s brings Black full equality, but
27 ... l:.b4!? is also possible.
After the text, White has no advan­
tage. For example:
a) 28 ii.xc3 ii.e3+ 29 'if xe3 .l:lxg2+!
30 'iii>h l l:.h2+ 31 'iii>g 1 .l:lg2+ is a draw.
b) 28 'ifxc3 ii.e3+ 29 'iii>h l ttJf2+ (or
29 ... 'ifd8 when 30 'ife1 is forced and
after 30...ttJf2+ 31 l:.xf2 'ifxh4+ 32 ii.h 3
ii.xf2 33 ii.xf2 'ifxf2 34 'ifxf2 l:.xf2 35
.l:lal d4 Black is fine) 30 l:.xf2 'ifxc3 31

ii.xc3 l:.xf2 and Black's activity gives
him at least enough for the pawn.
812) 17 l:.b4!?
...
This is a very enterprising move.
Black is willing to sacrifice more mate­
rial in order to create tactical complica­
tions as quickly as possible.
18 ii.cs
This seems to be the best move, but
White has tried other things as well:
a) 18 es is risky and after 18 ... ttJxds
19 ttJxds .l::i.x d4 20 'ikxd4 ii.xes 21 'if cs
kg3+ 22 'iii>fl cxds Black has good
compensation for the exchange. Then
23 l:.d1 'ikf6 24 'ifxds 'ifxb2 2s 'ifd2
"i!fes led to a black win in S.Jardorf­
U.Rinkis, correspondence 1993.
b) 18 ii.f2 cxds (here 18 ...ttJxe4? fails
to 19 fxe4 ii.xc3 20 'ifxc3 l:.bxe4 21 o-o
.:.xe2 22 dxc6 with a large advantage)
19 l:.d1 'ifas 20 ttJxds (20 o-o dxe4 21
a3 l:.b7 22 ttJxe4 'ifxd2 23 ttJxf6+ ii.xf6
24 �d2 l:.xb2 is equal) 20 ...ttJxds ttJxds
21 a3 .l:la4 (Black could also play
21 ... ii.h6 22 'ifxds 'ifxds 23 l:.xds .l:lxb2
with compensation, while 21 ....l:lxb2!?
Pa n n o Vari a t i o n : 9 h 4 h5 1 0 tLlc1
22 'ikxas ii.c3+ 23 'ifxc3 ttJxc3 24 l:.d2
.l:lbl+ 2S ii.d1 ttJxe4 26 fxe4 .l:lxe4+ 27
l:.e2 ii.g4 28 o-o is interesting but
probably better for White) 22 'ikxas
.l:lxas 23 .l:lxds l:.xds 24 exds ii.xb2 re­
sults in an equal endgame.
c) 18 ii.xf6 ii.xf6 19 o-o-o and here:
cl) 19...'ifas is risky, but playable: 20
a3 .l:lb8 21 f4 (21 dxc6 .l:le6!) 21 ...ii.xc3
22 'ifxc3 'if xc3+ 23 bxc3 .l:lxe4 24 ii.f3
.l:lc4! 2S dxc6 .ifs 26 'iii>d2 was drawn
here in M.Massimini Gerbino-R.Alvarez,
correspondence 2003.
c2) 19 ...cxds is an easy draw accord­
ing to Alvarez. It is not quite so simple,
as 20 ttJxds ii.xb2+ 21 'iii>c2 l:.b8 22 .l::i.b1
ii.e6 23 .l:lxb2 ii.xds 24 exds (not 24
.l:lxb8 'ife7+!) 24 ... .l:lxb2+ 2S 'iii>xb2 'ikb6+
26 'iii>al (Alvarez gives 26 'iii>a 3! ? .l:le3+
27 'iii>a4 'ifcs 28 'ifb2 'ifxds 29 ii.xa6
'ifd7+ with a draw) 26 ...'ikf6+ 27 'ikb2
'ifxb2+ 28 'iii> x b2 .l:lxe2+ 29 'iii>c3 �8 30
.l::i.dl 'iii>e 7 31 .l::i.d2 l:.xd2 32 'iii>xd2 'iii>d 6 33
'iii>e 3 'iii>x ds 34 �4 f6 3S g4 looks strong,
but then 3S ...hxg4 36 fxg4 'iii>e6 37 h s
�7 3 8 h6 'iii>g8 39 g s t s 40 'iii>e s 'iii> h 7
does indeed after all look like a draw.
299
A ttacking C h ess: Th e Kin g 's I n dian, Vo l u m e 1
18 ... lLixe41
Without this resource Black would
be busted.
19 fxe4 .Jlxc3 20 "ifxc3 �bxe4
Black wins back the piece and an
unusual opposite-coloured bishop
situation arises.
21 0-0
White can also play 21 0-0-0, but af­
ter 21 ...�xe2 Black has at least a draw:
22 dxc6 (or 22 d6 .Jlfs 23 �hel "ifxh4 24
J:xe2 �xe2 2S d7 "iff4+ 26 .Jle3 "ifxe3+
with a draw) 22...°ifC7 23 .Jld6 "ifb6 24
�hel .Jlfs 2S �xe2 �xe2 26 c7 was
L.Hesselberg-A.Serov, correspondence
2007. Here the simplest is 26 ...°ife3+ 27
"ifxe3 �c2+ 28 �bl J:d2+ with a draw.
21 ... �xe2 22 dxc6
Black is better after both 22 d6?!
"ifxh4 and 22 J:adl? cxds 23 "iff3 °ife7
24 .lld4 .llfs.
22 ..."ifxh4
Black should certainly take the
pawn. Worse are 22 ... "ifds 23 .Jlf2 and
22 ... .llfs 23 .llf2.
23 �f2
23 ...�2e6
300
This looks best. The advanced white
c-pawn is dangerous, but it need not be
fatal if Black stays active. Instead
23 ...�xf2 ?! 24 .Jlxf2 °ifg4 2S "iff6 "ife6 26
.Jld4 "ifxf6 27 .Jlxf6 .Jlg4 28 C7 gives
White an advantage, while 23 ...°ifgS?!
24 .llafl �xf2 2S .Jlxf2 also looks better
for White. Black can probably hold after
23 ..."ifd8: for example, 24 �xe2 �xe2 2S
�el �xel+ 26 "ifxel .Jle6 27 "ifes "ifds
28 "if xds .Jlxds 29 C7 .llb7, but it will
only be after some suffering.
24 J:af1 fs 25 .Jld4 "ife4
Black stays as active as possible. He
also managed to draw after 2S ... �h7
26 b4 °ife4 27 �C2 �d8 28 .Jlf2 �d3 29
"ifc4 "ifxc4 30 �xc4 J:e2 31 a4 �b3 32
J:e1 J:xel+ 33 .Jlxel �bl 34 �2 gs 3S
.lld4 f4 36 J:d8 J:b2+ 37 �1 .llfs 38
�h8+ and V2-V2 in K.Pecotic-F.Rubio
Dobias, correspondence 20os.
26 �c1 f4
Here a draw was agreed in J.Camp­
bell-F.Dutu, correspondence 2003. After
27 C7 gs Black has counterplay.
e2) 11 es

This move is very dangerous, but it
has not been seen much This is proba­
bly due to the fact that it has become
well known that this line causes Black
problems and thus the second player
has been avoiding it altogether.
11 ...lLid7
Instead 17... cs is creative, but looks
insufficient. White has:
a) 18 .i.xcs lLid7 19 .i.d6 lLixes!? 20
.i.xb8 lLixc4 21 °ifcl .i.fs was suddenly
agreed drawn in J.Lautier-1.Smirin,
European Club Cup, Rethymnon 2003.
White is up a rook, but Black has defi­
nite counter-chances.
b) 18 exf6 was suggested by Hazai.
After 18...cxd4 19 fxg7 dxc3 20 bxc3
�xg7 21 cxds °ifb6 22 �fl .i.fs 23 g3 it
is hard to believe that Black can prove
he has enough for the two pawns.
18 e61
White will exchange the dark­
squared bishops, castle queenside and
break with g2-g4. This move also mis­
places Black's pieces and is much
stronger than 18 f4 .i.h6! 18 cxds
lLixes.
18...J:xe6
Instead 18 ...fxe6 19 .i.xg7 �xg7 20
o-o-o °iff6 21 g4 hxg4 22 fxg4 was also
P a n n o Varia t i o n : 9 h4 h5 1 0 lLic1
much better for White in J.Lautier­
M.Golubev, Internet (blitz) 2004.
19 .i.xg7 �xg7 20 o-o-o
20...lLif6 21 g41 .i.b7
This is not a nice move to make, but
21 ... hxg4 is met by 22 h S with a strong
attack.
22 gs lLieB 2 3 .i.d3 "ifb6 24 lLia4
Instead 24 �hel looks better for
White.
24..."ifc7
Here 24...°ifd4 is a better try.
2s lLics
White has a clear advantage and
won after 2s ...�e7 26 �hel �g8 27
�bl lLid6 28 �xe7 "ifxe7 29 �el °ifC7 30
lLie6 fxe6 31 �xe6 lLie8 32 �xg6+ and
1-0 in R.Molo-P.Diaz Sjostrom, corre­
spondence 2006.
301
 Chapter 17
Pa nno Variation

9 !t:Jc1

1 d 4 l2Jf6 2 C 4 g6 3 l2Jc3 .Jt.g7 4 e4 d6 5

f3 o-o 6 .i.e3 lLic6 7 lLige2 a6 8 'ifd2 .rf.b8 A: 10 �b3
9 lLicl B: 10d5

A) 10 lLib3
This move is not very popular, partly
because White spends a long time ma­
noeuvring with his knight.
10 exd4 11 lLixd4
...

Instead of playing for a direct at­

tack, White simply prepares to con­
tinue his development. There are some
similarities between g lLicl and 9 h4 h S
1 0 lLicl, but very often the advances of
the h-pawns radically change the play.
9 es
... If we compare this to the Glek
Of course Black strikes at the weak­ Variation of the Classical (1 d4 lLif6 2 c4
ened d4-square. Now White can either g6 3 lLic3 .i.g7 4 e4 d6 5 lLif3 0-0 6 .i.e2
maintain the tension or advance in the es 7 o-o exd4 8 lLixd4 �e8 9 f3 lLic6)
centre. Black is up a couple tempi because

302

White has played lLige2-c1-b3xd4
rather than lLif3xd4. True, Black has
spent these tempi on ...�b8 and ... a6,
but these could be very useful moves if
he can initiate play on the queenside.
Black has also saved the move ...�e8,
which is usually played to force f2-f3.
This gives him another tempo to try to
create queenside counterplay.
At this juncture Black has more
than one way to continue. Usually he
will try to take advantage of his extra
moves on the queenside to aim for
... bs, normally preceded by either ...c6
or even ... cs. We have:
Al: 11•.•lbxd4
A2: 11••.lLleS
These two continuations are seen
with similar frequency. Black could also
consider 11...lLihS!? which is a common
idea in the aforementioned Glek.
Al) 11...lLixd4
Black chooses to simplify the centre.
12 .i.xd4 .i.e6 13 .i.e2
Pa n n o Va ria tion: 9 lLic 1
Now Black can prepare ...bs with
... c6 or he can try to achieve this ad­
vance tactically, aiming to blow up the
position while he is still ahead in de­
velopment.
All : 13 .c6
••
A12: 13 cs!?
•••
AU) 13 ...c6
A logical move. Black will follow up
with either ... bs or ... ds.
14 a4
The alternative is to allow ...bs with
14 o-o. After 14 ... bs 1 s b3 (1s cxbs
axbs 16 �fcl "ife7 17 a3 �fd8 18 b4
�bc8 was fine for Black in A.Ker­
E.Gufeld, Wellington 1988) 1s ...bxc4 16
bxc4 (instead 16 .i.xc4 .i.xc4 17 bxc4
"ifas 18 l:l.ab1 cs 19 .i.e3 lLid7 20 lLids
"ifxd2 21 .i.xd2 lLies gave Black good
counterplay in P.Korning-L.Fell, corre­
spondence 198S) 16 ... "ifas 17 �acl (17
�abl allows 17 ...cs 18 .i.e3 lLig4)
17 ...�fd8 Black intends ...cs followed by
...lLid7 with active play.
Here White has:
3 03
A ttacking C h e s s : The King 's I n dian, Vo l u m e 1
a) 18 �fd1 cs 19 .i.e3 lLid7 20 lLids
'ifxd2 21 �xd2 .i.xds 22 cxds (22 J:xds
J:b2) 22 ... as 23 a4 was A.Riazantsev­
E.Gousseinov, Istanbul 1998. Now
23 ...l:l.b4 with the idea of 24 .i.bs lLib6
would have given Black counterplay.
b) 18 �h l cs 19 .i.e3 'ifa3 (also good
is 19 ... lLid7, because 20 lLidS?! is well
met by 20 ... 'ifxd2 21 .i.xd2 �b2) 20 J:c2
lLid7 gave Black good play in
J. Hjartarson-J.Nunn, Rotterdam 1989.
14...ds
This forces the pace and while Black
should be able to make a draw, any
winning chances are rather lacking.
Black has also tried 14...'ifas lS o-o
'ifh4!? (other possibilities are lS ... cs
and 1s ... J:fd8) 16 as cs 17 .i.e3 (not 17
�a4 lLixe4! 18 fxe4 .i.xd4+ 19 'ifxd4 and
0-1 in A.Suarez-S.Kamberi, Dallas 2001)
17 ... lLid7 18 'ifcl ! .i.xc4 19 .l:.a4 .i.xe2 20
.l:.xb4 .i.xfl 21 l:l.a4 .i.bs which is un­
clear.
15 cxds cxds 16 es
White has to play this way if he
wants to obtain anything.
16...lLid7 17 f4 f6 18 exf6 lLixf6
3 04
Worse is 18 ....i.xf6 19 l:l.d1.
19 l:l.dl 'ifd7 20 0-0 l:l.bd8 21 .i.b6
Instead 21 'if e3 'iff7 (21 ...�fe8 looks
okay too) 22 .i.f3 .l:.fe8 23 'iff2 .lt.g4 24
.i.es .i.xf3 2s 'ifxf3 'if e6 26 .i.xf6 'ifxf6
27 lLixds 'ifxb2 28 fs gxfs 29 lLie7+ �h8
30 J:xd8 l:l.xd8 31 lLixfs .llf8 was level in
L.Brunner-Xie Jun, Sth matchgame,
Bern 199S.
21...d4!?
After 21 ...�c8 22 .i.f3 White has a
slight but persistent advantage.
22 'ifxd4 'ifxd4+ 23 .i.xd4
Black should meet 23 �xd4 with
23 ... lLid7 24 �d6 lLixb6 2S �xe6 .i.xc3
26 bxc3 lLixa4 (Be liavsky).
23 ... .i.b3 24 J:d3
Wisely avoiding 24 J:d2? lLie4 2S
lLixe4 �xd4.
24. ...lt.c4 25 J:d2
This was A.Beliavsky-J.Nunn, Reyk­
javik 1988. Black has sacrificed a pawn
and must play accurately to maintain
compensation.
25 ....i.xe2 26 lLixe2 lLie4 27 �ddl
.i.xd4+ 28 lLixd4
After 28 �xd4 �xd4 29 lLixd4 lLics

30 as .l:.d8 31 4Jf3 4Jb3 Black is fine.
28 ...tt:Jcs 29 as �ds 30 b4 �fd8 31 bxcs
J:xd4 32 �bl
Instead Beliavsky gives 32 J:xd4
�xd4 33 J:b1 .l:.c4 as equal.
32 ...J:c8 33 �fc1 �ds 34 ltxb7 �dxcs 3S
�el .l:.cl 3 6 J:be7 J:xel+ 37 �xel J:cs
38 �al 't;f7
Because of the difference in activity
of the rooks, Black should draw without
much difficulty.
A12) 13 ...cs!?
This is an enterprising but risky way
to play.
14 .i.e3 bs!?
Black goes all in, but 14...'ifas 15 o-o
�fd8 is also possible.
lS cxbs axbs 16 tt:Jxbs
Instead 16 o-o?! b4 17 tt:Jds (better is
17 tt:Jbs 4Je8 18 .l:.fdl 'ife7, although
Black is fine after 19 tt:Jxd6 'ifxd6 20
'ifxd6 tt:Jxd6 21 J:xd6 .i.xb2) 17 ... .i.xds
(or 17 ... tt:Jxds 18 exds .i.fs) 18 exds �e8
19 .lt.c4 4::l d7 leads to a good Benoni
structure for Black.
16... ds
Pa n n o Variatio n : 9 4J c1
The only consistent move. Black rips
the position open and is prepared to
offer more material.
17 .i.xcs
White takes everything on offer. In­
stead 17 exds tt:Jxds 18 .i.xcs is also
possible, but after 18....l:.e8 White must
avoid 19 0-0?! (a better try is 19 l:l.dl
when it is not clear that Black h as
enough after 19 ... .i.es 20 g3) 19 ... tt:Jc3 !
20 4Jd6 tt:Jxe2+ 21 'ifxe2 J:xb2 when
Black has the initiative, A.Kemmerer­
T.Marten, correspondence 1986.
11...dxe4 18 .i.xf8 'ifxf8
White is up the exchange and a
pawn, but with a development lead
and dark-square play, Black has certain
tactical chances. Some possibilities:
a) • 4::ld4!? is natural but has not
been seen in practice. After 19 ...tt:Jds 20
fxe4 .i.h6 Black has counterplay.
b) 19 'ifd6 hurries to exchange
queens: 19 ...'ifxd6 (or 19 ...exf3 20
'ifxf8+ 't;xf8 21 gxf3 tt:Jds 22 't;f2 tt:Jf4
23 a4 .i.xb2 24 �abl .i.f6 with reason­
able compensation for the exchange in
J .Brennin kmei jer-M.Bos boom, Hilver-
305
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
sum 1988) 20 lLixd6 J:xb2 21 lLixe4
lLixe4 (not 21 ... lLids 22 �dl, as in
J.Eising-A.Sznapik, Amsterdam 1984)
22 fxe4 .Jlc3+ 23 �fl J:c2 24 l:l.dl J:xa2
gives Black compensation.
c) 19 o-o exf3 (Black should not hurry
with 19 ...�xbs 20 .Jlxbs "ifcs+ because
after 21 �hl "ifxbs 22 "ifd8+ "ife8 23
"ifxe8+ lLixe8 24 fxe4 .Jlxb2 2S �abl
.lies 26 a4 the passed pawns give White
the advantage in the ending, N.Andri­
anov-A.Kuzmin, USSR 1984) 20 gxf3 (af­
ter 20 �3 Black should play the imme­
diate 20..."ifcs+, rather than 20...J:xbs?
21 .Jlxbs "ifcs+ 22 "iff2 and 1-0 in N.Le
Gore-R.Lifson, correspondence 1988),
and here 20.. J::t xbs 21 .Jlxbs "ifcs+ 22
�f2 "ifxbs 23 a4 looks better for White,
but 20 ... .Jlh6!? with the idea of ...lLids
gives Black active play.
A2) 11...lLies
Black elects to keep the knights on
the board.
12 J:d1
This deters Black from playing ...cs.
Other moves are also seen:
3 06
a) 12 o-o-o is uncommon, but inter­
esting. Then 12 ... c6 (worse is 12 ... cs 13
lLib3 .Jle6 14 "ifxd6 "ifxd6 lS �xd6 .Jlxc4
16 lLixcs when White has won a pawn)
13 h3 (13 lLib3 .Jle6) 13 ...°ifC7 14 lLib3
.Jle6 lS "ifxd6 "ifxd6 16 �xd6 leads to:
al) 16 ....Jlxc4 17 f4 (this was consid­
ered to be winning by Szuhanek, but
matters are not so clear) 17 ... hfl 18
fxes .Jlxg2 19 .l:.gl (or 19 exf6 .Jlxh l 20
fxg7 �fd8 with an unclear position, as
21 �xd8+ �xd8 22 .Jlh6? �d3 is good
for Black) 19 ...lLixe4 20 �xg2 lLixd6 21
exd6 b6 is unclear.
a2) 16 ... lLihS ! ? 17 lLid2 lLig3 18 �gl
lLixfl 19 �xfl .Jlxc4 (better is 19 ... �fd8!
with good compensation for the pawn)
20 .l:.dl .Jle6 21 f4 .Jlh6 was V.Varaciuc­
R.Szuhanek, Romanian Team Champi­
onship 1993. Here 22 �c2 ! ? looks a lit­
tle better for White due to his space
advantage.
b) 12 .Jle2 cs (bad is 12 ...c6 13 o-o
bS? 14 f4! b4 lS fxes bxc3 16 "ifxc3
lLixe4 17 "ifc2 with a clear advantage
for White) 13 lLic2 (13 lLib3 .Jle6 is
pleasant for Black) 13 ....Jle6 and now:

bl) 14 b3 scores terribly for White
after 14 ... 'ifas: for example, lS o-o bS
16 cxbs axbs 17 �fd1 (17 CZ'ixbs l:l.xbs
18 .i.xbs 'ifxbs 19 'ifxd6 CZ'ifd7 is also
good for Black) 17 ...�fe8 18 .i.f2 'Lied 7
with a very comfortable position for
Black in A.Beliavsky-E.Gufeld, USSR
1979.
b2) 14 CZ'ia3 CZ'ic6 lS o-o 'Lid4! (or
1s ...CZ'id7 16 f4 'Lid4 17 .i.d3 bS!? 18
cxbs axbs 19 .i.xbs 'Lif6 with counter­
play in A.Beliavsky-J.Hjartarson, Reykja­
vik 1991), and now both 16 �fdl 'Lid7
17 CZ'ic2 CZ'ixe2+ 18 'ifxe2 'ife7 19 'Lids?
l.xds 20 �xds CZ'ib6 21 J:d2 .i.xb2 22
.:b1 'Lia4 (M.Magomedov-K.Maslak,
Sochi 2oos) and 16 �acl 'Lid7 17 b3
it'as 18 CZ'iabl bS (Y.Gerasimov­
D.Yevseev, St Petersburg 200S) are very
good for Black.
12...c6 13 .i.e2 bs
14 cxbs
Instead 14 f4? b4! lS fxes bxc3 16
"lifc2 (or 16 'ifxc3 CZ'ixe4 17 'ifc2 'inl4+)
16 ... 'Lig4! 17 .i.xg4 .i.xg4 with a clear
advantage is analysis by Kasparov, but
after 14 o-o cs lS CZ'ib3 Black should
Pa n n o Va riation: 9 CZ'ic1
avoid his suggested 1s...b4 due to 16
'Lia4! and prefer 1S ...CZ'ixc4! 16 .i.xc4
bxc4 17 CZ'ixcs 'ifas which is unclear.
14...axbs
Black plans ...b4 and ...ds.
15 b 4 c5
Forcing the pace. Black can also play
1 s ... .i.d7 16 o-o �e8 17 CZ'ic2 .i.f8 18
�hl .i.e6 with equal chances.
16 ClJC2
Taking the pawn with 16 CZ'idxbs
cxb4 17 CZ'ib1 'ifas 18 CZ'ixd6 looks too
risky for White after 18...�d8.
White may also try 16 bxcs b4 17
'Lia4 (17 CZ'ib1! looks more critical, as
after 17 ... dxcs 18 CZ'ib3 'ifxd2+ 19
CZJ1xd2 .l:.d8 20 .i.xcs .i.e6 21 o-o Black
has to prove that he has enough for the
pawn), although after 17 ... 'ifas 18 CZ'ib6
dxcs 19 CZ'ixc8 cxd4 20 'Lie7+ �h8 21
.i.xd4 J:fd8 Black has good play:
a) 22 o-o CZ'ixe4! 23 fxe4 �xd4 24
'ifxd4 'Lif3+ 2S J:xf3 .i.xd4+ 26 �xd4
'ifcs with a clear advantage according
to Kasparov.
b) 22 'ifb2 �xd4 (instead 22 ...b3+ 23
'ifc3 'ifxc3+ 24 .i.xc3 bxa2 2 S J:xd8+
307
A ttacking C he s s : Th e Kin g 's I n dian, Vo l u m e 1

J:xd8 26 i.xes lLixe4! 27 i.xg7+ �xg7 A.Beliavsky-V.Loginov, Azov 1991, but

28 o-o lLic3 29 i.c4 .l:.d7 30 i.xa2 lLixa2 Kasparov's move looks simpler.
31 lLic6 is equal) 23 'ifxd4 b3+ 24 'ifd2 27 lLib6 �c6 28 J:b1 c3 29 �b4 i.xa2
'ifa7 25 axb3 'ifxe7 and Black's active Black had no problems in
minor pieces give him slightly the bet­ A.Beliavsky-G.Kasparov, Linares 1990.
ter chances.
16...cxb4 17 lLixb4 i.e6! 18 o-o B) 10 dS
After 18 lLixbs 'ifas! (even better More logical than 1o lLib3.
than Kasparov's 18 ... i.c4) Black is 10...lLid4
clearly for choice, while 18 'ifxd6 'ifxd6
19 J:xd6 .l:.fc8 gives him a strong initia­
tive for the pawn.
18 . 'if as 19 'ifxd6 lLic4!
..

This initiates a forcing sequence

that will lead to a balanced endgame.
20 i.xc4 bxc4 21 lLic6 'ifxc3 22 i.d4
lLixe4! 23 i.xc3 lLixd6 24 i.xg7 �xg7 25
lLixb8 lLif5
Black is down the exchange, but
White's knight is yet to escape. More­
over, the a2-pawn is weak and Black's White now has a choice of moves to
strong c4-pawn will give him enough challenge the black knight.
counterplay to hold the balance.
26 lLid7 B1t):l. �e:2
B2: U �J

Bl) 11 lLi1e2

26 ....l:.c8
Black also managed to draw after
26 ....l:.a8 27 lLics J:xa2 in the later game

308

This makes more sense than it does
with the moves 9 h4 hS included, be­
cause after 11 ... cs White can castle
kingside. We examine two lines here,
although Black really has three choices
because he can also play a pawn sacri­
fice that will be covered in Line B22.
8:11: 1�...lbke2·
81%: :1it..+t5:
811) 11...lLixe2
This is a simple way of playing.
Black aims for a normal position.
12 .i.xe2
12 ...lLihs
Having this move available is one of
the main advantages for Black com­
pared to the previous chapter. Instead
12 ... cs is less good because if we com­
pare the position to Line B21 below,
White does not have a weakness on b3
after 13 dxc6 bxc6.
13 0-0-0
Other moves are rare in practice
and against them Black will also play
13 .. .fs.
Pa n n o Va riation: 9 lLic1
13. .fs
.
Exchanging a pair of knights has
left Black with enough room for his
minor pieces. He has a flexible position
and may go .. .f4, .. .fxe4 or ...lLif4 on the
kingside, while playing on the queen­
side may also be a possibility.
14 �b1
Instead after 14 exfs gxfs or 14 g3
lLif6 (Black could also consider 14.. .fxe4
or 14.. .f4) 14 .i.d3 lLif4 lS .i.c2 bS (or
1s ..."ife7) 16 cs b4 17 lLia4 as 18 c6 fxe4
19 fxe4 .i.a6 Black has counterplay.
14 cs is White's main alternative:
a) 14.. .f4 is probably premature. Af­
ter l S .i.f2 .i.f6 16 �hfl .i.h4 17 .i.g1
White had a free hand on the queen­
side in L.Kaufrnan-W.Morris, USA 1979.
b) 14.. .fxe4!? lS lLixe4 (or lS fxe4
lLif4) 1s ....i.fs looks okay for Black.
c) 14 ... �h8 lS �bl b6 16 cxb6 cxb6
17 .l:.cl bS 18 .i.d3 lLif6 combined play
on both flanks in C.Ward-M.Hennigan,
British League 1998.
d) 14..."ife7 lS cxd6 (White's king
position looked a bit loose after lS b4
lLif4 16 .i.f1 fxe4 17 lLixe4 .i.fs in
3 09
A ttacking C he s s : Th e King 's I n dia n , Vo l u m e 1
A.Ward-J.Seia, correspondence 2006)
1s ...cxd6 16 .i.d3 bS 17 �bl .i.d7 gave
Black a fine position in l.Lutsko­
A.Fedorov, Minsk 1994.
14....i.d7
Black could also play 14..."if e7. There
is no real need to determine the place
of the hS-knight yet with 14 ... lLif4 or
14...lLif6, however.
15 .i.d3
Or lS .l:.cl "ife7 16 .i.d3 lLif6 with
chances for both sides. Instead lS a3
looks a bit pointless and 1s ... �h8 16
.i.d3 "ife7 17 l:l.c1 lLif4 18 .i.c2 cs 19
dxc6 bxc6 (with the idea of ... lLie6, tak­
ing aim at the d4-square) 20 cs? dS!
was good for Black in A.Dreev-Ye Jiang­
chuan, Shanghai 2001. White cannot
grab the pawn with 21 exds lLixds 22
lLixds cxds 23 "ifxds, because of
23 ... .i.e6 with the idea of ...e4.
15 ...�h8
Black might consider too 1s ..."ife7 ! ?.
16 .l:.hgl!?
White is thinking about playing g2-
g4 at some point.
16..."ife7 17 .i.c2 lLif6 18 h 3
310
After 18 exfs gxfs 19 g4?! fxg4 20
fxg4 lLixg4 21 .i.gs .i.f6 White may
have some play, but Black has a solid
extra pawn.
18 ... bs! 19 cxbs axbs 20 g4 b4 21 lLie2
fxe4 22 fxe4 cs! 23 dxc6 .i.xc6
Black has good counterplay against
the e4-pawn and another possibility is
to break in the centre with ... ds, to
open up the diagonal fort he g7-bishop.
24 .i.gs
White will take the f6-knight to re­
lieve the pressure on e4, but Black still
has good chances. It was probably pref­
erable to go in for 24 "ifxd6 °ifh7! 2S
lLig3 b3! 26 axb3 lLixe4 27 lLixe4 .i.xe4
28 .i.xe4 "ifxe4+ 29 °ifd3 "ifxd3+ 30
l:l.xd3 e4 31 l:l.d7 J:xb3 32 .i.cl when
White should hold the ending, but
Black still has some chances to press.
24...°ifa7 25 hf6 J:xf6 26 lLic1
This was l.lbragimov-A.Fedorov, So­
chi 1997. Now 26 ... �f2! 27 "ifxd6 b3
would have been strong: 28 .i.xb3
(Black wins after 28 lLixb3 �xc2 29
�xc2 .i.xe4+ or 28 axb3 l:l.a8 29 "ifa3
"ifb6 30 "ife7 .i.f8! 31 "ifxes+ .i.g7 32

'ife7 'ifas! 33 'ifa3 .i.xe4!) 28... .i.xe4+ 29
�al .i.c2 with the idea of ... e4 (Fedorov)
would have been very good for Black.
812) 11... cs 12 dxc6 lLixc6
Note that the pawn sacrifice
12 ... bxc6 13 lLixd4 exd4 14 .i.xd4 will be
considered under Line B22, below. The
text move looks better, however, as
Black gets good chances with much less
risk.
After the safer recapture, the posi­
tion resembles a Maroczy Bind where
Black has played the rather horrid­
looking ...es. The justification for this is
Black's development lead and possibili­
ties of breaking in Sicilian style with
...bs and even ...ds. Black will have to
play very actively to hold the balance,
but White's position is also not so easy
to pl ay.
13 �d1
White logically increases the pres­
sure on the d-file. Other moves have
also been tried, but they do not appear
to be too dangerous:
a) 13 o-o-o .i.e6 14 lLids bS 15 lLiec3
Pa n n o Va ria tio n : 9 lLic1
lLid4 and Black already has the initia­
tive.
b) 13 lLids bs (13 ... lLixds 14 cxds
lLias 15 lLic3 bS also looks okay) 14
lLiec3 in creases White's grip on dS, but
gives up the d4-square, and after
14 ... lLixds (14 ...lLid4 15 .i.d3 lLixds is
fine too) 15 cxds (or 15 lLixds lLid4)
15 ... lLid4 Black has a good position.
c) 13 lLicl .i.e6 (Black can also sacri­
fice a pawn with 13 ...bS!? 14 cxbS axb s
15 .i.xbs lLid4 with the idea of ... ds) 14
lLib3 lLias (14...lLie8! ? 15 .i.e2 fS is an­
other, more risky way to play) 15 lLixas
(instead 15 lLids lLixds 16 cxds lLixb3
17 axb3 .i.d7 is also level) 15 ...'ifxas 16
lLids (16 'ifxd6 bS 17 cxbs axbs gives
Black good counterplay) 16 ... 'ifxd2+ 17
�xd2 lLixds 18 cxds .i.d7 leads to a
sterile ending.
13 ....i.e6
Black develops and counterattacks
against the c4-pawn.
14 lLids
This natural move may not be best,
but White has trouble achieving any­
thing:
311
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
a) 14 'ifxd6 'ifxd6 l S l:txd6 .Jlxc4 is
certainly okay for Black.
b) 14 b3 bS (Black could also con­
sider 14 ... 'if as lS lLids and here not
1s ... 'ifxd2+ 16 J:xd2 when White is bet­
ter, but rather 1s ... 'ifa3!?) lS 'ifxd6 (or
lS cxbs axbs 16 'ifxd6 'ifc8 with coun­
terplay) 1s ...'ifxd6 16 �xd6 lLib4 17
lLids lLic2+ 18 �d2 lLixe3 19 �xe3 bxc4
20 bxc4 J:b2 21 J:xa6 lLid7 and Black's
activity combined with White's poor
development should give him compen­
sation for t he pawns.
c) 14 lLicl lLid4 (Black could also
consider 14...'ifc8! ? lS lLib3 bS 16 cxbs
axbs and now both 17 lLixbs dS and 17
.Jlxbs .l:.d8 followed by ...ds leave him
with good play for the pawn) lS lLib3
bS 16 lLixd4 (after 16 cxbs .Jlxb3 17
axb3 axbs 18 .Jlxd4 exd4 19 lLixbs dS
20 es �e8 Black has the initiative)
16 ... exd4 17 .Jlxd4 bxc4 18 .Jle2 lLihS ! ?
19 .Jlxg7 lLixg7 2 0 .l:.cl (Black also looks
to be fine after 20 o-o 'ifb6+ or 20 'ifxd6
'ifg s with counterplay against g2 and
b2) 20 ... lLihs (20...'ifb6!?) 21 lLid1?! (bet­
ter is 21 o-o 'ifb6+ 22 �hl 'ifxb2 23
'ifxd6 .l:.fd8, but Black is fine here too)
21 ...'inl4+ 22 lLif2 'iff6 23 �c3 lLif4 and
Black had the initiative in N.Mishra­
M.Waqar, India 1999.
14... b5 15 cxb5
After both lS lLixf6+ .Jlxf6 (but not
1s ... 'ifxf6?? 16 .Jigs) and 1s lLiec3 lLid4
Black is also doing well.
15 ...axb5 16 lLixf6+
Or if instead 16 lLiec3 then 16 ... lLid4
with good play.
312
16....Jlxf6
17 g3
White tries to get developed. In­
stead 17 lLiC3 lLid4 is very comfortable
for Black, while 17 'ifxd6 'if as+ 18 .Jld2
(even worse is 18 'ifd2 lLib4!) 18 ...'ifb6
gives Black a dangerous initiative for
the pawn.
11....Jlxa2 18 'ifxd6 .Jlb3
Black could also consider 18...lLid4
19 'ifxd8 �fxd8 20 lLixd4 exd4 21 .Jlf4
.l:.a8! 22 .Jlxbs .Jlb3 when if the rook
moves 23 ...d3 will give him a strong
initiative for the pawn.
19 'ifxd8 �fxd8
Gufeld's suggestion 19 ... .Jlxd8 is
also fine. After 19.. 1Hxd8 Black's posi­
tion was a little easier to play in
Z.Polgar-E.Gufeld, Wellington 1988,
although the game was still pretty
level.
82) 11 lLib3
This is a little more challenging, as
White does not clog up his own devel­
opment. Black can simplify or offer a
thematic pawn sacrifice.
 .
P a n n o Varia t i o n : 9 ltJ c1

pawns, while both 1 3 .i.e2 and 1 3 .i.d3

would be met by 13 ... ltJh s .

821) 11 r.-.
'"UXb3 12 axb3 cs
...

.
Black blocks th e queens1de. Also
.
worth considen·ng is 12 ...r.-.h
'"l.J S at once.

After the text, White can begin

queens1"de play or make a committal
.
kmgs1"de advance.

8211: �3 1>4
8212: 13 g4

Instead 13 dxc6 bxc 6 gives Black

.
enoug h play agamst the doubled b -

313
A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1
17 lLib6
After 17 lLic3 Black also has
17 ....i.h6! 18 .i.f2 (18 .i.xh6 "ifh4+)
18 ...°ifg5, which threatens mate and he
has counterplay with both 19 lLidl f5
and 19 .l:.dl .i.d7.
11 ....i.h6I
Here we see another resource that
is available to Black because of the
omission of 9 h4 and 9 ...h 5.
18 .i.f2
Instead 18 .i.xh6 "ifh4+ 19 �dl
"ifxh6 is fine for Black. After 20 lLixc8
�xc8 21 "ifd2 he can play 21 ..."ifxd2+ 22
�xd2 f5 (V2-V2 in V.Lukov-Z.Lanka, Ge­
neva 1993) or 21 ...'inl4 22 .i.xb5 lLig7
with compensation for the pawn.
18 ...�xb61
Black falls into the 'trap'.
19 "ifas
19 ....l:.a61
Only this queen sacrifice can justify
Black's play.
20 "ifxd8 J:xal+ 21 �e2 J:b1
This is better than 21 ....i.a6, which
has also been tried.
22 "ifxc8
314
After 2 2 .i.h4!? .i.a6 2 3 °ifb6 �xb2+
24 �el lLic7! 25 .i.e2 (not 25 "ifxc7?
.rf.c8) 25 ... �c8 26 �f2 Black has:
a) 26 ...�a2 27 J:e1 .i.d2 28 �bl .i.a5
29 "if xd6 b4 30 "ifxe5 .i.xe2 should fa­
vour Black, although White went on to
win in R.Ristorto-C.Blanco Gramajo,
correspondence 2000.
b) 26 ...b4! ? 27 �el .i.d2 28 .i.xa6
.i.xel+ 29 �fl (after 29 �xel? the ob­
vious 29 ... lLixa6 gives Black a winning
position, although 29 ...lLixd5! is even
stronger) 29 ... lLixa6 30 "ifxa6 .i.xh4 31
"ifxc8+ �g7 and only Black can be bet­
ter.
22 ....i::tx b2+ 23 �d3?
It is more prudent to play 23 �el
when Black should take the perpetual
check.
23 ...J:xf2 24 °ifh3 J:d2+ 25 �C3 b4+ 26
�xb4 .i.e3
Black had a strong attack in
l.lbragimov-A.Delchev, Linares 1997.
8212) 13 g4
This is more ambitious. White pre­
vents ... lLih5.

13 ... hs
This is much more enterprising
than 13 ...lLie8 (th is is better than in the
analogous line with 9 h4 hS - note 'b'
to Black's 14th in Line A of Chapter 16 -
but it still looks insufficient) 14 h4 fS
1s gxfs gxfs 16 exfs .i.xfs 17 hs and
White is better.
14 h3 lLih7
This is a typical idea in this struc­
ture and it is also seen in certain 6... es
lines of the Samisch.
15 gxhs
One point is that after lS o-o-o h4!
Black keeps the kingside closed and
plans ... .i.f6-g s. After 16 gs Black can
Pa n n o Varia tio n : 9 lLic 1
create counterplay on either side of the
board with 16 ... bs or 16 .. .f6.
1s ...°ifh4+
This is the other point. Black avoids
ruining his pawn structure. Still, his
queen may prove to be misplaced on
the edge, so care is needed.
16 "iff2 "ifxhs
This is better than 16..."ifxf2+ 17
�xf2 gxhs 18 .i.d3 h4 19 �e2 �h8 20
b4 when White has a nice edge in the
endgame, A.Karpov-Xie Jun, 3rd
matchgame, Guanzhou 2000.
17 .i.e2
Now Black's queen does not feel too
comfortable, but there are sufficient
resources.
11...lLif6!?
An interesting try. Worse is 17 ... .i.f6
18 h4!, but 17 .. .fs is more common
when White has:
a) 18 f4 °ifh6 and now:
al) 19 h4 exf4 20 'i'xf4 gs 21 "iff1 f4
(or 21 ... .i.es !? as suggested by Golubev,
while 21....i.d4!? with the idea of 22
hxg s "ifg7 looks good as well) 22 hxgs
"ifxgs 23 .i.d2 °ifg3+ 24 �d1 f3 2s .i.d3
315
A ttacking Ch ess: The King 's I n dian, Vo l u m e 1
lLif6 with sharp play that did not look
unfavourable to Black in A.Utnasunov­
V.Kuznetsov, Vladimir 2002.
a2) 19 exfs exf4 20 "ifxf4 gs 21 °ifg3
.i.xfs ( 2 1....i.es !? is again proposed by
Golubev) 22 h4 �be8! 23 �d2 .i.xc3+!
24 bxc3 lLif6 2s "ifxgs+ "ifxgs 26 hxgs
lLie4+ 27 �c2 was T.Roussel Roozmon­
K.Lahno, Montreal 2004, and now
27 ...lLixc3+! would give Black a clear
advantage.
b) 18 exfs! ? gxfs (Golubev suggests
18 ....i.xfs 19 f 4 °ifh6 20 fxes gs with
unclear play) 19 f4 'inl6 20 o-o-o e4 21
�dgl �h8 22 �h2 gave White the ini­
tiative in N.Vitiugov-A.Shomoev, Ulan
Ude 2009.
18 f4 °ifh8!
19 f5
After 19 fxes lLid7! Black has good
counterplay.
19...lLihs
Instead 19 ...bS!? with counterplay
may be even better.
20 .i.xhs "ifxhs 21 f6 .i.h6 22 �d2 �h7
The position is balanced but sharp,
O.Sanarov-P.Kotenko, Krasnodar 2001.
316
e22) 11 cs
•••
A familiar pawn sacrifice. Again the
omission of 9 h4 hS will change the
assessment of certain positions.
12 dxc6 bxc6 13 lLixd4 exd4 14 .i.xd4
�e8
Black has had little success with
moves like 14... cs and 14 ... "ifas. but
14... ds is another move order: l S cxds
cxds 16 es (or 16 exds �e8 17 .i.e2)
16 ...lLid7 (instead 16 ... �e8 17 .i.e2 is
Line B221 below, but here White has
the extra option of 17 o-o-o! ? to reckon
with as well) 17 f4 f6 and now both 18
e6 .l:.e8 19 .i.e2 and 18 exf6 �e8+ 19
.i.e2 transpose to the notes to Black's
18th in Line B221, but Black should
probably avoid this transposition.
15 .i.e2 dS 16 cxds
Here it is best for White to exchange
c-pawns immediately, as after 16 es
lLid7 17 f4 Black has 17 ....i.h6! with the
idea of meeting 18 "ife3?! With 18 ... cs!
(Golubev).
Black has the same choice as he did
in the analogous position in Chapter
16.

.·· .. ·IUJ.: $6�.ocd.S
a:22z:::t.��.:"4
8221) 16...cxds
17 e5
After 17 exds �b4 18 o-o (White
should avoid 18 .lies �xb2! 19 'ifxb2
lLixds when Black is better after 20 �cl
lLif4! or 20 o-o-o .Jlxc3 21 'ifxc3 lLixc3
22 l:l.xd8 �xd8 23 .Jlc4 .llfs with some
initiative) 18 ...lLie4 (capturing on d4 is
less effective) 19 fxe4 .Jlxd4+ 20 �h l
.lies Black has enough compensation
for the two pawns according to
Golubev.
17 ... lLid7 18 f4
White can also play 18 o-o lLixes 19
�adl, but Black was not without coun­
terplay after 19 ... .Jle6 20 �hl (or 20
.Jlxa6 'ifas 21 .Jle2 �b4) 20 ...J:b4 in
V.Vdovichenko-P.Kruglyakov, Lvov
2007. If 21 lLixds (White could instead
repeat moves with 21 .lies �b8 22
.Jld4) 21 ...'ifxds 22 'ifxb4 lLic6 23 .Jlxg7
lLixb4 24 l:l.xds .Jlxds 2s .i.C3 �xe2 26
.Jlxb4 �xb2 the endgame is level.
Pa n n o Va ria t i o n : 9 lLic 1
18 ....Jl h61?
In stead 18 .. .f6 tends to leave Black
with a passive position:
a) 19 e6 lLif8 20 fs gxfs may look
ugly, but the position is more unclear
than anything else.
b) 19 o-o fxes 20 fxes lLixes 21
�adl! 'ifd6 22 �hl .llfs 23 b3 was
slightly better for White in S.Atalik­
M.Golubev, Bucharest 1996.
c) 19 exf6 lLixf6 20 .lies �b4 21 o-o
'ifb6+ 22 �h l �xb2 was J.Borges
Mateos-A.Escobar Forero, Barcelona
20os, and now 23 'ifcl! would give
White some advantage.
19 'ife3
Golubev stopped his analysis here.
Indeed, Black will have to be resource­
ful to avoid a clearly worse position.
19...J:xb2 20 �dl
After 20 lLixds �xe2+! 21 'ifxe2 .llb 7
Black has definite counterplay.
20 ...lLifB! 21 .Jlf3 lLie6 22 lLixds lLixd4
2 3 'ifxd4 'ifas+ 24 �1 'ifbs+ 2 5 �g1
.Jlf8 26 l:l.cl �dB
With some active and precise play
Black had obtained reasonable com-
317
A ttacking C h e s s : The King 's I n dian, Vo l u m e 1
pensation in W.Brandhorst-A.Civitillo,
correspondence 2007. The game was
eventually drawn.
8222) 16 ...�b4
This is probably asking too much of
the position because White's kingside
has not been weakened with h2-h4.
11 .i.cs
Other moves are also possible:
a) 17 .i.f2 cxds 18 �dl "ifas is simi­
lar to the related position after 9 h4 h S
(note 'b' t o White's 18th move i n Line
B12 of Chapter 16) and should be okay
for Black.
b) 17 .i.xf6 .i.xf6 18 o-o-o cxds 19
lLixds .i.xb2+ 20 �c2 l:l.b8 21 l:l.b1 .i.es
22 l:l.xb8 .i.xb8 gave Black good com­
pensation for the pawn in S.Simanjun­
tak-J.lvanov, Thessaloniki 2009.
c) 17 es lLixds 18 lLixds .llxd4 19
"ifxd4 and here:
cl) 19 ... �xes 20 °ifb6 "if xds 21 �dl
"ife6 22 l:l.d8+ .i.f8 23 o-o was
T.Sammalvuo-J.Johansson, Turku 1997,
and after 23 ...�g7 24 �xf8 �xf8 25
.i.xa6 .i.xa6 26 "ifxa6 �e2 Black is just
318
hoping for a draw according to
Golubev. While this is true, Black is ac­
tive enough that he should be able to
hold without too much difficulty.
c2) 19 ... .i.xes isn't as good for Black
as when 9 h4 hS has been thrown in,
because he then has ... .i.g3+ available,
but here Black's compensation still ap­
pears sufficient to hold: 20 °ifb6 "ifxds
21 J:d1 "ife6 22 o-o .i.f4 23 °ifh4 "ifes
(23 ... °iff6 24 g3 l:l.xe2 25 "ifxf4 "ifxb2 26
°ifh6 °ifb6+ 27 �hl is somewhat better
for White according to Golubev) 24
�h l "ifgs 25 g3 �xe2 26 "ifxf4 "ifxf4 27
gxf4 28 l:l.b1 .i.xa2 29 �fel l:l.d2 30
l:l.edl l:l.e2 3 1 .l:.el V2-V2 was 1.Sharpe­
J.Slaby, British League 2010.
17 ...lLixe4
Again Black relies on this trick to
justify his play.
18 fxe4 .i.xc3 19 "if xc3 J:bxe4 20 o-o
J:xe2 21 dxc6
This position should be better for
White than it is in Chapter 16 (Line
B12), because there is no pawn hang­
ing on h4.
21. "if gs?
..
 P a n n o Va riatio n : 9 CZ'ic 1

This fails spectacularly, but Black is clearly worse by now in any case,
struggling in any case. Instead 21 ..."ifd5 though.
22 .i.f2 �8e3 23 "ifxe3 J:xe3 24 .i.xe3
"ifxc6 25 l:l.acl °ifh7 26 �fdl .i.e6 is in­
sufficient according to Van der Tak. The
best try looks to be 21 ...°ife7 ! ? 22 "iff3
(instead 22 .i.f2 .i.f5 with the idea of
... .lt.e4 looks okay, but 22 .i.d4!? would
test Black's resources) 22 ... .i.b7! 23
cxb7 "ifxc5+ 24 �hl �2e7 2 5 �ael
V2-V2 was A.Barreras Garcia-C.Blanco
Gramajo, correspondence 2007. After
25 ... �xel 26 "ifxf7+ �h8 27 °iff6+ it's
perpetual check. 23 .i.f8!!
22 l:l.f2 A fantastic, highly-creative shot.
This is stronger than 22 .i.f2 .l:.8e3 White threatens mate on g7.
23 °ifc4 �e4 24 °if d3 J:.4e3 25 "ifc4 l:l.e4 23 ...�2es
26 °ifh3 (it is probably better to allow a No better are any of 23 ...J:xf8 24
repetition with 26 °ifd3 �4e3) 26 ... .i.e6 J:xe2, 23 ...�xf8 24 'inl8+ �e7 25 �xe2+
27 C7 .i.xb3 28 c8°if+ .l:.e8 29 °ifcl "ifxcl or 23 .. .f6 24 °ifc4+ J:.2e6 25 c7 when
30 .l:.fxcl .i.e6 and Black had a slight White wins.
initiative in the ending in J.Ehlvest- 24 .i.d6! �e3 25 .i.f4! J:xc3 26 .i.xgs
1.Smirin, Connecticut 2003. �cs 27 .i.f4 .i.cB 28 c7
22 ....i. h3? White's advantage was decisive in
Consistent, but losing. Black is J.Michenka-E.Hagara, Trinec 1998.

319
 Cba,pter 18
Panno Variation
" ,'' ' ' ' '

1 d4 tt:'lf6 2 c4 g6 3 tt:Jc3 �g7 4 e4 d6 5
f3 o-o 6 �e3 tt:Jc6 7 tt:Jge2 a6 8 'tlt'd2 l:tb8
Bl, while 10 �h6 bxc4 (10 ... �xh6 11
'tlt'xh 6 es 12 h4 is Line (1, below) 11 h4
tt:'lb4 12 tt:'lg 3 and now 12 ... �xh6! (this
looks even stronger than 12 ... cs 13
�xg7 @xg7 14 dxcs �e6, which is also
satisfactory) 13 'tlt'xh6 cs! 14 es (14
dxcs runs into 14 ... tt:Jxa2+! 1S tt:Jxa2
'tlt'as 16 @b1 c3) 14 ... tt:'lg4!! 1s fxg4
�6! 16 l:td2 cxd4 gives Black a strong
attack

In this chapter we will look at White's

attacking tries other than 9 h4.

A; 9 .0·0•.0
8� 9 14
C: t khri

A) 9 0-0-0
This move often transposes to other 10 es
...

lines. Instead 10...hs takes play back into

9 bs 10 h4
... the realm of Chapter 1s, while 10 ...tt:Jas
Instead 10 g4 is covered under Line 11 tt:Jg 3 (or 11 tt:Jf4 tt:Jxc4 12 �xc4 bxc4

320
 P a n n o Varia tio n : 9 0 - 0 - 0, 9 il h 6 a n d 9 g4
13 hS tt:Jd 7!? with the idea of ...cs)
11...tt:Jxc4 12 Ji.xc4 bxc4 13 hS c6 sud­
denly reaches note 'b1' to Black's 10th
in Line B of Chapter 14. Black could also
consider 10 ... bxc4 with the idea of 11
hS tt:Jxh S !? (usually this move is bad,
but there are a few exceptions) 12 g4
tllb 4!, which is interesting.
11 d5
11 hS could also come about from
the move order 9 h4 bS 10 h S es 11
o-o-o and might be met by 11... tt:Jxh S!?
with the idea of 1 2 g4 exd4.
11 ...tt:Jas 12 tt:Jg3 b4
Black should avoid 12 ... hS?! 13 cxbs,
but 12 ...Ji.d7 ! ? is possible. Then 13 h S?
would be met with the thematic trick
that occurs in the main line following
13 ...b4 14 tt:Jb1 Ji.a4 1s l:te1 Ji.b3!.
13 tt:'lbl Ji.d7
This looks preferable to both 13 ... c6
14 dxc6 and 13 ... hs 14 cs!.
14 ilh6?
Instead 14 cs would also be met by
14... Jla4!, but White should probably
settle for 14 b3.
14...Ji.a4! 15 Ji.xg7 �xg7 16 l:te1
Even worse is 16 b3 Ji.xb3!.
16...Ji.b3! 17 hs Ji.xa2 18 'tlt'gs tt:'lb3+ 19
�C2 tt:Jd4+ 20 �Cl h6 21 'tlt'e3 l:th8 22
tt:Jd2 cs
Black had a strong attack in T.Meinert-
1.Gaponenko, Polanica Zdroj 2000.
8) 9 g4
A sharp but committal move. White
discourages ... h s and now tt:Je2-g3 will
not block the g2-pawn.
9... bs
Here White can castle or try to at­
tack with his king still in the centre.
Bt: �00-o-Q
82: 10 h4
Other moves are dubious:
a) 10 tt:Jg3? is a common mistake:
10 ... es 11 dS (the ugly 11 tt:Jce2 fares no
better after 11 ... exd4 12 tt:Jxd4 tt:Jxd4 13
Ji.xd4 cs! 14 Ji.e3 bxc4 1S Ji.xc4 tt:Jxg4!
16 fxg4 l:txb2 and White's position is
falling apart) 11...tt:Jd4 12 Ji.xd4 (both
12 Ji.g2 and 12 Ji.e2 are simply met by
11...bxc4, while 12 �2 bxc4 13 Ji.xc4
321
A ttacking C h e s s : Th e Kin g 's In dian, Vol u m e 1
�xg4! 14 fxg4 tt:Jxg4+ 1S �g2 l:txb2! is
crushing) 12 ... exd4 13 tt:Jce2 d3 14
'tlt'xd3 bxc4 1S 'tlt'xc4 l:txb2 and Black is
much better.
b) 10 cxbS?! is not much better:
10 ... axbs 11 tt:Jg3 (or 11 h4 hS!, while 11
�h6 �xh6 12 'tlt'xh6 b4 13 tt:Jds es!
gave Black a good game in
R.Schlindwein-R.Schmaltz, Viemheim
199S) 11 ... es (11 ... b4 12 tt:Jce2 tt:Jas 13
b3 cs is also good) 12 dS tt:'ld4 13 �xd4
exd4 14 tt:'lxbs tt:Jxds! 1 s exds 'tlt'e8+ 16
�2 l:txbs 17 �xbs 'tlt'xbs gives Black
excellent compensation for the ex­
change.
Bl) 10 0-0-0
This can also arise, of course, from
the move order 9 o-o-o bS 10 g4.
10... bxc4
Worse is 10 ... es 11 dS tt:Jas 12 tt:'lg3
when White has reached the set-up he
was aiming for: 12 ... �d7 (White is also
better after 12 ...bxc4 13 h4 c6 14 g S! or
12 ... tt:Jxc4 13 �xc4 bxc4 14 h4 with the
initiative) 13 cs (White must avoid 13
h4 b4 14 tt:'lb1 �a4!) 13 ...b4 14 c6!
322
(again 14 tt:Jb1? �a4!) 14...bxc3 1S
'tlt'xc3 tt:Jxc6 (1S...tt:Jxg4 might be a bet­
ter chance to complicate matters) 16
dxc6 �e6 17 gs has occurred several
times. White looks a little better to me,
while 17 �c4!? also looks favourable to
him.
11 h4
This is the best move. Instead 11
tt:Jg3?! runs into 11 ...es! again, while
other moves also do not impress:
a) 11 gs ?! (now White will not be
able to exchange of the g7-bishop)
11 ... tt:'ld7 12 h4 tt:'lb4 13 tt:'lg3 cs! is good
for Black, as 14 dxcs? loses to 14 ... �xc3
(14 ... tt:Jes and 14 ... 'tlt'as also win) 1s
bxc3 tt:'ld3+ 16 �xd3 'tlt'as! with a deci­
sive attack
b) 11 �h6 could be met with the
untried 11 ... �xh6!. After 12 'tlt'xh6 es 13
dS tt:Jas 14 tt:Jg3 'tlt'e7 1s h4 tt:'ld7 16 h s
tt:Jcs, with ideas like ... g s, closing lines,
as well as ...�d7 and ....l:tb4 Black has
excellent counterplay.
11 ... hs
Instead 11 ...tt:'lb4 looks less trust­
worthy: 12 tt:'lg3 cs 13 h S (better than
 Pa n n o Va riation: 9 0 - 0 - 0, 9 il h 6 a n d 9 g4

13 dxcs tt:'ld7! ? 14 Ji.xc4 tt:Jes 1 5 Ji.e2
'tlt'as 16 �bl Ji.e6 when Black's attack is
very strong) 13 ...cxd4 14 Ji.xd4 (White
looks faster in this Dragon-like posi­
tion) 14...tt:Jc6 15 hf6 ! ? (15 Ji.e3 Ji.e6 is
unclear) 1s ...Ji.xf6 16 hxg6 hxg6 and
now 17 tt:Jfs! ? Ji.xfs 18 'tlt'h2 l:te8 19
gxfs gives White a strong attack.
12 tt:'lg3
White is probably better off with 12
Ji.h6, taking play into the variation 9 h4
hS 10 o-o-o es 11 Ji.h6 bxc4 12 g4 (Line
C2 of Chapter 15), which is fairly unex­
plored and quite unclear.
placed here, but f3 is weak as well, so
White must plough forward.
14 gxhs tt:Jxf3 15 'tlt'g2 tt:'ld4
1S ... Jlg4!? is also possible.
16 Ji.gs?!
The pin is not dangerous. Better was
16 h6 Ji.h8 17 Jlxc4 tt:'lg4 with an un­
clear position.
16...Jlg4 17 h6 Ji.h8 18 Ji.e2

18...l:txb2! 19 @xb2 'tlt'b8+ 20 �cl 'tlt'b4

By now Black had a winning attack
in J.Philippy-J.Schuller, correspondence
1997.

12 ...es! 82) 10 h4
It is better to strike in the centre
here than to play 12 ... hxg4 13 h S when
White's attack looks very strong.
13 d5
Instead after 13 dxes tt:Jxes 14 gxhs
(Black is clearly better after 14 Ji.e2
hxg4 15 f4 tt:Jf3) 14...tt:Jxf3 15 'tlt'g 2 tt:Jes
16 hxg6 fxg6 17 hS tt:Jfg4 Black was bet­
ter across the whole board in A.Dunne­
B.Hlavica, correspondence 2001.
13 ...tt:Jd4
Not only is the knight excellently The theory of this line is mostly

323
A ttacking C h e s s : The King 's I n dian, Vol u m e 1
based on a 1992 game between Chris
Ward and Joe Gallagher. Interestingly
enough, both players subsequently
wrote books on the Samisch Variation
and annotated this game.
10...hs
White has the initiative after both
10 ... bxc4 11 h S es (or 11 ... tllb4 12 tllg 3)
12 dS tllb4 13 tll g 3 and 10...tll a s 11
tll g 3 tllxc4 12 �xc4 bxc4 13 hs.
11 tllg 3
Instead 11 o-o-o would transpose to
Line Bl, while it is harder for White to
attack the Black king after 11 gs tlld7.
Instead after 11 gxh s tllx hs it is diffi­
cult to eliminate the h S-knight as mov­
ing the knight from e2 weakens d4.
Finally, 11 �h6 transposes to note 'b2'
to White's 10th move in Chapter 1S,
which is not really sound for White.
11...esl 12 dxes
Not 12 dS? tlld4.
12 ...tllxes 13 �e2 hxg4
White would have a free attack af­
ter 13 ...tllxc4 14 �xc4 bxc4 1S gxhs, as
in A.Gupta-N.Vinuthna, Mumbai 2004.
14 f4
324
14...tllf3+
The alternative 14 ... tllxc4 looks risky
after 1s �xc4 bxc4 16 h S gxh s 17
tllxhs when he has tried:
a) 17 ... tllxhs 18 l:txhs fs 19 o-o-o
gives White the initiative. This has
been played a couple of times, but it
just looks too dangerous to me.
b) 17 ...tll x e4 is an interesting try,
but it should fall short: 18 tllxe4 l:txb2
19 tllhf6+! �xf6 20 tllxf6+ 'tlt'xf6 21 'tlt'd4
'tlt'xd4 (21 ... 'tlt'g7 22 'tlt'xg7+ �xg7 23
�d4+ �g6 24 �xb2 �b7 may be a
slight improvement) 22 �xd4 l:te8+ 23
�fl f6 24 �xb2 l:te4 was C.Ward­
G.Buckley, Guildford 1991. White's ex­
tra rook should outweigh the pawns,
especially after 2S l:th6 l:txf4+ 26 �g1.
15 �xf3 gxf3 16 hs tllg4 17 o-o-o
11... b4
This was suggested by both Galla­
gher and Ward in their respective
books on the Samisch. There are a cou­
ple of other moves:
a) 17 ...tllx e3 was played in the
aforementioned game between the
opposing experts. After 18 'tlt'xe3 b4 19
 Pa n n o Va ria t i o n : 9 0 - 0 - 0, 9 il h 6 a n d 9 g4
tt:Jds c6 20 h6! Ji.f6? (instead 20...cxds
21 hxg7 �xg7 22 fS gives White a win­
ning attack; best is probably 20... Ji.xb2+
21 @xb2 cxds, but even here White
looks better) 21 tt:'lxf6+ 'tlt'xf6 22 fS aS
there is:
al) 23 tt:'lhs? is a bit overzealous:
23 ... gxhs 24 l:txd6 Ji.e6 (24 ... 'tlt'es 2s
'tlt'gs+ �h7 26 l:tf6 l:tg8 27 l:txf7+ �h8 28
l:tg7 intending 'tlt'g6 is dangerous for
Black according to Gallagher) 2S 'tlt'xf3
(instead 2S fxe6 fxe6 gives Black coun­
terplay with his passed f-pawn) 2S ... �h7
26 'tlt'xhs l:tbd8 27 l:txd8 l:txd8 28 fxe6
fxe6 29 'tlt'h4 'tlt'd4 30 'tlt'e7+ V2-V2 was
C.Ward-J.Gallagher, London 1992.
a2) 23 l:tdf1 a4 24 l:txf3 a3 2s es!
was what worried Gallagher. White will
maintain some advantage, even in the
endgame.
a3) 23 h7+ �h8 24 'tlt'h6 'tlt'g7 2S
'tlt'xg7+ �xg7 26 l:txd6 with advantage
for White is given by Ward, who admit­
tedly points out that Black should look
to improve earlier.
b) However, 17 ...bxc4!? has gone
unmentioned.
It appears to be logical to me, as
Black opens the b-file instead of chas­
ing White's knight into the centre of
the board.
Now 18 tt:Jds? l:txb2 is bad for White,
and Black defends after both 18 Jld4
hd4 19 'tlt'xd4 'tlt'f6 and 18 hxg6 fxg6
19 fS 'tlt'f6. Another possibility is 18
l:tdf1 tt:Jxe3 19 'tlt'xe3 Jlg4 20 h6 Ji.f6 21
es!? dxes 22 tt:Jge4 exf4 (22 ...�h8!?) 23
'tlt'xf4 Ji.fs and Black appears to at least
be holding his own in a sharp position
(and he is up three pawns at the mo­
ment).
18 tt:Jds c6 19 h6 tt:Jxh6
Ward points out that after
19 ... Ji.xh6 20 fS tt:Jxe3 21 l:txh6 tt:'lg4 22
l:tdh 1 White has a strong attack. Black
could instead try 20 ... Ji.g7, but after 21
f6 tt:Jxf6 22 Ji.h6 tt:'lg4 23 Ji.xg7 @xg7 24
tt:'le3 tt:'lxe3 2 S 'tlt'xe3 l:th8 26 es White
has pretty good compensation for the
pawns.
20 f5
This is Ward's idea. Instead Galla­
gher discusses some complicated lines
starting with 20 tt:Jxb4 'tlt'as 21 tt:Jxc6
325
A ttacking C h e s s : The King 's I n di a n , Vo l u m e 1

'tlt'xa2 22 tt:'le7+ �h7 23 l:txh6+ �xh6 24

�d4, but Ward's suggestion looks like a
better try.
20 tt:Jg4 21 f6
...

(1: 9 .txh6
.•.

c2: 9 bs•••

21 ...�xf6 Black can also pl ay in the centre

White has serious compensation af­
ter 21 ...tt:Jxf6 22 �h6 tt:Jg4 (or 22 ... tllxds
23 �xg7 �xg7 24 exds l:th8 25 'tlt'd4+ f6
26 l:txh8 'tlt'xh8 27 dxc6) 23 �xg7 �xg7
24 tt:Je3 because of the open h-file and
pressure on the d6-pawn (Ward).
22 tllxf6+ 'tlt'xf6
Of course, not 22 ...tt:Jxf6?? 23 'tlt'h2.
2 3 �d4 tt:'les 24 cs l:tbs!?
Instead Ward points out that
24... dxcs? loses to 25 'tlt'h2 l:te8 26 'tlt'h7+
with 9 ... es and then :
a) 10 �xg7 �xg7 11 dS tt:Jas 12 tt:Jc1
cs is satisfactory for Black.
b) 10 dS is well met by 10 ... �xh6 (or
10 ... tllas) 11 'tlt'xh6, which transposes to
the notes to White's 11th in Line C1.
c) 10 o-o-o is probably White's best.
Now both 10 ... �xh6 and 10 ... bs 11 h4
�xh6 12 'tlt'xh6 lead to Line C1.
Cl) 9 .hh6 10 'tlt'xh6
•••

�8 27 �xcs+.
25 cxd6 l:td8
Black intends ....lii. g4 in this sharp
position.

C) 9 �h6
Very direct play, as White aims to
eliminate the King's Indian bishop.
Black can now invite the white queen
into h6 or he can simply begin his own
play on the queenside:

326
 Pa n n o Va ria tion: 9 0 - 0 - 0, 9 il h 6 a n d 9 g4
We have seen this idea before. Black
hopes that the white queen's absence
from the centre and queenside are
more important than its ominous­
looking post near the black king.
10 ... es
Having exchanged the dark-squared
bishops, Black logically places his
pawns on dark squares.
11 0-0-0
White can also play 11 dS tt:'ld4 (or
11 ... tt:Jas 12 tt:'lg3 cs 13 h4 Ji.d7 14 hs bs
which transposes to Line A of Chapter
14) and here:
a) 12 tt:Jxd4 exd4 13 tt:Je2 cs is much
already better for Black.
b) 12 'tlt'd2 cs (or 12 ... tt:Jxe2+ 13 Ji.xe2
ild7 14 g4 bS with counterplay) 13
dxc6 bxc6 (13 ... tt:Jxc6! ? 14 cs Ji.e6 1 S
cxd6 bS is a decent alternative) 14
tt:Jxd4 exd4 1s 'tlt'xd4 l:txb2 16 Ji.e2 tt:'lhs!
17 g3 'tlt'as 18 o-o Ji. h3 19 l:tfd1 .&1.c2 20
l:!.d3 l:!.b8 21 'tlt'xd6 1:!.bb2 gave Black the
initiative in A.Lewis-J.Hall, British
Championship, Morecambe 1981.
c) 12 o-o-o cs 13 dxc6 bxc6
(13 ... tt:Jxc6! ? 14 l:td1 Ji.e6 1s b3 'tlt'as
also looks quite comfortable for Black)
14 tt:Jxd4 exd4 1S 'tlt'xd4 l:txb2! and,
compared to the position with 9 h4 and
9 ... hs inserted (Chapter 1S, note 'b3' to
White's 10th move), Black does not
need to fear 16 cs l:tb8 17 l:txd6 'tlt'as 18
@c2 Ji.e6! because the g6-pawn is se­
curely defended and l:txe6 is not effec­
tive here.
11... bs 12 h4
Instead 12 dS tt:Jas 13 tt:'lg3 (even
worse for White is 13 cxbs axbs fol­
lowed by ... b4) 13 ...b4 (13 ...bxc4 14 h4
ild7 1S hS transposes to note 'b' to
Black's 13th in Line B of Chapter 14) 14
tt:Jb1 and now the simplest is 14...tt:'lb7
with the idea of ... tt:Jcs (Watson).
12 ... bxc4
Black has also often played 12 ... exd4
13 tt:Jxd4 tt:Jxd4 14 l:txd4 and here:
a) 14...'tlt'e7 1S hS Ji.e6 16 hxg6 fxg6
17 tt:Jds (not 17 cxbs axbs 18 Ji.xbs
Ji.xa2) 17 ... Ji.xds 18 cxds c6 19 dxc6
l:tbc8 20 @b1 l:txc6 21 l:td1 may leave
White with a slight edge.
b) 14... tt:Jhs 1s g4 tt:'lg3 16 l:tg1 tt:Jxf1
17 l:txf1 Ji.b7 (17 ... Ji.e6 looks better to
me) 18 hS 'tlt'f6 19 'tlt'e3 bxc4 20 l:txc4 cs
21 l:td1 Ji.c6 is unclear - White's c4-
rook is oddly placed.
c) 14... cs!? 1s l:td2 b4 16 tt:Jds tt:Jxds
17 l:txds Ji.e6 18 hs l:tb7! 19 l:td2 'tlt'e7
was solid for Black in M.Szymanski­
A.Mista, Trzebinia 2000.
13 hs 'tlt'e7
The position is similar to the varia­
tion 9 h4 bS (Line B of Chapter 14) and
could even transpose.
327
A ttacking C h e s s : The King 's I n d i a n , Vo l u m e 1
14 g4
Instead 14 dS tt:Jas 1s tt:'lg3 �d7 16
�e2 l:tb6!? (or 16 ...l:tb4, which is the 9
h4 bS main line) 17 l:td2 cs 18 tt:Jf1 (in­
stead 18 dxc6 l:txc6 19 tt:Jf1 �e6 20 tt:Je3
l::t c s is also unclear) 18...tt:Je8!? 19 tt:Je3
tt:'lc7 20 �d1 tt:Jbs 21 tt:'la4 l::t b 7, with a
double-edged position, has occurred in
a couple of games.
14 exd4 15 tt:Jxd4 tt:Jxd4 16 l:txd4 �e6
•••
17 �XC4!
This is a very creative idea. White
sacrifices a whole rook. The alternative
17 hxg6 fxg6 18 �xc4 �xc4 19 l:txc4
cs! leaves the rook looking funny on c4.
17 cs 18 hxg6 cxd4
•••
After 18 .. .fxg6 19 �xe6+ 'tlt'xe6 20
l:td2 White's rook was well placed and
he had some advantage in H.Lahlum­
F.Elsness, Norwegian League 1998.
19 tt:'lds �xds 20 �xds d 3
This move is necessary. Now Black
threatens 21 ... 'tlt'C7.
21 l::t h 2
Instead 21 @b1 'tlt'es 22 b3 'tlt'c3
leads to a draw after 23 l:!.c1 d2 24 l::t x c3
d1'tlt'+ or 23 gs 'tlt'c2+ 24 �a1 'tlt'c3+.
328
21 'tlt'es
•••
Black should not be tempted by
21 ... l:tbc8+ 22 �bl l:tc2 because 23 l:th1
l:te2 24 gs wins for White.
22 �b1
White could try 22 gS!. Then Black
should avoid 22 ... 'tlt'f4+ 23 �bl 'tlt'xf3 24
a3 ! when White wins and instead play
22 ... 'tlt'd4 23 gxf7+ l:txf7 24 �xf7+ �xf7
2S 'il¥xf6+ 'tlt'xf6 26 gxf6 �xf6, but after
27 l:td2! he still has to work to hold the
game.
22 d2 23 l:txd2 hxg6 24 'tlt'xg6+ �h8 25
•••
l:th2+ 'tlt'xh2 26 'tlt'xf6+
Here a draw was agreed in Yaplian­
A.Kuzmin, USSR 1982.
C2) 9... bS
This move is a bit riskier than
9 ... �xh6.
10 h4
Instead 10 o-o-o bxc4! was consid­
ered in the notes to White's 10th in
Line A.
A different kind of position is
reached after 10 �xg7 �xg7 11 cxbs
(11 0-0-0 bxc4 12 g4 es! looks good for
 Pa n n o Va riation : 9 0 - 0 - 0, 9 il h 6 a n d 9 g4
Black) 11...axbs 12 dS and now:
a) 12 ... tll e s 13 tlld4 b4 14 tllcbs e6
(after 14 ... Ji.d7 1S 'tlt'xb4 e6 Black has
good play, but White could consider 1S
.l:tc1 ! ?) 1 S tllc6 tllxc6 16 dxc6 es 17
'tlt'xb4 tllh s and Black has counterplay
brewing on the kingside. Then 18 o-o-o
Ji.a6 with the idea of ...tllf4-e6 gives
Black compensation for the pawn.
b) 12 ...tlla s 13 tlld4 b4 14 tllcbs es
1S 'tlt'xb4 (worse was 1S tllb3 tllxb3 16
axb3 Ji.d7 17 'tlt'xb4 tllh s! 18 g3 'tlt'gs
with a big advantage for Black in
T.Sohal-Zhe Quan, Edmonton 2009)
1s ... cs 16 dxc6 exd4 17 .l:tc1 'tlt'b6 18 c7
was F.Ludvigsen-K.Trygstad, Norwegian
League 1993. Now after 18 ... .l:tb7 it is
not clear that White has enough for t he
piece.
10 es
...
Instead 10...Ji.xh6 11 'tlt'xh6 trans­
poses to Line C1 after both 11 ... es 12
o-o-o and 11 ...bxc4 12 o-o-o es 13 hs.
Another move order which Black
can use is 10 ... bxc4!?, although he
should be careful after 11 Jlxg7 (11 hS
tllb4! is note 'a3' to Black's 10th move
in Line B of Chapter 14, while 11 o-o-o
tllb4 was covered in the notes to
White's 10th in Line A of this chapter)
11 ... �xg7 12 hS es 13 tllds and now:
a) 13 ... �h8 leads to the main line,
below.
b) 13 ... tll g 8 has also been played a
few times, but this looks too passive.
c) 13 ...tllxds 14 exds (14 hxg6 is not
sound after 14...tllf6 or 14...tllf4)
14... tllb4 transposes to note 'b' to
Black's 12th, below.
d) 13 ...Ji.e6 14 tllxf6 'tlt'xf6 1S dS tllb4
16 tll c3 intending Ji.xc4 looks much
better for White.
11 Ji.xg7 �xg7 12 hs
12 �hB!?
...
This leads to some spectacular
complications and was played in a
well-known game. Alternatively:
a) 12 ...tllxd4 13 tllxd4 exd4 and now
14 'tlt'xd4 gives White some advantage
(but 14 tllds cs, as in A.Ker-B.Spassky,
Wellington 1988, is less clear).
b) 12 ... bxc4 13 tlld s! tll x ds 14 exds
tllb4 1S tllc3 (not 1S hxg6? tlld 3+)
1S ...exd4 16 hxg6 (or 16 'tlt'xd4+ �g8 17
o-o-o Ji.fs) 16 ... 'tlt'e7+ 17 tlle4 fxg6 18
Jlxc4 and White is better because of
the open h-file and Black's unstable b4-
knight.
c) 12 ...tllxh S !? may actually be
Black's best: 13 g4 exd4 (not 13 ...tllf6 14
'tlt'h6+ followed by 1s tlld s) 14 tllxd4 (or
14 gxh s dxc3 1s tllxc3 gs with unclear
play) 14 ... tllxd4 1S 'tlt'xd4+ 'tlt'f6 16
'tlt'xf6+ tllxf6 17 cxbs axbs 18 Ji.xbs
ild7 is level.
329
A ttacking C h e s s : Th e King 's Indian, Vo l u m e 1
13 tt:Jds bxc4 14 hxg6 fxg6 15 'tlt'h6
After 15 o-o-o tt:Jxd5 16 exd5 Black
has the clever 16 ... c3! 17 bxc3 (worse
are 17 tt:Jxc3 tt:Jxd4 and 17 'tlt'xc3 tt:'lb4)
17 ... tt:'le7 when he has good chances.
1s ...tt::i h s
This is forced. Instead 15 ... .&!.f7? 16
'tlt'xg6 'tlt'g8 loses to 17 'tlt'xf6+!.
16 g4
After 16 o-o-o l:tf7 17 g4 tt:'lf6 18
'tlt'xg6 'tlt'g8 19 'tlt'xg8+ (now 19 'tlt'xf6+?
l:txf6 20 tt:Jxf6 fails to 20 ...'tlt'g5+)
19 ... tt:Jxg8 the ending is equal accord­
ing to Gufeld, but 20 �g2 ! ? may leave
White a little bit better.
16 ... l:txb2 17 gxhs
Black next two moves are forced -
he must keep lines closed.
11 ... gs 1s l:tg1 g4 19 o-o-o
Instead 19 dxe5 tt:Jxe5 20 o-o-o l:txa2
21 tt:Jef4 �g8 22 tt:Jg 6 hxg6 23 hxg6
'tlt'd7 ! led to a draw after 24 l:thl l:tal+
25 @b2 �5+ 26 �xal 'tlt'a4+ 27 �b2
'tlt'b3+ 28 �cl 'tlt'a3+ 29 �bl in Meyer­
G.Antoszkiewicz, correspondence 1984.
19 .. Axa2 20 tt:Jef4
White has to find a way to continue
his own attack others:
a) 20 �h3 l:txe2 21 �xg4 l:tf7 22
�xc8 'tlt'xc8 23 tt:'lf6 �8 24 l:tg8+ 'tlt'xg8
25 tt:Jxg8 tt:'lb4 26 l:td2 l:te1+ 27 l:td1 (not
27 @b2 l:txf3) 27 ... l:te2 with a repeti­
tion.
b) 20 dxe 5!? and here:
bl) 20...�d7? 21 tt:Jef4 l:txf4 (Black is
also not saved after 21 ... l:tal+ 22 �b2
l:txdl 23 tt:'lg6+ �g8 24 tt:Jde7+ tt:Jxe7 25
�xc4+) 22 tt:Jxf4 and White was win-
330
ning in Bowers-Frankema, correspon­
dence 1988.
b2) 20...l:txf3? 21 tt:Jef4 l:ta1+ 22 �b2
l:txdl 23 tt:'lg6+ �g 8 24 �xc4+ wins.
b3) 20 ... tt:Jxe5 21 tt:Jef4 �g8 22 tt:Jg6
hxg6 23 hxg6 (or 23 �xc4 tt:Jxc4 24
'tlt'xg6+ and White delivers perpetual)
23 ... 'tlt'd7 ! 24 l:th 1 l:tc2+! with impend­
ing perpetual check
20 ... exf4 21 tt:Jxf4?
White could have draw with 21
�xc4 l:ta1+ 22 �b2 l:txd1 23 l:txd1 l:tg 8!
24 tt:'lf6 l:tg7 25 �g8 ! 'tlt'e7 26 �xh7
l:txh7 27 tt:Jxh7 'tlt'xh 7 28 'tlt'f8+ and
again it's perpetual check
21 ...l:txf4 22 'tlt'xf4 C3! 23 �C4 J:ta3! 24
fxg4 tt:'lb4 2 5 �bl
2s ...�e6!
A fantastic clearance move. Black's
queen will come to b8 with devastating
effect.
26 �xe6 tt:Jd3 27 'tlt'f7 'tlt'b8+ 28 �b3
l:txb3+ 29 �c2 tt:'lb4+ 30 �xb3 tt:Jds+ 31
@c2 'tlt'b2+ 32 �d3 'tlt'bs+
Here White resigned in V .Bagirov­
E.Gufeld, USSR 1973, in view of 33 �c2
'tlt'e2+ 34 @b3 �2+ 35 @c4 �5 mate.
 Chapter 19
Pa nno Variation

9 a3, 9 I:.b1 and 9 I:.c1

1 d4 tt:'lf6 2 c4 g6 3 tt:'lc3 �g7 4 e4 d6 5

f3 o-o 6 �e3 tt:Jc6 7 tt:Jge2 a6 8 'tlt'd2 1:1.bB A: 9 a3
8: 9 nb1
C: 9 .:.c1

The first two variations are very

similar, but there are some small dif­
ferences. Line C is currently quite fash­
ionable and gives both sides a fair bit
of scope for creative play.

A) 9 a3

In this chapter we examine plans

where White plays to gain space on the
queenside. Even though he is slow to
get his kingside developed, White has a
lot of extra space and the position is
not exactly wide open. Still, White
should take some care because if the
position blows up, he may find his lack
of development is indeed an issue.
Black must play with intent, as even
though White's play is slow, it's not White simply wants to expand on
always so easy to create counterplay. the queenside with b2-b4. The down-

331
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
side to this move is that it weaken s b3,
which in turn can leave c4 vulnerable.
9 �d7
...
Black makes a useful waiting move
before committing to a queenside ad­
vance. There are several other option s
to consider as well:
a) 9 ...bs is also common and can
transpose to the main line, but there
are a couple of independent possibili­
ties after 10 cxbs axbs 11 dS (or 11 b4
�d7 12 dS tt:Jes 13 tt:'ld4 which trans­
poses to the main line) and here:
al) 11 ... tt:Jes 12 tt:'ld4 �d7 and now
in practice White always plays 13 b4,
which transposes to the main line.
However, he could consider grabbing
the bS-pawn at once when I do not see
anything special for Black. By playing
9 ...�d7 Black avoids this possibility.
a2) 11 ...tt:Jas is an alternative. It does
not have a very good reputation, but it
is not completely clear. After 12 tt:'ld4 es
(Black has also tried 1 2...�d7 1 3 b4 cs 14
dxc6 tt:Jxc6 1S tt:Jdxbs tt:Jes, but this looks
too speculative) 13 tt:'lc6 tt:Jxc6 14 dxc6
�e6 1S tt:Jxbs dS 16 �a7 Black has:
332
a21) 1 6...l:ta8 1 7 �cs l:te8 1 8 �e2
'tlt'b8 19 �4 did not give Black enough
compensation in K.Schmitzer-H.Alber,
Doernigheim 1994.
a22) 16 ... l:tc8 17 �cs l:te8 (after
17 ... dxe4 18 'tlt'xd8 l:tfxd8 19 �e7 l:te8
20 �xf6 �xf6 21 fxe4 White had the
better ending in M.Dlugy-J.Howell,
Sharjah 198S) 18 tt:Jc3 d4 19 tt:Jd1 and
now, instead of the overzealous
19 ... �dS?!, Black could try 19 ...l:tb8 or
19 ... tt:Jhs with some play for the pawn.
b) 9 ... es 10 dS tt:'le7 (it is hard to be­
lieve that the move 9 a3 hurts White
enough for this plan to suddenly be
good) and now both 11 b4 and 11 g4
look pleasant for White, while practice
has also seen:
b1) 11 tt:'lg3 tt:'ld7 12 �d3 fs 13 exfs
tt:Jxfs 14 tt:Jxfs gxfs 1S o-o-o (1s o-o is
probably better) 1S ...tt:Jcs 16 �c2 bS
and Black had decent counterplay in
N.Situru-E.Gufeld, Honolulu 1996.
b2) 11 o-o-o tt:'ld7 12 �bl fS 13 �1
tt:'lf6 14 h3 f4 1S �a7 l:ta8 16 �f2 tt:'lhs
17 �d3 gs 18 cs �d7 19 tt:Jb3 gave
White some initiative in L.Zsinka­
S.Cigan, Austrian League 199S, as his
queenside play is more dangerous than
anything Black has going on the other
side of the board.
c) 9 ... l:te8 is probably not the most
efficient use of a tempo. After 10 b4 bS
11 cxbs axbs 12 ds tt:Jes 13 tt:Jd4 �d7
both 14 �e2 and the immediate 14
tt:Jdxbs have scored well for White.
d) 9 ...e6 is a typical flexible move tn
these variations. Here the idea is little
 Pan n o Va ria tion: 9 a 3, 9 l:! b 1 a n d 9 l:!c 1
explored, although Black was doing
fine after 10 b4 �e7 11 l:!dl bS 12 cxbs
axbs 13 tt:Jc1 es 14 dS tt:'ld4 1S tt:J1e2
tt:Jxe2 16 �xe2 �d 7 17 o-o tt:'lh S 18 l:!c1
fS in J.Brenninkmeijer-R.Douven, Wijk
aan Zee 1988.
e) 9 ...tt:'ld7!? is a very interesting
idea. This flexible move gives Black sev­
eral possibilities: ...tt:'lb6 will attack the
c4-pawn and ...tt:Jas becomes feasible
because Black can now follow up with
... cs.
White has tried a few things here:
el) 10 �h6?! es! 11 �xg7 �xg7 12
dS tt:'le7 13 g4 c6 14 dxc6? tt:Jxc6 was
already better for Black in N. S ituru­
V.Tkachiev, Jakarta 1996. If 1S �xd6
then 1S ... �h4+ 16 tt:Jg3 l:!d8 gives Black
excellent compensation for the pawn.
e2) 10 h4!? is logical because Black's
knight has left his king's defence, but
after 10...tt:Jas (10...bs is also possible)
11 tt:Jc1 cs Black has good counterplay
according to Golubev.
e3) 10 l:!d1 ! ? tt:Jas 11 tt:Jc1 cs?! 12
dxcs dxcs 13 �xcs did not give Black
enough for the pawn in D.Vigorito-
R.Bumett, U SA 1994, so the flexible
10 ... e6!? could be considered.
e4) 10 b4 is the consistent move,
but after 10 ... tt:'lb6 the attack on the c4-
pawn is annoying:
e41) If White protects it with the
natural 11 tt:Jc1, then 11 ... es exploits
the weakened d4-square because after
12 dS (12 tt:Jb3 exd4 13 tt:Jxd4 tt:Jes is
good for Black as 14 �a2? just loses to
14 ... cs) 12 ... tt:'ld4 White's lack of control
over the b3-square is telling .
e42) 11 �d3 could be met with
11...e6!? intending ...dS!. Instead
11 ...�e6 is well met by 12 tt:Jds, but
Golubev has suggested 11 .. .fs.
e43) White, however, has another
way to cover the c4-pawn with 11 �a2.
This looks awkward, but White just
wants to consolidate his space advan­
tage: 11...e6 12 cs (12 l:!d1 dS! gives
Black good play) 12 ...tt:'ld7 (Black could
consider 12 ... dxcs!? with the idea of 13
bxcs tt:'ld7 intending ...b6, but 13 dxcs is
also possible, because the tempting
13 ...tt:'lc4 14 �xc4 tt:'les 1S �a2 tt:'ld3+ 16
�dl tt:Jf2+ 17 @cl tt:Jxhl looks too risky
333
A t tacking C h e s s : Th e King 's Indian, Vo l u m e 1
as the hi-knight cannot escape; here
13 ... ttJd7 is more sensible for Black) 13
.l:td1 dxcs (worse is 13 ...b6 ?! 14 cxd6
cxd6 15 dS!) 14 bxcs (14 dxcs �e7 is
unclear) 14 ... �e7 (Black could also con­
sider the immediate 14 ...b6!? because
15 dS exds 16 exds tt:Jces 17 c6 tt:Jcs is
very strong, while 15 cxb6 tt:Jxb6 looks
okay) 15 es (15 dS tt:Jces 16 d6 cxd6 17
cxd6 �d8 15 �c4 .l:td8 16 'itf2 tt:Jas 17
�b4 tt:'lc6 18 �c4 �f8 was unclear in
J.Sriram-V.Saravanan, Sangli 2000)
1s ... b6 16 cxb6 tt:Jxb6 17 tt:'lg3 .l:td8 18
�e2 �f8 ! and Black had a slight initia­
tive in B.Alterman-E.Sutovsky, Haifa
1996.
Now we go back to 9 ...�d7.
10 b4 bs
Black can still try to play without
this move:
a) 10 ... .l:te8 11 J::td1 and now 11 ... �c8
and 11 ... e6 are both possible, while
11 ...bS 12 cxbs axbs 13 ds tt:Jes 14 tt:'ld4
e6 15 dxe6 fxe6 16 tt:Jdxbs dS was a
speculative attempt in E.Godin­
C.Chase, Somerville 1996.
b) 10...es (Black switches plans now
334
that White has committed his queen­
side) 11 dS tt:'le7 and then:
bl) 12 g4 c6 13 tt:Jc1 bS gave Black
counterplay in A.Makarov-V.Aleshnya
correspondence 2002.
b2) 12 tt:Jc1 tt:'lhs 13 tt:'lb3 fs 14 o-o-o
has been played a few times. Now
14 ... b6!? discourages White's cs­
advance and looks reasonable for Black.
b3) 12 tt:'lg3 tt:Je8 (12 ... h s !?) 13 �d3 c6
14 o-o cxds was L.Zsinka-J.Nunn, Lon­
don 1987, and now 15 cxds should
leave White with an edge.
11 cxbs axbs 12 ds tt:Jes 13 tt:'ld4
This is the structure that White is
generally aiming for after both 9 a3
and 9 l::t b 1. The bS-pawn is a perma­
nent weakness, so Black must try to get
counterplay by breaking in the centre
with either his c- or e-pawn.
13 ...�eB
Black protects bS. Instead the im­
mediate 13 ... e6 is sometimes played
when White has:
a) 14 dxe6 fxe6 15 tt:Jdxbs (White
should avoid 15 �xbs �xbs when both
16 tt:Jcxbs tt:Jxe4! 17 fxe4 �h4+ 18 �d1
 Pan n o Va riatio n : 9 a3, 9 1:! b 1 a n d 9 1:! c 1
cs and 16 tt:Jdxbs tt:Jxe4! 17 fxe4 'tlt'h4+
18 @d1 l:txbs! bring him trouble)
1S ... tt:Jxe4 16 fxe4 �xbs (or 16 ... l:txfl+
17 l:!xf1 �xbs) 17 tt:Jxbs (even worse is
17 �xbs l:!xbs 18 tt:Jxbs 'tlt'h4+ 19 �d1
tt:Jc4 20 'tlt'e2 tt:Jxe3+ 21 'tlt'xe3 �xa1)
17 ...l:txfl+ 18 l:txf1 tt:Jc4 19 'tlt'e2 tt:Jxe3
20 'tlt'xe3 �xa1 and Black has excellent
play.
b) 14 �e2 ! is similar to the main
line and scores very well for White. In­
deed, after 14...exds 1 s exds l:te8 16
o-o tt:Jc4 (16 ...'tlt'c8 17 tt:Jdxbs is also in­
sufficient) 17 �xc4 bxc4 18 bS 'tlt'c8 19
a4 he is certainly better.
14 �e2
14...e6
Instead 14 ... tt:'lc4 1S �xc4 bxc4 16
o-o es 17 tt:Jde2 tt:'lhs was B.Alterman­
M.Hennigan, Santiago 1990. Here
White could just go after the c4-pawn
with the typical idea of 18 'tlt'a2! when
Black probably does not have enough
counterplay.
Black's best is likely 14... c6!? 1S dxc6
�xc6 16 o-o (or 16 tt:Jxc6 'tlt'xc6 17 o-o
tt:Jc4 when Black looks to be okay)
16 ... �d7 when White is still probably a
little better, but Black is in the game
and can try to play ... e6 and ... ds or
.. ltb7, ...'tlt'b8 and ...l:!c8.
15 o-o exds
Alternatively, 1s ...tt:Jc4 16 �xc4 bxc4
17 l:tfe1 es 18 tt:Jde2 'tlt'd8 19 'tlt'a2! was
also insufficient for Black in M.Dlugy­
J.Fedorowicz, Sth matchgame, New
York 1984.
16 exds tt:Jc4 17 �xc4 bxc4 18 l:tfe1
Black is in difficulties - he lacks
counterplay and the c4-pawn is weak.
White may also continue with bS and
a4, with a grip on the queenside.
B) 9 l:tb1
This little rook move is similar con­
ceptually to 9 a3. White simply intends
to play b4. There are a couple of differ­
ences though. By playing 9 l:tb1 White
is no longer able to castle queenside.
Also, the rook will be tied to the de­
fence of the b4-pawn in most positions,
so bringing it to a more active square
such as cl or d1 is less likely. Another
odd little point is that White's queen
335
A ttacking Ch ess: Th e King 's I n dian, Vo l u m e 1
lacks the a2-square - we saw in Line A
that White can often attack a pawn on
c4 with 'tlt'd2-a2. On the plus side,
White has avoided weakening the b3-
square, so playing tt:Je2-c1 becomes
possible, because after the sequence
... es; dS tt:'ld4 White can challenge Black
with tt:'lb3.
All in all, it is difficult to say which
of 9 a3 and 9 l:tb1 is 'better', but I will
try to point out the subtle differences
in the play that arises.
9 �d7
...
Again, there are several alterna­
tives:
a) 9 ... es perhaps makes a little more
sense here than in Line A, as White
cannot castle queen side, but 10 dS tt:'le7
11 tt:'lg3 (or even 11 g4) should still fa­
vour White.
b) 9 ... e6!? is rare, but rather interest­
ing.
Then 10 b4 tt:'ld7 11 tt:Jc1 (probably
not the best) 11 ...'tlt'h4+!? (this looks
better than 11 .. .fs, which has been seen
a couple of times in practice) 12 �f2
'tlt'f6! shows the downside to 9 .l:tb1 -
336
the b4-pawn needs to be tended to.
After 13 tt:J1e2 (not 13 tt:Jb3 tt:'lxb4)
13 ...tt:'lb6 Black has good play.
c) 9 ... tt:'ld7 is possible here as well: 10
h4! ? (after 10 b4 tt:'lb6 11 tt:Jc1 Black
could play 11 ...e6!? or 11 ...eS 12 dS tt:'ld4
13 tt:'lb3 tt:Jxb3 14 .l:txb3 fS with decent
play) 10 ... hs (worse is 10 ... es 11 dS tt:'le7
12 hS fS. while 10 ...tt:Jas makes less
sense here because b3 is not weakened;
White might continue, for instance,
with 11 tt:Jc1 cs 12 dxcs dxcs 13 h S!) 11
�gs ! ? �f6 12 f4 (this is a rather
strange move) 12 ... es (Golubev sug­
gests 12 ... �g7 ! ? 13 fs tt:'lb6) 13 tt:Jds
�xg s 14 hxgs tt:Jxd4 1s tt:Jxd4 exd4 16
g4 c6 17 gxhs cxds 18 hxg6 fxg6 19
'tlt'h2 �7 20 'tlt'h7+ @e8 21 'tlt'xg6+ .l:tf7
was E.Ghaem Maghami-A.Volokitin,
Lausanne 2001. Now 22 'tlt'xd6 looks
best: for example, 22 ... 'tlt'e7 23 .l:th8+
l:tf8 24 .l:txf8+ @xf8 2S 'tlt'xe7+ @xe7 26
cxds tt:Jcs 27 es and here White is bet­
ter.
d) 9 ... bs is the most direct. After 10
cxbs axbs White has:
d1) 11 dS tt:Jas (or even 11 ...tt:Jes 12
tt:'ld4 b4 13 tt:Jcbs e6!?) 12 tt:'ld4 b4 13
tt:Jcbs es 14 tt:Jc6 (not 14 dxe6 fxe6 1S
'tlt'xb4? cs) 14...tt:Jxc6 1S dxc6 when
Black can simply play the straightfor­
ward 1s ... ds or even 1S ...tt:Jxe4!? 16
fxe4 'tlt'h4+ 17 @d1 'tlt'xe4, with sharp
play that does not look unfavourable
for him.
d2) 11 b4 and now 11 ...�d7 trans­
poses to the main line, but Black often
prefers 11 ...es!? 12 dS tt:'le7.
 Pa n n o Va r i a t i o n : 9 a 3 , 9 l:! b 1 a n d 9 l:!c1
Here White has:
d21) 13 tt:'lg3 �d7 and now:
d211) 14 �d3 c6 1S dxc6 �xc6 16
o-o dS 17 �cs d4 18 tt:Jce2 tt:'ld7 19 �d6
l:!c8 20 f4 was A.Dreev-Z.Efimenko,
Moscow 2002. Here Dreev considers
20... hs to be unclear, while Hazai's
suggestion of 20 ...exf4 21 'tlt'xf4 fs 22
exfs tt:Jes! 23 �xes 'tlt'ds is interesting
as well.
d212) 14 �e2 c6 1S dxc6 �xc6 16
o-o dS 17 �cs (or 17 exds tt:Jexds 18
�cs l:!e8 with unclear play in
1.Khenkin-J.Balcerak, German League
2001) 17 ... d4 (17 ... l:!e8! ?) 18 tt:Jd1 tt:'ld7
(worse is 18 ... tt:Je8, as in S.Savchenko­
A.Fedorov, St Petersburg 1996) 19 �d6
l:!c8 20 tt:Jb2 tt:'lb8! 21 �cs tt:Ja6 and
Black's impending possession of the
bishop-pair gave him good chances in
S.Jenkinson-C.McKenzie, correspon-
dence 2001.
d22) 13 tt:Jc1 is a favourite of Haba's:
13 ... �d7 14 tt:Jb3 tt:'lh s (Black could also
consider 14... c6 1S dxc6 �xc6 16 tt:Jas
�a8 with the idea of ... dS) 1s �d3
(White does well to get himself castled;
instead 1s tt:Jas fs 16 �d3 fxe4 17 fxe4
c6 left him with an uncomfortably­
placed king in P.Haba-P.Bobras,
Koszalin 1999) 1s ... tt:Jf4 (Golubev sug­
gests 1s .. .fs 16 o-o f4 17 �a7 l:tb7 18
�f2 gs 19 tt:Jas l:!b8 when Black may
lose the bS-pawn, but he has some
kingside counterplay) 16 o-o tt:Jxd3 17
'tlt'xd3 fs 18 tt:Jas f4 19 �a7 l:!a8 20 �f2
has been reached a couple of times
with by Haba. He won both games, but
was considerably higher rated in each.
White has made some progress on
the queenside, but Black has the
bishop-pair and potential kingside
play:
d221) 20 ... .rf.b8 21 a4 bxa4 22 tt:Jc6
tt:Jxc6 23 dxc6 �e6 24 tt:Jds l:!f7 2s bS
gave White the initiative in P.Haba­
F.Voelzke, Kiel 2001.
d222) 20...'tlt'e8 21 l:!fc1 gs 22 l:!b2
�f6 23 tt:Jb1 was P.Haba-M.Raubal,
Loosdorf 1993, and now 23 ...\ib8 24
l:!bc2 l:!c8 would be unclear.
e) 9 ...l:!e8 is more sly here than it
was in Line A because White cannot
castle long. After 10 b4 Black has:
337
A t tacking C h e s s : The King 's In dian, Vo l u m e 1
el) 10... �d7 is rather vague. White
could play 11 g3, 11 tt:Jc1, 11 a4 or even
11 bS!?.
e2) 10 ... es 11 dS tt:'ld4 (11 ... tt:Je7 is
better, but then ... .rf.e8 looks silly) does
not work after 12 tt:'lxd4 exd4 13 �xd4
tt:Jxe4? 14 fxe4 l::txe4+ 1S tt:Jxe4 'lli'h4+ 16
tt:Jf2 �xd4 17 �d3, winning.
e3) 10...bs leads to play similar to
what we have already seen. White is bet­
ter after 11 cxbs axbs 12 dS tt:Jes 13 tt:Jd4
�d7 14 tt:Jcxbs e6 1S dxe6 fxe6 16 �e2.
e4) 10 ... tt:Jd7 !? sees Black aim for
similar play to 9 a3 tt:'ld7 (note 'e' to
Black's 9th move in Line A).
White has:
e41) 11 cs?! dxcs (11 ...bs!?) 12 bxcs
(even worse is 12 dxcs tt:'ldeS) 12 ...b6 13
dS tt:Jas 14 tt:Jf4 was Zhang Zong­
Y.Kruppa, Moscow 2004. Now 14...tt:Jes
with the idea of 1S .l:tb4 bS, controlling
the c4-square, is pleasant for Black.
e42) 11 �gs was suggested by
Golubev. Here 11 ... tt:Jb6 12 'tlt'd3 (not 12
tt:Jc1? tt:Jxd4) 12 ... h6!? 13 �h4 (or 13
�e3 e6!) 13 ... as!? 14 a3 (14 bS? tt:'lb4)
14...axb4 1S axb4 .l:ta8 is unclear.
338
e43) 1 1 tt:Jc1 e s 1 2 dS tt:'ld4 1 3 tt:'lb3
tt:'lxb3 (alternatives are 13 .. .fs and
13 ... 'tlt'M+ 14 �f2 'tlt'f6 when Black is
supporting d4) 14 l::txb3 fS and al­
though the e8-rook is misplaced, Black
has a fairly normal and playable posi­
tion.
e44) 11 a3 tt:'lb6 12 tt:Jc1 es (in stead
12 ... e6 13 cs tt:'ld7 14 tt:'lb3 b6 1S cxd6
cxd6 16 bS was a little better for White
in V.Malakhatko-J.Radovanovic, Port
Erin 2oos) 13 dS tt:'ld4 14 tt:'lb3 tt:'lxb3
(again 14.. .fs and 14 ... 'lli'h 4+! ? 1s �f2
'tlt'f6 are alternatives) 1S .l:txb3 fS is
similar to variation 'e43'.
e4S) 11 h4!? tt:'lb6 (Black could con­
sider 11 ... hs, but 11 ... es 12 dS tt:Je7 13
hS fS 14 hxg6 hxg6 1s �h6 tt:'lf6 16 exfs
was a little shaky in J.Santos-E.Agdes­
tein, Thessaloniki Olympiad 1984) 12
tt:Jc1 (12 tt:'lg3 could be met with 12 ... hs
or 12 ... es 13 ds tt:'ld4 14 hs 'tlt'e7) 12 ... es
13 dS tt:'ld4 14 tt:'lb3 probably favours
White somewhat, e.g. 14... tt:'lxb3 1S
l::t x b3 hS 16 �e2 and the position is
similar to some we saw in Chapter 16.
Finally, we return to 9...�d7:
 Pan n o Varia ti o n : 9 a 3 , 9 :t:! b 1 a n d 9 :t:! c 1
10 b4 b s
After 10...es 11 dS CfJ e 7 both 1 2 CfJg3
and 12 g4 look good for White.
11 cxbs axbs
Black should be very conscious of
the type of passive positions that we
saw in the main line of 'A' and so try to
at least cause White some problems.
12 ds CiJes 13 CiJd4 'tlt'e8
After 13 ... e6 14 �e2 (14 dxe6 fxe6
15 CiJdxbs CiJxe4 is less good than the
analogous operation in Line A because
here there is no rook hanging on al!)
14...exds 15 exds .rf.e8 16 o-o White has
a familiar advantage.
14 �e2
14 ...e6
Black should consider too 14 ... cs!?
1 5 dxc6 �xc6 16 o-o �d7.
15 0-0 CfJc4 16 �XC4 bXC4 17 .l:tfel e5
White has a choice:
a) 18 CiJc2 CfJhS 19 CiJa3 fS 20 exfs
�xfs 21 CfJe4 was 5 .Savchenko­
M.Golubev, Alushta 1995. Now Golubev
gives 21 ...CfJf6! 22 �gs (22 CiJxf6+ �xf6
23 .l:tbc1 'tlt'a4 is good for Black) 22 ... C3!
with good play for Black.
b) In Line A the position after 18
CiJde2 was good for White. The only
difference here is that White has
played .l:tb1 instead of a3, but now he
does not have 'tlt'a2 at least. Here
18 ...CfJh S 19 g4 CfJf6 20 CfJg3 is still bet­
ter for White according to Golubev, so
perhaps Black could try 19 ...CfJf4!? 20
CiJxf4 exf4 21 �xf4 .l:ta8 with counter­
play; his pieces are active and White's
king position has been loosened.
C) 9 l:tc1
This line is very fashionable. White
does not initiate any action just yet,
but he does prevent 9 ...bs? because 10
cxbs axbs 11 CiJxbs .rtxbs 12 .rtxc6
would win a pawn. Moreover, if Black
plays too passively, White will slowly
develop his kingside and look forward
to a nice space advantage in the mid­
dlegame, so Black must play with some
purpose.
9 �d 7
...
Here too this is a good, flexible
move. Black also protects the c6-knight
which will allow him to play ...bs. In-
339
A ttacking C h e s s : Th e King 's I n d i a n , Vo l u m e 1
stead both 9 ... tLias and 9 ... tLld7 allow
White to happily play his intended 10
t2Jd1!. Other moves also seem insuffi­
cient:
a) 9 ...ifd7 (the novel idea is to pro­
tect the c6-knight with the queen,
while ...�d8 becomes a possibility)
10 tLlds (after 10 t2Jd1 bS 11 cs dxcs
12 .l:!.xcs .l:!.d8 13 tLif2 e6 14 g3 l2Je8
Black has achieved his desired set-up
and stands well according to Kaufman)
10 ....l:!.e8 (not 10...bs? 11 cxbs axbs 12
�xc6!, while 10...e6 11 tLixf6+ il.xf6 12
t2Jc3 is slightly better for White accord­
ing to Kaufman) 11 h4 h S 12 g3 lLih7 13
il.g2 es 14 o-o ifd8 1s .l:!.fd1 lLif8 16 cs
il.e6 17 cxd6 cxd6 18 dxes il.xds 19
ifxds dxes 20 ifxd8 �exd8 21 iLh 3 and
the bishop-pair gave White a healthy
advantage in A.Dreev-A.Cabrera, Albox
200S.
b) 9 ... es 10 dS lLie7 11 lLig3 (or 11 cs)
11... l2Je8 12 il.d3 fs 13 exfs tLixfs
(13 ... gxfs 14 tLihs) 14 tLixfs gxfs 1s o-o
looks like a worse version of the classi­
cal 6... es line - Black's queenside moves
are pointless here.
340
c) 9 ... e6!? was Black's choice in a
game between two members of the
world's elite and could certainly use
further testing.
White has:
cl) 10 g3 can be met by 10...ifd7!?
when Kaufman gives both 11 tLlf4 es
12 dxes tLixes and 11 il.g2 bS!. Instead
10 ...t2Je7 11 il.g2 bS 12 b3 c6 13 o-o as
14 es! l2Je8 1S f4 likely gave White an
edge in R.Ponomariov-S.Ganguly, Span­
ish Team Championship 2010, al­
though even here Black wasn't without
his counter-chances and might have
preferred 1S ...bxc4!? 16 bxc4 il.a6 to
the stodgy 1s .. .fs of the game.
c2) 10 t2Jd1 could be met by
10 ... t2Jd7 (A.Gupta-A.Lahiri, Bhubanes­
war 2009) or 10...t2Je7 (V.Malakhatko­
K.Neumeier, Oberwart 2006).
c3) 10 il.g s was suggested by
Golubev. Here Black should play 10 ...bS!
11 cxbs axbs 12 tLixbs �xbs 13 il.xf6
(or 13 .l:!.xc6 tLixe4!) 13 ...il.xf6 14 �xc6
il.b7 1s �c2 il.gs 16 ifd3 �b6 when he
has compensation for the pawn
according to Kaufman.
 Pa n n o V aria t i o n : 9 a 3, 9 I:. b 1 a n d 9 I:. c 1
c4) 1 0 b3 tLid7 11 h4! ? (Kaufman
suggests 1 1 g3 tLie7 12 il.g2 bS 13 tLid1
cs 14 dxcs dxcs 1s o-o lLic6 16 f4)
1 1...hs 12 g3 tLie7 13 g4! ? hxg4 14 fxg4
lLif6 1S gS lLihS 16 il.g2 es 17 il.f3 bS 18
dxes il.xes 19 t2Jf4 il.xf4 (or even
19 ...tLixf4 20 il.xf4 bxc4) 20 il.xf4 bxc4
21 bxc4 il.e6 was unclear in V.Anand­
A.Grischuk, Linares 2009.
1o tLid1
This is another important part of
White's plan. Now ... bs can be met with
cs and the knight will swing over to f2.
Instead 10 dS tLies with the idea of
...cs should be fine for Black, while 10
b3 bS (Black could also try 10 ... es 11 dS
tLie7) 11 cxbs axbs 12 dS tLies 13 lLid4
ife8 14 il.e2 (14 a3 b4 looks okay for
Black) 14 ... e6 (14...b4! ?) 1S b4 leads to:
a) 1s ...exds 16 tLixds tLixds 17 exds
l2Jc4 18 il.xc4 bxc4 19 I:.xc4 il.bs was
V.Raicevic-M.Golubev, Belgrade 1991.
Here White should have played the
simple 20 I:.xc7 according to Golubev
when Black does not have enough for
the pawn.
b) 1s ...l2Jc4 16 il.xc4 bxc4 17 bS exds
18 exds ifes 19 lLic6 il.xc6 20 dxc6
I:.fe8 21 �f2 was D.Krumpacnik­
M.Muhutdinov, Budapest 1991, and
now 21 ... .i::txbs!? 22 tLixbs ifxbs would
give Black very good compensation for
the exchange.
10 bs
...
Thi s is consistent, even though
White can 'push by' the black b-pawn.
Others:
a) 10... es 11 dS l2Je7 12 cs (or 12 tLif2
lLie8 13 g3 cs 14 b4 b6 1s il.g2 fs 16 o-o
lLif6 17 I:.b1 with some advantage to
White in J.Rowson-K.Arakhamia Grant,
British Championship, Scarborough
2004) 12 ... tLle8 (here Black should con­
sider the rather unusual 12 ... dxcs 13
il.xcs I:.e8 with the idea of ... c6, trying
to use his development lead) 13 b4 fs
14 tLiec3 l2Jf6 1S il.e2 �h8 16 tLib2 lLic8
17 l2Jc4 was very good for White in
A.Graf-J.Nunn, German League 2001.
White's position looks like a very good
version of a Classical Variation.
b) 10 ... e6 11 tLif2 I:.e8 (11 ...tLle8 12
g3 fs 13 il.g 2 lLif6 14 o-o fxe4 1s fxe4
ife7 16 b4 es 17 dS lLid8 18 cs was
341
A ttacking C h e s s : The King 's I n dian, Vo l u m e 1
good for White in E.Tomashevsky­
A.Fier, World Team Championship,
Bursa 2010, but Black might retain the
tension with Schmaltz's 14...ife7) 12
iLgS ife7 13 g3 es 14 dS l2Jd4 1S t2Jxd4
exd4 16 il.g2 iff8 17 o-o h6 18 il.xf6
il.xf6 19 iLh3 il.xh3 20 tLixh3 c6 was
fine for Black in A.Dreev-D.Kokarev,
Mumbai 2010.
c) 10...�e8 11 tLif2 bS 12 cs dxcs 13
.l:!.xcs es 14 dS lLie7 1S ifc2 c6 (a better
try is 1s ... il.f8!? with the idea of 16
.l:!.xe7 tLiexdS! 17 exds tLixds with active
play) 16 dxc6 tLixc6 17 �xc6 il.xc6 18
ifxc6 and Black's compensation was
insufficient in A.Graf-J.Nunn, German
League 2003.
d) With 10... as Black switches to
more traditional King's Indian methods
and tries to hold White up on the
queen side.
Practice has seen:
d1) 11 t2Jdc3!? es 12 dS lLie7
(12 ...t2Jb4!? may be a better try) 13
lLibS ! b6 14 g4! ? l2Je8 1S lLig3 c6 (Black
could also lash out with the thematic
1s ... fs) 16 dxc6 tLixc6 17 h4 tLle7 and
342
now, instead of 1 8 tLixc7 ifxe7 19 h S
l2Jd4 which gave Black reasonable play
in 1.Khenkin-Z.Efimenko, German
League 2008, White could try the an­
noying 18 il.gs ! f6 19 il.e3 with a
pleasant advantage.
d2) 11 g3 b6 12 il.g2 es 13 dS lLie7
14 o-o tLihs 1S tLif2 (Hazai suggests 1S
g4 lLif6 16 lLig3 l2Je8 17 t2Jc3 fs 18 gxfs
gxfs 19 f4 with a clear advantage)
1s ...fs 16 f4! (Black had sufficient play
after 16 b3 .l:!.f7 17 a3 iff8 18 t2Jd3 h6 19
b4 axb4 20 axb4 gs 21 exfs tLixfs 22
il.f2 lLif6 in l.Novikov-Z.Efimenko,
Montreal 2004) 16 ... exf4 17 tLixf4 lLixf4
18 il.xf4 and White's space advantage
gave him the more pleasant position in
E.Gasanov-A.Fedorov, Minsk 2006.
11 C5
11 cxbs axbs has also been played,
but this looks strange and should not
cause Black any problems.
11 e6
...
Alternatively, 11 ...dS 12 es l2Je8
gives up a lot of space, as does 11 ... es
12 dS tLle7 13 c6, while 11 ... dxcs 12
�xcs leaves Black a bit looking a bit
 Pa n n o Va r i a t i o n : 9 a 3 , 9 I:. b 1 a n d 9 �c 1
soft on the c-file, but the position is still
not so clear after 12 ...e6 13 g3 ike7.
12 l2Jf2 b4
Instead 12 ... ike7 is just met by 13
g3, while 12 ... t2Je7 13 t2Jd3 makes it dif­
ficult for Black to move his queenside
pawn s. However, the text move gains
some space and will allow Black to ac­
tivate his minor pieces on the queen­
side.
13 g3
White has a couple of other moves
worth considering here.
a) 13 t2Jd3 as 14 g3 t2Je7 1s il.g2
il.bs 16 o-o t2Jd7 17 �fd1 �a8 18 ikc2
ikc8 19 a4 il.a6 20 b3 .rf.d8 21 f4 dS and
Black held firm in 1.Khairullin­
E.Gorovykh, St Petersburg 2010.
b) 13 lLig3 as 14 il.d3 t2Je7 1S 0-0
il.bs 16 tLie2 t2Jd7 17 cxd6 cxd6 18 b3
il.xd3 19 t2Jxd3 dS 20 es ikb6 was also
solid for Black in M.Nikolov-K. Berbatov,
Kyustendil 2010.
13 ...tLie1 14 il.g2 il.bs
Black has less space, but his position
is flexible and not without dynamic
potential.
15 0-0 tLld7
Also possible is 1s ...l2Jc6! ?.
16 cxd6
White does not need to hurry with
this exchange, but it is also not clear
what else he should do. Indeed, in prac­
tice Black has found good counterplay:
a) 16 f4 dxcs 17 dxcs tLie s! 18 fxes
'i!t'xd2 19 il.xd2 il.xe2 20 �fe1 il.bs 21
l2Jg4 tLic6 22 l2Jf6+ �h8 23 il.f4 gs 24
il.xg s tLixes gave Black good play in
l.Gromova-E.Akatova, Chelyabinsk
2008.
b) 16 I:.fd1 as 17 ikc2 ?! ikc8 18 f4
ika6 19 I:.d2 es (perhaps even better is
19 ... ds!? 20 es tLlfs) 20 fxes dxcs 21
dxcs tLixes and again Black had good
play in A.Stanoev-K.Berbatov, Plovdiv
2010.
16 ...cxd6
The Croatian IM Srjdan Sale and the
Belorussian GM Alexei Fedorov have
debated this position a couple of times.
Black has an interesting Hedgehog­
type position:
a) 17 I:.fd1 as 18 il.f1 �c8 19 t2Jd3
�xc1 20 I:.xc1 ika8 21 b3 �c8 22 tLib2
'i!t'a6 23 t2Jf4 il.xf1 24 :i'.xf1 dS left Black
doing well in S.Sale-A.Fedorov, Abu
Dhabi 200S.
b) 17 .rtc2 and now Black tried
17... dS 18 �fc1 es in S.Sale-A.Fedorov,
Abu Dhabi 2006, but White looks better
prepared for the opening of the centre.
Instead Kaufman proposes 17 ...ikas 18
b3 �fc8 19 �fc1 ikd8 with equality.
343
 c'hapter 20
·pa:nno .variation
Other Lines

1 d4 lLif6 2 c4 g6 3 l2Jc3 il.g7 4 e4 d6 5 W e begin our coverage of White's

f3 o-o 6 il.e3 lLic6 early deviations with Line A, where
White delays or omits playing tLige2.
Line B instead covers lines where White
delays or omits playing ii'd2. In Line C
we come back to the main line 7 tLige2
a6 8 ii'd2 .l::tb 8 and here we look at what
happens when White plays something
we have not covered yet on move 9,
namely 9 �d1 and 9 dS. Line D is the
Westerinen system with 7 ...�b8 and
8 ... �e8. This is a tricky line and if White
is inaccurate then Black gets a wonder­
In this chapter we look at some odds ful version of the ... cs pawn sacrifice.
and ends, as well as a couple of other Direct approaches are more dangerous,
systems for Black. however. Then in Line E we look at a
kind of mixed system where Black
A: 7 1i'd2. a6 8 0-0-0 · omits ... .l::tb8 and instead combines ...a6
9� 7 �ge2 a6 8 �Cl with ....l::te 8. Finally, Line F covers 6 il.gS
C: 7 �e2 a6 8 •d2 :b8 9 lc11 l2Jc6.
and 9 d5 Note that in the Panno the fairly
D: 7 �ge2 llb8 8 •d2 :ea fashionable move 6 tLige2 does not
E: 7 �ge..2 a6 8 •dz :ea have any real independence after
F: 6 .igs �c6 6 ... l2Jc6, .as 7 il.e3 just inverts White's
usual 6th and 7th moves while 7 il.gs

344

transposes to Line F. Instead White
could play 7 dS, but after 7 ... tLies nei­
ther 8 lLig3 c6 nor 8 l2Jd4 cs look dan­
gerous. Black can also meet 6 tLige2
with 6 ... a6!?. Then White has hardly
anything better than 7 il.e3 or 7 il.gs
when in both cases 7...l2Jc6 just brings
us back to normal lines.
A) 7 ii'd2 a6 8 0-0-0
There is really no reason to castle so
soon, but this move has been played by
some strong players. Instead 8 il.d3 also
looks a bit naive, but it is not so terrible.
After 8 ... es 9 dS l2Jd4 Black may not be
better but he should not be worse: 10
tLige2 tLihs!? (10 ... t2Jd7 11 o-o cs 12 dxc6
bxc6 13 b4 was a little more pleasant for
White in G.Kamsky-J.Hjartarson, Biel
lnterzonal 1993) 11 o-o and now Black
could play 11 .. .fs or 11 ...cs.
s .. bs!?
. with 9 h4.
Black seeks to punish White for his
committal play. It is also possible to
play 8 ...�b8 when White should proba­
bly just play 9 tLige2, which transposes
to Line A of Chapter 18, although he
Pa n n o Va ria tio n : O t h e r L in es
might try one of:
a) 9 g4 e s 10 dS l2Jd4 11 t2Jce2 (not
11 tLige2? tLixf3) allows the interesting
11 ...tLixf3! ? 12 tLixf3 il.xg4 13 il.g2
tLixe4 with good compensation for the
piece.
b) 9 il.h6 il.xh6 10 ii'xh6 es 11 dS
(instead 11 tLige2 bS transposes Line C1
of Chapter 18) 11...t2Jd4 12 tLige2 cs (or
even 12 ... tLixe2+ 13 il.xe2 il.d7) 13 dxc6
bxc6 14 t2Jxd4 exd4 1S .i::txd4 .i::tx b2! is
good for Black, as we saw in note 'c' to
White's 11th move in Line C1 of Chap­
ter 18.
c) 9 h4 and here 9 ...es seeks to pun­
ish White's omission of tLige2. After 10
dS l2Jd4 11 tLige2 cs 12 dxc6 bxc6 13
t2Jxd4 exd4 14 il.xd4 il.e6 Black has
managed to play the pawn sacrifice
with White's king committed to the
queenside. Still, I am not sure if I be­
lieve in this version, as White's king is
not in terrible danger and he has not
spent a lot of time moving his king's
knight, which is usually some justifica­
tion for offering the pawn. Instead of
all this, Black could simply play 9 ... h s
with a likely transposition t o the main
lines of Chapter 1S.
9 cxbs
White should probably consider ig­
noring the offer and getting on with it
9... axbs 10 il.xbs tLias!
Black already has excellent play.
With the kings castled on opposite
wings, the initiative is worth more than
the pawn.
345
A ttacking Chess: Th e King 's Indian, Vo l u m e 1

1 2 il.xa6 .i::txa6 1 3 ifd3 ifa8! 14 tLige2

.i::tb 8

11 �b1
White has also tried 11 il.h6 c6 12
il.xg7 �xg7 13 il.d3 il.e6! (this was Black clearly has an excellent game.
queried by Hjartarson in his notes, but It looks as if White has castled queen­
it looks good to me; instead 13 ... ifb6 14 side in a Benko Gambit.
es tLlds 1s exd6 ! exd6 16 tLixds cxds 17 15 il.c1 e6!
tLie2 gives White better chances to Black creates an outpost on c4 for
weather the storm) 14 b4? (this is the his knight.
only move according to Hjartarson, but 16 h4 ds 17 hs l2Jc4 18 hxg6 hxg6 19 b3
it looks insane; that said, 14 �bl 'i!t'b6 cs! 20 dxcs tLld7!
gives Black excellent compensation) All of Black's pieces are participat­
14...l2Jc4 1S il.xc4 il.xc4 16 ifb2 ifb6 17 ing in the attack.
lLih3 �fb8 18 .l:!.d2 (or 18 a3 ifa6 19 21 exds exds
tLib1 cs - Hjartarson), and now 18 ... es Even stronger is 21 ... .l:!.xa2 ! !
1 9 .l:!.hd1 ifxb4 20 ifxb4 .i::txb4 gave
Black some advantage in
L.Christiansen-J. Hjartarson, Szirak ln­
terzonal 1987, but with 21 dxes dxes
22 tLif2 .l::ta3 23 �c2 White could have
put up a lot of resistance. Instead
18 ...ifxb4 19 ifxb4 .l::txb4 is more flexi­
ble and looks good for Black, but
18 ... ifa6! with the idea of ... cs looks
best of all.
11...il.a6
Also strong is 11 ... c6 12 il.d3 il.e6 2 2 tLixa2 (instead 2 2 ifxc4 tLies!
with excellent play. wins the queen which is tied to the de-

346

fence of the b3-pawn) 22 ... t2Ja3+
(22 ... ifa4!? also looks good, but the text
is simpler) 23 il.xa3 ifxa3 24 l2Jd4 tLixcs
2S ifc2 tLixb3 26 'i!t'b2 t2Jd2+ 27 �xd2
�xb2+ 28 .l:!.xb2 il.xd4 and although the
material balance is similar, Black has a
much better version of the game con­
tinuation.
22 ifxds .i::tx b3+ 23 axb3 .!:!.al+ 24 �c2
�xcl+ 25 tLixcl lLie3+ 26 �bl tLixds 27
.i::t x ds il.xc3 28 �xd7 ifa3
Black still has the initiative, but
White managed to hold out in M.Petur­
sson-E.Gufeld, Hastings 1986/7.
B) 7 tLige2 a6
8 tLlcl
Here we look at lines where White
delays playing ifd2. Moves like 8 .i::tb1
and 8 a3 are likely to transpose to
Chapter 19, although the latter gives
Black the interesting option of 8 ... il.d7
9 b4 'ir'b8!? supporting ... bs with the
queen.
After 8 h4 there are many transposi­
tional possibilities. Here 8 ... es 9 dS tLias
10 lLig3 cs 11 hS :i'.b8 12 if d2 and
Pa n n o Va ria t i o n : O t h e r L i n e s
8. . ..l::tb8 9 h S bs 1 0 ifd2 transpose to the
main lines of Chapter 14. Instead 8...h s
9 ifd2 �b8 takes play t o the start of
Chapter 1S, while White can also play 9
tLic1 with an odd mix of set-ups that
was tried three times in the 1992
Fischer-Spassky match:
a) 9 ... es 10 dS l2Jd4 (Black was left
with a passive position after 10 ...tLle7?!
11 il.e2 lLih7 12 t2Jd3 fS 13 a4 lLif6 14
lLif2 as 1S ifc2 cs?! 16 0-0-0 b6 17
.l:!.dg1 in B.Spassky-R.Fischer, 12th
matchgame, Sveti Stefan 1992) 11 t2Jb3
tLixb3 12 ifxb3 �h7 13 il.e2 il.h6 14
il.xh6 �xh6 1S o-o-o �g7 was pretty
solid for Black in B.Spassky-R.Fischer,
28th matchgame, Belgrade 1992, al­
though with a clear plan of �dg1 and
g4, White's play is easier.
b) 9 ... tLld7!? 10 t2Jb3 as 11 a4 t2Jb4 12
il.e2 b6 13 g4 hxg4 14 fxg4 cs 1S hS
cxd4 16 t2Jxd4 tLics 17 tLlds il.b7 gave
Black a dynamic position in B.Spassky­
R.Fischer, 30th matchgame, Belgrade
1992.
8 es 9 ds
...
This time 9 t2Jb3 exd4 10 t2Jxd4
34 7
A ttacking C h e s s : T h e King 's I n dian, Vo l u m e 1
t2Jxd4 11 il.xd4 il.e6 12 il.e2 c6 should
favour Black, as he has saved a tempo
on the superfluous ... �b8.
9 ...l2Jd4 10 t2Jb3
Again, after 10 tLi1e2 tLixe2 11 il.xe2
tLih s Black has saved time by omitting
....l::tb8.
10...tLixb3
Black can also play 10 ...cs 11 dxc6
bxc6 12 l2Jxd4 exd4 13 il.xd4 .l::tb8 and
now 14 ii'd2 transposes to the main
line of the pawn sacrifice with 9 tLic1
(Line B22 in Chapter 17). The text move
gives the position its own character.
11 ii'xb3
After 11 axb3 Black can play 11...cS
to block the queenside or the immedi­
ate 11 ... tLih s.
11...cs
This was Kasparov's choice, but
11 ...tLih s, 11 ...t2Jd7 and 11 ...b6!? are
other possibilities.
12 dxc6 bxc6 13 o-o-o ii'e7
14 ii'b6
This is the only way to cause Black
any trouble.
14...il.b7 15 g4 �ab8
348
Black could also try 1 s...�fb8.
16 h4 �fc8 11 hs tLlds?!
This should not work. Instead
11 ... tLld7 18 ii'b4 tLics may leave White
with a small edge, but the position is
still quite unclear and with ideas like
...l2Je6-d4 1 do not think Black should be
too unhappy.
18 exds cxds 19 �xds?!
Instead White could refute Black's
play with 19 cxds! il.a8 (not 19 ... il.xds?
20 ii'xb8! ) 20 ii'as e4 21 hxg6 hxg6 22
f4! according to Kasparov.
19 ... il.xds 20 tLixds ii'e6 21 ii'a7
Perhaps Timman missed 21 ii'as �bS!.
21.. AaS 22 ii'b7 �ab8 23 ii'a7 .l::t a 8 24
ii'b7 �ab8
With a draw by repetition in
J .Timman-G. Kasparov, Moscow 1981.
C) 6 il.e3 l2Jc6 7 l2Jge2 a6 8 ii'd2 .i::t b 8
Now we come back to the main line
to tie up some loose ends. We examine:
('1:9 Jlcl2.
C2:'t'd5

Otherwise, 9 lLig3 looks funny, but
after 9 ... es 10 dS l2Jd4 11 tLige2 we have
just transposed back to Line Bl of
Chapter 17. Instead 9 a4 also looks
strange. Black could play 9 ... es (or even
9 ... as!? to fix the queenside) 10 dS and
now both 10 ... tLias 11 tLic1 cs and
10 ...l2Jb4, with the idea of 11 as cs, look
fine.
C1) 9 �d1
This is kind of an odd move. It is cer­
tainly not clear that the rook belongs
on the d-file, while White's develop­
ment still lags and he can no longer
castle queenside. This move does, how­
ever, serve to illustrate some of Black's
dynamic possibilities quite well!
9...�e8
I like this move, but 9 ...bs and
9 ...il.d7, with the idea of 10 tLic1 es, are
possible as well.
10 tLic1 es 11 dxes
After 11 dS l2Jd4 12 tLi1e2 tLixe2 13
il.xe2 tLih s Black has a normal position
and White has spent a move on the
strange �dl.
P a n n o Variatio n : Oth e r L in e s
11...tLixes 12 il.e2 bsl
Very aggressive. The simple 12 ...il.e6
is another option.
13 cxbs axbs
14 .il.xbs
If White cannot take the pawn at all,
then Black has achieved his strategic
aims very easily and has an improved
version of the variation 9 tLic1 es 10
t2Jb3 exd4 11 t2Jxd4 tLies (Line A2 of
Chapter 17).
The alternative is 14 tLixbs tLixf3+ !
1S gxf3 tLixe4 1 6 fxe4 '1lr'h4+ 17 il.f2
ifxe4 18 ifds (18 �g1 �xbs) 18 ... ifb4+
(better than 18 ... �xbs 19 ifxe4 .i::txe4 20
il.d4 il.xd4 21 �xd4 .i::txd4 22 il.xbs �b4
when Black is really playing only for a
draw) 19 �fl (19 ifd2 ifxbs) 19 ... il.b7
and Black is much better.
14...tLixe4! 15 fxe4 �xbs!
It is one blow after another.
16 tLixbs t2Jc4 17 iff2
17 il.gs ifd7! does not solve White's
problems.
11...�xe4 18 o-o tLixe3 19 ifxf7+ �h8
20 ifxc7 ifxc7 21 tLixc7 tLixdl 22 .i::tx dl
il.d4+
349
A tt a c k i ng C h e s s : Th e King 's I n d i a n, Vo l u m e 1
23 �fl?
This loses quickly. Instead 23 �h1
il.g4 24 �fl il.xb2 was a better try, al­
though Black's dominating bishops
give him excellent winning chances.
23 ...il.g4 24 .l::t d2
24 Ite1 �f4 is mate.
24 ...il.e3 25 �C2 �d4
White resigned here in F.Gheorghiu­
L.Kavalek, Amsterdam 1969.
C2) 9 dS
This advance looks logical enough,
but if it were that good the Panno
would never be seen. Play often heads
into a Benoni structure that gives Black
350
his fair share of the chances.
9...tLies
Also possible is 9 ...tLias!? when Black
will meet any of 10 lLig3, 10 tLic1 and 10
l2Jd4 with 10 ...cs, aiming for queenside
play.
10 l2Jd4
After 10 lLig3 cs 11 il.e2 e6 the play
heads for a Benoni position, but
10 ... c6! ? keeps the tension and gives
the game a more original flavour. The
same can be said for 10 tLic1 cs (again
there is 10 ... c6!?) 11 il.e2 e6 12 o-o (12
a4 exds 13 cxds is a decent enough
Modem Benoni for Black) 12 ... exds 13
exds (13 cxds allows 13 ... bs) 13 ... .i::te8
14 a4 h S ! ? with an unusual Benoni.
10...cs
With White's knight centralized,
10 ... c6 is untried, probably because of
the simple 11 il.e2, rather than 11 f4!?
tLieg4 12 dxc6 tLixe3 13 ifxe3 es! with
excellent play for Black ..
11 tLic2
After 11 dxc6 bxc6! (White has a
pleasant Maroczy Bind after 11 ... tLixc6
12 il.e2 il.d7 13 0-0) 12 f4? cs! Black is
doing well after either 13 t2Jb3 l2Jc6 14
il.e2 (or 14 h3 tLih s) 14 ... l2Jg4 or 13 fxes
cxd4 14 il.xd4 dxes when b2 will fall
after the impending exchange of
queens.
11...e6
Also possible is 11 ... il.d7 12 a4 e6 13
t2Ja3 (13 il.e2 exds 14 cxds bS gives
Black a good version of the Modem Be­
noni) 13 ... exds 14 cxds tLih s !? 1s il.e2
ifh4+ (or 1s .. .fs!?) 16 il.f2 iff4 when 17

il.e3 �4+ repeats.
12 a4 exds 13 cxds
Black has quite a decent Modem
Benoni because White has spent so
much time moving his king's knight.
13 ...tLihs
Black begins active operations im­
mediately, but 13.. Ji.d7 should also
suffice.
14 .il.e2 fs
Black has good counterplay.
D) 7 t2Jge2 .l::tb 8
8 ii'd2
Note that 8 dS does not win a pawn
after 8 ... tLies 9 il.xa7?! .l::ta8 because the
Pa n n o Variatio n : Oth e r L i n es
c4-pawn is loose. If Black wishes to play
this system against 7 ii'd2, however,
then he should employ the move order
7 ... .l::te 8 (7 ... .l::tb8? 8 dS does indeed win a
pawn because c4 is still protected) 8
l2Jge2 .l:!.b8.
8 ....i::te 8
This sly system was christened the
Westerinen Variation by John Watson
in recognition of its development by
the Finnish Grandmaster. Black can
now avoid the exchange of his dark­
squared bishop following il.h6, but
playing 8 ... �e8 does delay Black's
queenside counterplay by a tempo. The
really clever point of this line is when
White plays an early tLic1, as we shall
see.
9 h4
This direct move is probably best, as
it is in the main lines. Of course,
though, White has a broad choice:
a) 9 o-o-o is similarly direct and dan­
gerous: 9...a6 10 g4 (or 10 h4) 10...bs 11
h4 tLias (instead 11...hs 12 lLidS! ? has
scored terribly for Black) 12 lLig3 tLixc4
13 il.xc4 bxc4 14 hS and now, instead of
351
A ttacking C h e s s : Th e King 's I n d i a n, Vo l u m e 1
14...cs 1s dxcs (also very good for White
is 1s hxg6 hxg6 16 dxcs ii'as 17 il.h6
il.h8 18 il.f4, as in T.Marinelli-F.Portisch,
Budapest 1989) 1s ... ii'as 16 hxg6 fxg6
17 cxd6 il.e6 18 il.d4 when White had a
huge advantage in Nguyen Anh Dung­
F.Donguines, Manila 2001, Black should
consider 14...c6!?.
b) 9 g4 is relatively unexplored. Af-
ter 9 ... a6 10 h4 bS (or perhaps 10 ... hs)
11 h S Black could try 11 ... bxc4, 11 ...tLias
or 11 ... es.
c) 9 .i::t d1 a6 is more likely to arise
from 8 ... a6 9 �d1 �e8 and this position
was considered in Line Cl, above.
d) 9 dS will lead to similar play to
that of Line C2. Black should not take
this plan too lightly, though.
For example, 9 ...tLies 10 l2Jd4 cs 11
tLic2 e6 12 il.e2 exds 13 cxds il.d7 (ob­
vious, but perhaps too slow) 14 a4 a6
1S t2Ja3 and White's long knight tour
proved rather purposeful by clamping
down on the bS-square in S.Brudno­
D.Vigorito, Providence 2008. Very inter­
esting, however, would have been
13 ...bs!? 14 il.xbs ii'as! 1s il.xe8 tLixe8
352
1 6 t2Jd1 ii'a6 with tremendous activity
for the exchange.
e) 9 a3 (of course White can pursue
queenside plans as well) 9...t2Jd7!?
(more interesting than 9...a 6 10 b4
which could transpose to Line A of
Chapter 19) 10 b4 tLlb6 11 tLic1?! (this is
pleasant for Black, but both 11 ii'a2 e6!
and 11 ii'd3 e6! give Black counterplay
as well) 11...eS 12 dS l2Jd4 (White's lack
of control over the b3-square leaves
him struggling for equality) 13 il.d3 c6
14 o-o cxds 1s cxds il.d7 16 il.xd4 exd4
17 tLibs il.xbs 18 il.xbs �e7 19 il.d3
.i::tc7 20 tLie2 �bc8 gave Black some ini­
tiative in D.Vigorito-J.Watson, Phila­
delphia 1993. I may have lost this
game, but at least I made a good
friend!
f) 9 .i::t b 1 is probably a better way of
implanting the queenside plan:
9 ... a6 (9 ...tLld7 is possible, but now
after 10 b4 lLib6 11 t2Jc1 es 12 dS l2Jd4
13 t2Jb3 Black would rather have his
king's rook on f8 to support the .. .fs­
advance) 10 b4 bS (it is probably better
to temporize with 10 ... il.d7) 11 cxbs

axbs 12 ds tLies 13 l2Jd4 il.d7 14 tLicxbs
(the patient 14 il.e2 ! ? is also possible,
saving the consumption of the bS­
pawn until White completes his devel­
opment) 14 ... e6 1S dxe6 fxe6 16 il.e2
t2Jxf3+ 17 gxf3 es was L.Polug aevsky­
E.Gufeld, Riga 197S. Now 18 il.c4+ �h8
19 o-o exd4 20 t2Jxd4 would leave Black
with very little for the pawn.
g) 9 tLic1 is quite common, but Bl ack
will now reveal a key point: 9 ... es 10
dS?! (10 t2Jb3 exd4 11 t2Jxd4 is safer,
when Black could head for the lines of
Chapter 17 with 11 ...t2Jxd4 12 il.xd4 a6
or try 11...ds!? 12 cxds tLixds 13 tLixds
t2Jxd4; then 14 o-o-o cs 1s il.gs ii'd6 is
unclear and 14 il.xd4 ii'xds 1S il.xg 7
ii'xd2+ 16 �xd2 �xg7 is level) 10...t2Jd4
11 t2Jb3 cs! 12 dxc6 bxc6 13 t2Jxd4 exd4
14 .il.xd4.
Compared with the pawn sacrifice
from Chapter 17 (after 8...a6 9 tLic1),
Black is much better off having played
... �e8 instead of ... a6. Following 14 ... ds
(even 14... cs!? is possible, with the
point 1 s il.e3 dS! 16 cxds tLixds and
now 17 tLixds .l::txb2 18 ii'd3 il.e6 17
Pa n n o Va ri a t i o n : O t h e r L i n es
ii'xds il.xc3+ 18 bxc3 ii'xds 19 exds
�xe3+ gives Black plenty for the pawn,
but 17 �d1 �xb2 18 ii'xb2 il.xc3+ 19
ii'xc3 tLixc3 20 �xd8 .l::txd8 21 il.xcs or
21 �g1 should prove more testing) 1S
cxds cxds White has:
g1) 16 il.bs runs into 16 ...�xbs! 17
tLixbs tLixe4! 18 fxe4 .i::txe4+ 19 �2 (al­
ternatively, 19 �fl il.a6 or 19 �d1
il.g4+ 20 �cl when after both
20...il.xd4 21 t2Jxd4 ii'c7+ 22 ii'c3 ii'f4+
and 20 ....i::te 2 21 ii'f4 ii'c8+ 22 t2Jc3 �e4
Black wins) 19 ... ii'M+ 20 g3 .i::txd4 21
t2Jxd4 il.xd4+ and White can resign.
g2) 16 es tLih s! (16 ...t2Jd7 has also
been played, but the text is better) and
now:
g21) 17 f4 is well met by 17 ...f6,
which leaves:
g22) 17 o-o-o is the safest and then
11 ... il.xes 18 il.xes �xes is about equal.
g23) 17 il.bs (ambitious, but Black
has a wonderful idea) 17 ...�xes+! 18
�f2 (18 il.xes il.xes gives Black a very
strong initiative for the exchange)
18 ... a6! 19 il.e2 �xb2 (or 19 ...ii'h4+ 20
g3 �xb2 !) 20 ii'xb2 ii'h4+ 21 �fl ii'xd4
353
A t tacking Ch e s s : The King 's I ndian, Vo l u m e 1
and Black was winning in A.Saidy­
H.Westerinen, Netanya 1969.
Now we return to 9 h4:
9 a6
...
Black has also tried 9 ... hs, but 10
o-o-o (not 10 tLic1?! es with similar play
to variation 'g' above) 10 ... a6 11 il.h6
(White has also fiddled with all of 11
ii'e1, 11 lLif4 and 11 tLlds) 11 ...il.h8
(11...bs 12 g4!) 12 g4 (or 12 ii'e3 ! ? es 13
dS tLias 14 lLig3 cs when both Watson's
suggestion of 1S il.d3 bS 16 tLif1 and 1S
il.e2 il.d7 16 tLif1 bS 17 .i::td2 tLixc4 18
il.xc4 bxc4 19 ii'e2 l2Jh 7 20 g4, as in
J.Tarjan-A.Whiteley, Hastings 1976/77,
look promising for White) 12 ... es (Black
could try 12 ...b s or even 12 ... hxg4!?, but
the whole position looks precarious for
him) 13 iLg S! looks promising for
White.
10 hs bs
Black could consider too 10 ... es 11
hxg6 fxg6 12 dS tLias 13 lLig3 cs or even
10 ... lLixh S !? 11 il.h6 (11 g4 lLif6 12 il.h6
il.h8 13 il.g s also looks dangerous;
Black could try 13 ... bs or 13 ... e6) 11 ... es
12 il.xg7 �xg7 13 g4 and here:
354
a ) 1 3 ... t2Jxd4 1 4 t2Jxd4 exd4 with a
further split:
al) 1S tLie2 ii'f6 (1S ... tLif6!? 16 ii'xd4
was feared by Watson, but 16 ...il.xg4!
17 fxg4 �xe4 gives Black excellent play)
16 gxh s ii'xf3 17 h6+ �8 18 �h4 .i::txe4
19 .i::txe4 ii'xe4 20 ii'xd4 ii'xd4 21 t2Jxd4
il.g4 22 �2 and White had the better
chances in W. Browne-J.Watson, Berke­
ley 1977.
a2) 1S lLidS! looks strong, since
1s ... c6 16 gxh s cxds 17 hxg6 hxg6 18
ii'xd4+ f6 19 ii'xds il.e6 20 ii'd2 must
favour White.
b) 13 ... exd4 is probably best: 14
gxh s (14 tLlds tLies 1s il.g2 lLif6 16
ii'h6+ �g8 gives Black counterplay, as
there is no mate) 14... dxc3 1S ii'xc3+
�g8 16 o-o-o gives White some com­
pensation, but this still looks playable
for Black; ... tLie s is available, while after
hxg6 fxg6 Black can defend along his
second rank. However, if White delays
the exchange on g6, then ... g s may
even be possible to keep the h-file
closed.
11 hxg6 fxg6
 Pan n o Variation: O th e r L i n e s

Certainly 11...hxg6?! looks too risky. 14. ..axbs 1 5 tLlds

12 il.h6 Now 1s ...l:!.f8? 16 if c2 (16 �cl also

Untried is 12 cxbs axbs 13 dS, but looks good) 16...il.d7 17 'ir'b3 il.e6 18
this may well be good for White as the tLixf6+ 'ili'xf6 19 il.xf6 il.xb3 20 il.xh8
bS-pawn is weak. l2Jb4 21 �d2 tLixa2 22 dxes �xh8 23
12 ... il.hS 13 il.gs es l2Jd4 il.g8 24 e6 cs 2 5 l2Jc6 �b6 26 tLle7
Instead 13 ... �f8 has been the most was winning for White in D.Studzinski­
popular choice, which seems strange. S.Szczepaniec, correspondence 1991.
Both 14 o-o-o and 14 cxbs axbs 15 lLif4 However, Black can improve with
look promising for White. However, De 1S ... t2Jxd4! when the position is un­
Firmian and Watson's 1980 suggestion clear: for example, 16 t2Jxd4 exd4 17
of 13 ... e6!? is still untried! ifxd4 tLixe4! with complications that
14 cxbs are not unfavourable for Black.
Browne and Watson considered 14
tLlds t2Jxd4 15 t2Jxd4 exd4 16 il.xf6 (in­ E) 7 tLige2 a6 8 ifd2 �es
stead 16 ifxd4 tLixdS! 17 ifxds+ il.e6 18
il.xd8 il.xds 19 il.xe7 .l::tb7 20 il.xd6
il.xc4 is fine for Black) 16 ...il.xf6 17 'i!t'h6
il.g7 18 ifxh7+ 'itf7 19 �h4 to be win­
ning for White (we must keep in mind
that this was in 1980), but 19 .. Ji.fS!
(worse is 19 ....l::th 8?! 20 .l::tf4+ il.fs 21
.i::txfs+! gxfs 22 'ili'xfs+ �g8 23 0-0-0,
which gives White good compensation
for the exchange) 20 �f4 ifgs 21 '1!r'h4
(and not 21 o-o-o? .l::th 8) 21 ... il.f6! 22
'1lr'h7 + il.g7 leads to a draw by repetition. Here Black omits ....l::tb8 rather than

355
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
... a6. This is kind of a half-measure be­
tween the main lines and the
Westerinen Variation.
9 h4
This is always critical, but White has
the usual array of ideas. The queenside
ideas 9 a3 and 9 .l::tb1 will likely trans­
pose to the notes to White's 9th in Line
D, if not to Chapter 19, while 9 �cl
il.d7 10 b3!? (instead 10 tLid1 could be
met by the immediate 10 ... bs when
Black has benefited from omitting
... .l::tb8) 10 ... e6!? was J .Piket-M.Golubev,
German League 2002.
Other attacking moves are less
good: 9 o-o-o allows 9 ... bS! immedi­
ately; the unexplored 9 g4 �b8 was
mentioned in note 'b' to White's 9th
move in Line D; and the direct 9 il.h6
il.h8 10 h4 e s ! allows Black to use the
e8-rook with 11 dS l2Jd4 12 tLixd4 exd4
13 tLie2 (13 ii'xd4? tLixe4) 13 ... cs 14
dxc6 bxc6 1S t2Jxd4 tLixe4! 16 fxe4
.i::txe4+ 17 tLie2 .l::tb 8, which gives him a
strong initiative for the piece.
White can also play 9 tLic1 es
10 dS (after 10 tLlb3 exd4 11 tLixd4
356
Black has the usual choice between
11 ...tLixd4 and 11 ... tLies) 10...t2Jd4 11
tLlb3 (11 tLi1e2 cs 12 dxc6 gives Black
the additional option of 12 ... tLixc6 !? 13
lLidS bS 14 il.b6 ii'd7 1S lLie7 .l::tb8 16
tLixe8 ii'xe8 with some compensation
for the exchange in the well-known
game A.Beliavsky-G.Kasparov, Moscow
1981) 11 ... cs (instead 11 ... tLixb3 12
axb3 may be okay too, but Black has
played ... .l::te 8 instead of ... �b8, which
hardly favours him) 12 dxc6 bxc6 13
tLixd4 exd4 14 il.xd4 dS!? (14 ... .l::tb8
transposes to the main-line pawn sac­
rifice - Line B22 of Chapter 17) 1S cxds
cxds 16 es tLld7, which has been seen a
few times. This version of the pawn
sacrifice should help Black because
... .l::te8 is often more useful than ....l::tb8.
9 ... hs 10 o-o-o
10 tLic1 again allows the thematic
10 ... es 11 dS l2Jd4 12 t2Jb3 (after 12
tLi1e2 cs 13 dxc6 Black has the extra
possibility of 13 ... tLixc6!?) 12 ... cs!. Once
more, compared to lines with ... �b8,
Black probably has a slightly improved
version of the pawn sacrifice with his
rook on e8 instead: 13 dxc6 bxc6 14
tLixd4 exd4 1s il.xd4 dS!? (1s ....l::tb8
would transpose to the main lines of
the pawn sacrifice covered in Line B22
of Chapter 17) 16 cxds cxds 17 es tLld7
18 f4 f6 19 e6! lLif8!? (instead
19 ... �xe6+ 20 il.e2 has been played a
couple of times with little success) 20
fS gxfs (Black has little choice; he must
try to use piece play to compensate for
his deranged structure) 21 il.e2 tLixe6
 Pa n n o Varia t i o n : O t h e r L in es

22 o-o? (White probably felt that this

was the safest, but 22 o-o-o! ? looks bet­
ter; his king is pretty safe on the
queenside and castling long brings
pressure against d5 more quickly)
22 ...t2Jxd4 23 ifxd4 il.h6! and Black had
serious counterplay in S.Kasparov­
D.Bocharov, Bhubaneswar 2010.
10...bs

14...gs!?
Black keeps the h-file closed. Instead
14 ... e5 15 hxg6 fxg6 16 d5 l2Jd4 was the
risky continuation of A.Komev-
1.Kumosov, Tomsk 2001. Here Gallagher
gives 17 t2Jxd4 exd4 18 il.xd4 il.xd4 19
ifxd4 ifg5+ 20 �d2 if e5 21 ifxc4 gxf3
22 ifxe7 with the initiative for White
and Golubev continues 22 .. lte7 23
Black has achieved this advance ifd8+ �e8 24 ifh4! .l::ta7 25 .i::td h2. This
without the preparatory move ... �b8, looks scary, although I do not see any­
but this is not such a great accom­ thing clear for White after 2 5 ....l::t g 7.
plishment because as long as White 15 il.xgs es 16 il.xf6 ifxf6
does not capture on b5, Black would
benefit from having his rook on the b­
file after all.
11 tLlds!
This is considered critical. White
cannot easily exchange off Black's g7-
bishop, so he trades off Black's defen­
sive knight instead.
11..bxc4 12 tLixf6+
Instead 12 g4 hxg4 13 tLixf6+ il.xf6
14 h 5 transposes, whereas 12 ...tLixg4?!
13 fxg4 il.xg4 14 il.h6 did not hold up 17 fxg4
in J.Lautier-C.Maier, Versailles 2006. Instead 17 d5 gxf3 18 dxc6 fxe2 19
12 ...il.xf6 13 g4 hxg4 14 hs il.xe2 iff4 20 il.xc4 il.e6 21 .i::t dg1+ �h8

357
A ttacking Chess: Th e King 's I n d i a n , Vo l u m e 1
was given as equal by Lautier, but 21
.i::th g l+!? with the idea of �dfl looks
promising for White. Black may im­
prove earlier with 17...ifxf3 !?, intend­
ing 18 ...l2Jd4 with counterplay as the
e4-pawn is weak.
17...exd4 18 gs iff3 19 g6 il.g4!
A suggestion by Golubev as an im­
provement over 19.. .fxg6?! 20 t2Jxd4
t2Jxd4 21 ifxd4 with a winning position
in J.Lautier-J.Piket, Cannes 1990. Later
he got to try it against Lautier himself!
20 .l::!.gl
This could well be a mistake.
Golubev mentions a couple of alterna­
tives:
a) 20 t2Jxd4 ifxdl+ 21 ifxdl il.xdl
22 il.xc4 �g7!? 23 .l:!.g1 il.g4! (not
23 ... .i::tx e4?! 24 h6+! �6 25 g7! when
White is magically better) 24 tLixc6 (24
.i::t x g4 tLies! was Black's point) 24 ... .i::txe4
25 il.ds �ae8 26 il.xe4 �xe4 and Black
is certainly okay.
b) 20 gxf7+ is probably best:
20 ... �xf7 21 .i::tg1 tLies 22 ifxd4 iff6
and now one possibility Golubev gives
is 23 �bl! ? .i::tab8 24 t2Jc3 !? il.xdl 25
358
il.xc4+ tLixc4 26 ifxc4+ ife6 2 7 ifxc7+!
if e7 28 ifc4+ ife6 when Black is some­
how holding on.
20...�xe4?
Tragic. Instead 20 ...tLies 21 gxf7+
�xf7 transposes to variation 'b' above,
but Black's best looks to be 20 .. .fxg6!
and here:
a) 21 �el (to take on d4) 21 ...c3 !? 22
bxc3 il.xhs 23 t2Jxd4 t2Jxd4 24 ifxd4
iff4+ 25 �b2 ifes gives Black good
chances.
b) 21 hxg6 and now Golubev gives
21 ...�g7 (not 21 ... c3 22 bxc3 .rf.ab8 23
t2Jxd4!) 22 �el d3 23 il.g2 (23 .l::t g 3
ifxe4 24 il.g2 iffs is good for Black)
23 ...iff2! 24 .rf.gfl! ifxe2 25 �xe2 dxe2
with compensation for the queen.
21 gxf7+ �xf7 22 il.g2
And White won significant material
in J.Lautier-M.Golubev, Odessa (rapid)
2006.
F) 6 iLgS
This move can be annoying to those
who like to play 6 il.e3 es, but it en­
courag es Black to play 6 ... cs without

offering a pawn. While completely sat-
isfactory, 6 ... cs does not exactly fit into
our repertoire and there is nothing
wrong with continuing with our fa-
vourite system.
6 l2Jc6
...
Just like after 6 tLige2, Black could
also consider 6 ... a6 and only then 7
ii'd2 l2Jc6. Against 6 il.e3 this would
give White the extra option of 7 il.d3!?,
but here the d4-pawn would be
weaker, while after 6 tLige2 a6 the
move il.d3 is simply not possible.
7 tLige2
There is nothing to be gained by de­
laying this move: for example, 7 ii'd2
a6 8 tLige2 tran sposes, while 7 dS tLies
8 ii'd2 (8 f4 l2Jed7 9 lLif3 tLics with ideas
like ... c6, ... ii'b6/aS and ... l2Jg4 is good
for Black) 8 ... c6 is comfortable for Black,
as it is not easy for White to develop his
kingside naturally without the loosen­
ing f4-advance.
1 a6 8 ii'd2 �b8
...
So now we have reached our main
set-up, except that White's bishop is on
gs instead of e3. If White plays a quick
Pa n n o Va ria ti o n : O t h e r L i n e s
il.h6, we will generally transpose to 6
il.e3 lines with a quick il.h6. Indeed, the
immediate 9 il.h6 would take play into
Line C of Chapter 18. Independent play
will be examined here, chiefly after the
options:
F1: 9 l:tc1
F2: 9 d5
Instead 9 tLic1 does not mix well
with il.gs because d4 lacks protection:
for example, 9 ... es 10 dS l2Jd4 11 t2Jb3
cs 12 dxc6 bxc6 13 t2Jxd4 exd4 14 ii'xd4
.rtxb2 with a great position for Black.
White can also play 9 h4 hS (riskier
is 9 ... bs 10 hS) when both 10 il.h6 and
10 o-o-o bS 11 il.h6 transpose to the
main lines with 6 il.e3 (Chapter 1S).
Instead White has tried 10 tLlds, but
after 10 ... bs (also possible is 10 ...l2Jh 7!?
11 il.h6 il.xh6 12 ii'xh6 e6 13 t2Je3 bS
with unclear play in J .Piket-F.Nijboer,
Groningen 1992) 11 o-o-o bxc4 12 g4
tLixds 13 exds l2Jb4 14 t2Jc3 cs! Black
enjoyed good play in E.Vegh-L.Vogt,
Eger 1984.
359
A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1
Fl) 9 �cl
We have seen this move before, but
here White actually intends to head
into a Benoni structure rather than
expand on the queenside.
9 Ji.d1 lo ds
...
Instead 10 llld1 follows the recipe
from Line C of Chapter 19, but it is not
clear that White's bishop should be on
gs in this kind of position. Black should
continue 10 ... bs (or even 10 ... e6!?) 11 cs
e6.
The subtle 10 b3 is more popular.
Then after 10 ... bs 11 ds lll e s
(11 ...llla7! ? 12 lllg 3 looks strange and
11 ...llla s?! 12 cxbs axbs 13 llld4 looks
good for White, as 13 ...b4 14 lllcbS! is
strong) we have:
a) 12 cxbs axbs 13 llld4 'ii'e 8 (not
13 ...b4?! 14 lllcbS!) 14 f4 b4! 1s fxes
bxc3 16 'ii'e3 (16 'ii'xc3 lllxe4) 16 ... dxes
and Black has good counterplay accord­
ing to Golubev.
b) 12 f4 llleg4 13 lllg 3 (13 h 3 b4!)
13 ...h6 14 Ji.h4 and now, instead of
14 ... lll h7, as in A.Dreev-M.Golubev,
Alushta 1994, when lS Ji.e2 would
3 60
have been good for White, Black should
play 14... b4! lS llldl (or lS lllce2 es! 16
dxe6 fxe6 17 h3 gs 18 fxg s hxgs 19
Ji.xg s lllf2! - Golubev) 1s ... es! 16 dxe6
fxe6 17 h3 gs 18 fxgs lll h7 19 lllh s
hxg s 20 hxg4 gxh4 21 lll x g7 'it>xg7
which looks okay for Black according to
Golubev.
10 lll a s!?
...
Of course 10...lll e s is also possible,
but after 11 b3 Black will have some
concern over the stability of his es­
knight.
11 lll g3
The e2-knight will have to move at
some point anyway, so there is no rea­
son to go 11 b3?! cs 11 llld4 cs 12 lllc2
bS and Black's play is well underway.
11 cs
...
A good alternative is 11 ...bs!? 12
cxbs axbs 13 lllx bs (White could also
try 13 b4 lllc4 14 Ji.xc4 bxc4, but this
looks like a decent version of this kind
of structure for Black - White's pieces
are oddly placed) 13 ... Ji.xbs 14 'ii'xas
Ji.xfl lS 'it>xfl �xb2 16 'ii'xe7 'ii'a8! with
excellent play.
 Pa n n o Va riati o n : O t h e r L i nes

12 Ji.d3 Another possibility is 9 ...lllas 10

No better is 12 Ji.e2 bs 13 cxbs (or lll c 1 cs. After the text, we have a split:
13 b3 bxc4 14 bxc4 �b4 15 llld1 Ji.a4!)
13 ... axbs with good play for Black. F2a.: 10 .!tla3
12 ...bs 13 b3 bxc4 F22: 1042Jd4
Hazai suggests 13 ... es here, but I

prefer to focus on the queenside. F21) 10 lllg3 c6

14 bxc4 �b4 1s llld1 The alternative 10... cs is fine too,
but the text is the more challenging
idea.

This was S.Mamedyarov-1.Chepa­

rinov, Antalya 2004. Now 1s .. .Ji.a4!? 16
o-o lll d 7 with ideas like ...lll b 6 and 11 Ji.e2
...llle s, attacking the c4-pawn, would Black has good play after 11 f4
give Black good counterplay. llle d7 11 a4 'ii'a s.
11. ..bs 12 cxbs cxds!
F2) 9 dS

A nice pawn sacrifice. It is not easy

9 llle s
... for White to keep control.

361
A ttacking Ch e s s : Th e King 's Indian, Vo l u m e 1

13 exds
White has also tried 13 Ji.xf6 Ji.xf6
14 lllxd5 Ji.g7 15 a4 axb5 16 axb5 (16
Ji.xb5 Ji.d7! gives Black good compen­
sation) and here:

15 Ji.xf6
This is a concession, but 15 lllc 3
Ji.a8 with the idea of 16 o-o 'ii'b6+
(Smirin) gives Black counterplay.
1s ...exf6!
a) 16 ...Ji.e6 17 o-o Ji.xd5 was Worse is 15 ...Ji.xf6 16 llle4 Ji.g 7 17
L.Polugaevsky-J.Nunn, European Team o-o and White has organized his forces.
Championship, Plovdiv 1983. Now 18 16 f4 lll d 7
exd5 'ii'b6+ 19 'it>h1 �fc8 20 �acl with Black intends ...f5 with good play.
the idea of f4 and !k6 would give 17 fs 'ii'b6 18 llle4 Ji.a6! 19 lllbxd6 llle s
White the upper hand according to With White's king stuck in the cen­
Polugaevsky. tre, Black had good compensation in
b) 16 ... Ji.d7!? 17 f4 (after 17 'ii'a 5 e6 l.Novikov-1.Smirin, Las Vegas 1999.
18 'ii'xd8 !Ifxd8 19 lll C 3 both 19 ... d5
and 19 ...�dc8 give Black good compen­ F22) 10 llld4
sation for the pawn) 17 ...lllg4 18 'ii'a 5
'ii'e8 19 J:.d1 (not 19 o-o? Ji.d4+) 19 ... h5
(19 ...Ji.e6!?) 20 o-o Ji.xb5 21 Ji.xb5 !Ixb5
22 'ii'a7 !Ixb2 23 h3 lllf6 24 lllxe7+ 'it>h7
25 !Ixd6 'ii' a8! gave Black good coun­
terplay in M.Sadler-L.Vogt, Altensteig
1992.
c) 16 ... e6!? 17 lllc 3 'ii'c 7 18 o-o l:.d8
also looks fine for Black. The two bish­
ops promise him good long-term com­
pensation for the pawn.
13 .. axbs 14 lll x bs Ji.b7
. 10 ...cs

362
 Pa n n o Va ria t i o n : O t h e r L in es

Black kicks the knight with gain of 11 ... ct:Jhs!?, with the idea of .. .fs, are
time. Another idea is 10 ... �e8!? 11 a4 legitimate alternatives.
(or 11 Ji.e2 cs 12 ct:Jc2 bs!) 11 ... e6. 12 a4
11 ct:Jc2 To prevent ...bs.
12...e6

11 ...'ii'e B!?
Black must play very deliberately.
Both 11 ... e6?! 12 f4 and 11 ...'ii'a s 12 a4
miss the mark, while after 11 ... bs 12
cxbs axbs 13 Ji.xbs (13 ct:Jxbs ct:Je8 ! ?)
13 ... !Ixbs 14 ct:Jxbs ct:Jc4 1s 'ii'e 2 �as+
16 'it>f2 ct:Jxb2 (the main point is that
16 ... 'ii'xbs fails to 17 b3) 17 a4 Black did
not have enough in A.Dreev-1.Smirin,
Biel lnterzonal 1993.
I like the clever text move, but
11 ... ct:Je8!?, aiming to support ... bs, and
13 Ji.e2
Instead 13 ct:Ja3 clamps down on the
bs-square, but White's development is
lacking and 13 ...exds 14 cxds Ji.d7 1s
as ct:Jhs with the idea of .. .fs gives Black
counterplay.
13 ... exds 14 cxds
Now 14 exds? runs into 14 ... ct:Jxc4.
14... bs
Black has achieved his goal and has
good play.

363
 Index of Variations

1 d4 lllf6 2 c4 g6 3 lllc 3 Ji.g7 4 e4 d 6 and now:

Part I: s lllf3 0-0 6 Ji.e2 es 7 0-0

Part II: s lllf3 0-0 6 Ji.e2 es without 7 0-0
Part Ill: S f3

Part I: 1 d4 lt:Jf6 2 C4 g6 3 lt:Jc3 Ji.g7 4 e4 d6 S lt:Jf3 0-0 6 Ji.e2 eS 7 0-0 lt:Jc6 8 dS

8 dxes - 1s6
8 Ji.e3
8...!Ie8 - 1S8
8 ...lllg 4 - 161
8 llle 7 9 lllel
...

9 llld2 as 10 a3 Ji.d7 11 b3
11 ...lll c8 - 80
11...lll e8 - 83
11 ...c6
12 !Ibl - 86
12 !Ia2 - 88
12 Ji.b2 - 89
9 b4 lllh s
1o lll d2 - 93
10 'ii'c2 - 9S
10 cs - 97
10 g3 - 101
10 �e1 fs 11 lllg s lllf6
12 Ji.f3 c6
13 'ii'b3 - 107
13 bs - 110
13 Ji.b2 - 113

364
 In dex of Varia t i o n s

13 Ji.e3 - 118
12 f3
12 ... c6
13 'it>hl - 126
13 Ji.e3 - 128
12 ...'it>h8
13 Ji.e3 - 132
13 lll e 6 - 134
13 cs - 137
13 �bl - 139
9 Ji.gs - 14S
9 Ji.d2 - 149
9 a4 - lSl
9 'it>hl - 1S3
9 lll d 7 10 Ji.e3
•.•

lo lll d 3 ts 11 Ji.d2 lllf6 12 f3 f4

13 cs lllg 6
14 !kl - 29
14 cxd6 cxd6
lS �Cl - 20; lS lt:Jf2 - 27
13 g4 - 34
lo f3 fs 11 g4 'it>h8
12 Ji.e3 - 39
12 lll d3 - 40
12 lll g2 - 42
12 h4 - 44
10... fs 11 f3 f4 12 Ji.f2 gs 13 a4
13 b4 - 49
13 llld3 - Sl
13 lll bs - ss
13 �Cl - S8
13 ... as
13 ...lll g 6 - 64
14 lt:Jd3 b6 15 b4
lS Ji.el - 67
1s...axb4
16 lll bs - 69
16 lllxb4 - 73

365
A ttacking C h e s s : The King 's I n d i a n , Vol u m e 1

Part II: 1 d4 lt:Jf6 2 C4 g6 3 lt:Jc3 Ji.g7 4 e4 d6 S lt:Jf3 0-0 6 Ji.e2 eS 7 Ji.e3

7 ds as
8 llld2 - 204
8 Ji.e3 - 207
8 0-0 - 210
8 h3 - 214
8 Ji.g s h6 9 Ji.h4 llla6 10 llld2 'ii'e8
11 a3 - 223
11 o-o lllh 7 12 a3 Ji.d7 13 b3 hs 14 f3 Ji.h6
lS Ji.f2 - 229
lS 'it>hl - 230
lS l:.bl - 234
7 dxes dxes 8 'ii'xd8 l:.xd8 9 Ji.gs
9 ... !Ie8 - 241
9 ...lllbd7 - 247
9 ... c6 10 lll xes !Ie8 11 o-o-o lll a6
12 lllf3 - 2S3
12 f4 - 2 S S
1 2 !Id6 - 2S9
1 ...lllg4 s Ji.gs f6 9 Ji.h4
9 Ji.cl lll c 6 10 ds
10 ... llle 7 - 171
10 ... llld4 - 171
9...gs
9 ...lll c 6 10 ds llle 7 11 llld2
11 ... lllh 6 - 176
11 ... hs - 180
10 Ji.g3 lll h 6 11 ds
11 cs - 184
11 h3 - 188
11 dxes - 190
11 ...fs - 194
11 ...lll d7 - 197

Part I l l : 1 d4 lllf6 2 c4 g6 3 lll c 3 Ji.g7 4 e4 d6 S f3 0-0 6 Ji.e3

6 lllg e2 - 3S9
6 Ji.gs lllc 6 7 lllg e2 a6 8 'ii'd2 !Ib8
9 !Icl - 360

366
 I n dex of Varia t i o n s

9 ds - 361
6 lllc6 7 lllge2
.••

7 'ii'd2 - 34S
1 ... a6
7 ...!Ib8 - 3S1
8 'ii'd 2
8 lll c l - 347
8 h4 - 347
8...!I b8
8 ... l:.e8 - 3SS
9 h4
9 lllc1 es
10 lll b 3 exd4 11 lllxd4
11 ...lllxd4 - 303
11 ...lll e s - 306
10 ds llld4
11 lll 1 e2 - 308
11 lllb3
11 ... lllxb3 - 313
11...c s 1 2 dxc6 bxc6 1 3 lllxd4 exd4 1 4 Ji.xd4 !Ie8
1s Ji.e2 ds 16 cxds
16 ... cxds - 317
16 ...l:.b4 - 318
9 0-0-0 - 320
9 g4 - 321
9 Ji.h6
9 ... Ji.xh6 - 326
9 ...bs - 328
9 a3 - 331
9 !Ibl - 33S
9 �c1 - 339
9 �d1 - 349
9 ds - 3so
9... hs
9 ... es - 266
9 ...bs - 268
10 0-0-0
10 lllc1 es 11 ds llld4 12 lllb3

367
A ttacking C h e s s : Th e King 's I n d i a n , Vo l u m e 1

12 ... lllxb3 - 294

12 ... cs 13 dxc6 bxc6 14 lllxd4 exd4 15 Ji.xd4 !Ie8 16 Ji.e2 ds
11 cxds - 297
17 es - 300
10 bs 11 Ji.h6
...

11 llld s - 216
11 lllf4 - 280
11 es
...

11...Ji.xh6 - 284
11 ...bxc4 - 285
12 Ji.xg7 - 288
12 lll d s - 281

368
 2free updates
of this book available
within a year of publication at
www.everymanchess.com