High School Secrets

Viet Quoc Hoang

The University of Auckland
Mathematics graduate
August 13, 2013
Abstract
Many high school students understand the textbook concepts and nd them a bit boring to
just do the usual practice exercises. However, if they choose to take on a Scholarship in Calculus
or Statistics, often a case, the questions in those exams may simply change their opinions about
a particular topic or subject. Why is it a case? This is basically a problem in writing textbook in
which there is a little or no advanced mathematical techniques required. In order to unveil this
mystery, I have opened this article for curious minded teachers, students, parents or educators
to have a chance to look back at what we could have done further, out side the textbook context
and dig into more of advance Mathematics. This is dedicated to my parents with their greatest
support that I could not ask for more and my brother for his endless interest in Mathematics
and his new laptop sent over to me, without it, the script could not have been written.
1 Translation and Rotation matter
Occasionally, we deal with several tough problems such as rotating a graph about a point that is NOT
the origin and so forth. In order to accomplish this mission, we need to develop the tools beyond
high school knowlegde (what I mean here is the concept of Matrices). And yes, having coped with
matrix is not a pleasant experience at all, in fact, we need to understand basic matrix operations such
as addition, subtraction, multiplication, inverse of a matrix, eigenvalues, eigenvectors, determinants
and so on. I will let your textbook handle this critical part. Lets talk a bit more about the change
in the origin.
Theorem 1. As usual, our origin should be at (0, 0) by mathematical convention. Now, here comes
an interesting story, suppose I do not want my origin to be at (0, 0) but rather at some chosen points,
how would I do this? The answer is simple. We perform the following transfromation, in fact, to be
precise, coordinate-wise transformation. Let (a, b) be our new origin, we consider the transformation
under the previous coordinate system as follows
(x, y) (x +a, y +b)
So, our rst tool is seemingly understandable and applicable. However, the second theorem I will
be stating below is not that simple to comprehend at the rst sight and the proof for this is omitted
due to a basic reason: it is everywhere in standard Calculus textbooks for rst-year university courses.
1
Theorem 2. Let (x, y) be the point in the Cartesian coordinate system and we want to rotate that
point about the origin (0, 0) with the angle , counter-clockwise. The transformation is given in
the following formula
_
x

_
=
_
cos sin
sin cos
__
x
y
_
where (x

, y

) denotes the transformation point of (x, y).

Similarly, it follows from the above theorem with a little spin of trigonometry cos = cos() and
sin() = sin .
Theorem 3. Let (x, y) be the point in the Cartesian coordinate system and we want to rotate that
point about the origin (0, 0) with the angle , clockwise. The transformation is given in the following
formula
_
x

_
=
_
cos sin
sin cos
__
x
y
_
Yes, we are right there to get ready for the rst problem.
Problem 1 (Modied NCEA Calculus Exam 2011, Question 2(ii)).
Consider the function g(x) =
_
1

x. The curve y = g(x) rotated 180 degrees around the point

_
1
2
,
1
2
_
. The two curves intersect only at the points (0, 1) and (1, 0). Find the area of the region
enclosed by the two curves.
Needless to say, the important part of this question is to determine the rotated equation of g(x).
To do this, we would like to implement the above theories correctly. Lets go over this for once.
Firstly, we have got to move the origin to the point
_
1
2
,
1
2
_
. Thus,
(x, y)
_
x +
1
2
, y +
1
2
_
Therefore, the original function becomes
y +
1
2
=

1
_
x +
1
2
Furthermore, we need to rotate g(x) around the new origin. Thus, let us see how transformation
matrix works
_
x

_
=
_
cos sin
sin cos
__
x
y
_
=
_
1 0
0 1
__
x
y
_
=
_
x
y
_
As a matter of fact, the above function becomes
y +
1
2
=

1
_
x +
1
2
y =
1
2

1
_
x +
1
2
2
Now, we convert back to our normal origin (0, 0) by doing a backward transformation (This is
my favorite part of this technique).
(x, y)
_
x
1
2
, y
1
2
_
Therefore, our rotated function becomes
y
1
2
=
1
2

_
1

_
x
1
2
_
+
1
2
= 1
_
1

1 x
Putting everything together, our original and new functions are, respectively, y = g(x) =
_
1

1 x
and y = h(x) = 1
_
1

1 x. One can show that these two functions meet at (0, 1) and (1, 0).
Finding area between these two curves is left for interested readers. (Hint. Substitution technique).
2 Inequality Tales
Inequality is often regarded one of the hardest topics at high school curriculum level. Indeed, it is a
combination of rigorous algebra and a little bit of intuition (this can be achieved by regular practice,
at least to my own experience).
Problem 2 (IMO 1969 B3). Given real numbers x
1
, x
2
, y
1
, y
2
satisfying x
1
> 0, x
2
> 0 and x
1
y
1
>
z
2
1
, x
2
y
2
> z
2
2
. Prove that
8
(x
1
+x
2
)(y
1
+y
2
) (z
1
+z
2
)
2

1
x
1
y
1
z
2
1
+
1
x
2
y
2
z
2
2
Give necessary and sucient condition for equality.
Theorem 4 (A tender version of Holders inequality).
Let a
i
, b
i
, c
i
be postive numbers for i = 1, n, one has
_
n

k=1
a
k
__
n

k=1
b
k
__
n

k=1
c
k
_

_
n

k=1
3
_
a
k
b
k
c
k
_
3
Solution.
We may rewrite the inequality as
8
_
(x
1
+x
2
)(y
1
+y
2
) (z
1
+z
2
)
2

_
1
x
1
y
1
z
2
1
+
1
x
2
y
2
z
2
2
_
Again, we need a little help from Cauchy-Schwarz inequality
(x
1
+x
2
)(y
1
+y
2
) (z
1
+z
2
)
2
(

x
1
y
1
+

x
2
y
2
)
2
(z
1
+z
2
)
2
Thus, it suces to show that
_
(

x
1
y
1
+

x
2
y
2
)
2
(z
1
+z
2
)
2

_
1
x
1
y
1
z
2
1
+
1
x
2
y
2
z
2
2
_
8
3
or
_
1
x
1
y
1
z
2
1
+
1
x
2
y
2
z
2
2
_
(

x
1
y
1
+z
1
+

x
2
y
2
+z
2
)(

x
1
y
1
z
1
+

x
2
y
2
z
2
) 8
This is the traight application of Holder inequality above and thus, we are done.
There is a list of generalization and related inqualities for you to enjoy proving
Problem 3 (1
st
generalization).
If x
i
> 0 and x
i
y
i
> z
2
i
then prove that
n
3
_
n

k=1
x
i
__
n

k=1
y
i
_

_
n

k=1
z
i
_
2

n

k=1
1
x
i
y
i
z
2
i
Problem 4 (High School Mathematics, China Tianjin No.3, 1992).
Given positive real numbers x
1
, x
2
, y
1
, y
2
satisfying x
1
y
1
> z
2
1
, x
2
y
2
> z
2
2
, prove that
(x
1
+x
2
)(y
1
+y
2
) (z
1
+z
2
)
2
+ 4
_
(x
1
y
1
z
2
1
)(x
2
y
2
z
2
2
)
Problem 5 (Strengthened version).
Given positive real numbers x
1
, x
2
, y
1
, y
2
satisfying x
1
y
1
> z
2
1
, x
2
y
2
> z
2
2
, prove that
4
(x
1
+x
2
)(y
1
+y
2
) (z
1
+z
2
)
2

1
_
(x
1
y
1
z
2
1
)(x
2
y
2
z
2
2
)
Here are the selelected inequalities from dierent contests around the world, the list will be regularly
updated for the sake of reference
Problem 6 (Moldova TST 2013).
For x, y, z > 0, show that
x
y
+
y
z
+
z
x

z(x +y)
y(y +z)
+
x(y +z)
z(x +z)
+
y(x +z)
x(x +y)
Problem 7 (Moldova TST 2013).
Consider the obtuse angled-triangle ABC and its side lengths a, b, c. Prove that
a
3
cos A +b
3
cos B +c
3
cos C < abc
Problem 8 (Moldova TST 2013).
Let x, y, z > 0. Show that
(xy +yz +zx)
_
1
x
2
+y
2
+
1
y
2
+z
2
+
1
z
2
+x
2
_
>
5
2
Hint. Let z = min{x, y, z}. Consider f(x, y, z) f(x, y, 0). Equality occurs when (x, y, z) = (k, k, 0)
and its permutations. Moreover, a new and nice inequality pops up as a present
xy
x
2
+y
2
+
x
y
+
y
x

5
2
4
Problem 9 (High School Mathematics, China Tianjin, No.4, 2013).
Let a, b, c be positive real numbers such that abc = 1. Prove that
1
b +c +
+
1
c +a +
+
1
a +b +

3
2 +
Hint. Let x
3
= a, y
3
= b and z
3
= c and use the following useful inequality x
3
+ y
3
xy(x + y) to
complete the proof.
Problem 10 (Romania Junior Balkan Team Selection Tests , 2013).
Let a, b, c > 0 such that a +b +c = 3. Prove that
a
2
(b + 1)
ab +a +b
+
b
2
(c + 1)
bc +b +c
+
c
2
(a + 1)
ca +c +a
2
Problem 11 (Iran TST 2013, Day 2, Problem 2).
Let a, b, c be sides of a triangle such that a b c, prove that
_
a(a +b

ab) +
_
b(a +c

ac) +
_
c(b +c

bc) a +b +c
Extension: The condition a, b, c are sides of a triangle is not necessary here. In fact, by the rst
script is written, there is no submitted solution on the famous problem solving website http://www.
mathlinks.ro. Here is the sketch of the proof in the general case.
Solution.
Firstly, we need to play around a bit with algebra. Thus, a strong algebraic manipulation skill is
required here. The readers are encouraged to double-check these computation in order to convince
themselves about the validity of the proof.
_
a(a +b

ab)

a(

a +

b)
2
=

a
_
_
a +b

ab

a +

b
2
_
=
3

a(

b)
2
2(2
_
a +b

ab +

a +

b)
Thus, it suces to show that
3

a(

b)
2
2
_
a +b

ab +

a +

b
+
3

c(

c)
2
2
_
b +c

bc +

b +

c
+
3

b(

c)
2
2
_
a +c

ac +

a +

c
(

b)
2
+ (

c)
2
There is a good chance that the LHS could be compared with an expression containing simpler terms,
in fact, we may need the following estimation
2
_
a +b

ab < 2(

a +

b)
Thus, we only need to show that

a(

b)
2

a +

b
+

b(

c)
2

a +

c
+

c(

c)
2

b +

c
(

b)
2
+ (

c)
2
5
or
(

c)
2

a +

c

(

b)
2

a +

b
+
(

c)
2

b +

c
This is exactly what we are looking for, a nice, neat and good looking form of inequality. Now, it is
the time for some serious algebra, lets take a walk...
(

b)
2
+ 2(

b)(

c) + (

c)
2

a +

c

(

b)
2

a +

b
+
(

c)
2

b +

c
or
(

b)
2
(

c)

a +

b
+ 2(

b)(

c)
(

c)
2
(

b)

b +

c
Thus, we only need to check that

a +

b
+ 2

b +

c
However, it is correct since the RHS is less than 1 and the LHS is more than 1. Therefore, needless to
say, we just complete the proof here. Throughout the entire proof, we have not used triangle condition
yet so this version is clearly stronger.
Problem 12 (Archimedes Contest).
Prove that
cos
2
x + cos
2
y + cos
2
z +|cos x|+|cos y|+|cos z|+cos
2
(x +y) + cos
2
(y +z) + cos
2
(z +x) 3
for all x, y, z R.
Solution.
The following trigonometric formulas may be handy sometimes
cos
2
(a) =
1 + cos(2a)
2
cos(A) + cos(B) =
1
2
_
cos
_
A B
2
_
+ cos
_
A +B
2
__
Using these formula, we may rewrite the LHS as follows
LHS = 3 +
x,y,z

cyc
|cos(x)|+
1
2
x,y,z

cyc
{cos[2x] + cos[2(x +y)]}
= 3 +
x,y,z

cyc
|cos(x)|+
x,y,z

cyc
cos(2x +y) cos(y)
3 +
x,y,z

cyc
[|cos(y)||cos(2x +y)|+cos(2x +y) cos(y)]
3
Thus, the proof is complete here.
6
The following inequality is updated from website www.mathlinks.ro and personally, I think it is
extremely suitable for any olympiad context.
Problem 13 (Mathlinks forum, Jayjayniboon).
Let a, b, c be positive numbers, prove that
a
1 +bc
+
b
1 +ca
+
c
1 +ab

a
1 +a
2
+
b
1 +b
2
+
c
1 +c
2
Hint. We make the common denominators on both sides respectively and prove that the numerator
on the LHS is more than or equal to the RHS with notice that
a,b,c

cyc
(1 +a
2
)
a,b,c

cyc
(1 +bc)
by using Cauchy Schwarz (1 +a
2
)(1 +b
2
) (1 +ab)
2
.
The proof above leads us to the generalized inequality once suggested on Mathlinks forum as follows
Problem 14 (Generalized, Michael Rozenberg).
Let a, b, c and k are non-negative real numbers such that
a,b,c

cyc
(k
2
+a
2
) = 0. Prove that
a
k +bc
+
b
k +ca
+
c
k +ab

a
k +a
2
+
b
k +b
2
+
c
k +c
2
Problem 15 (Poland 2013 nal round P5).
Let k, m and n be three dierent positive integers. Prove that
_
k
1
k
__
m
1
m
__
n
1
n
_
kmn (k +m+n).
Solution.
The inequality is equivalent to
2(m
2
+n
2
+p
2
1)
k,m,n

cyc
k
2
(mn)
2
If the distances |k m|, |mn| and |nk| are all more than or equal to 2 then the problem is solved
because we can give the following estimation.
k,m,n

cyc
k
2
(mn)
2
4(k
2
+m
2
+n
2
) > 2(m
2
+n
2
+k
2
1)
Suppose that one of |k m|, |m n| and |n k| equals 1, we will see what happens to the above
inequality. WLOG, assume that m = n + 1. The inequality becomes
_
k
1
k
__
n + 1
1
n + 1
__
n
1
n
_
kn(n + 1) (k +n + 1 +n)
_
k
1
k
_
(n + 2)(n 1) k(n
2
+n 1) (2n + 1)
(k
2
1)(n + 2)(n 1) k
2
(n
2
+n 1) k(2n + 1)
0 k
2
k(2n + 1) + (n + 2)(n 1)
0 [k (n 1)][k (n + 2)]
7
We need to use a bit of reasoning here. If k n + 2 then k > n 1. Phew! It is not so hard here.
However, if k n 1 and k < n + 2 then n 1 k < n + 2 so k must be n or (n + 1) but m
is already (n + 1) (by the assumption). Therefore, k must equal m or n and this is a contradiction
(m, n and k are distinct postive integers).
Thus, the inequality holds for k, m, n postive distinct integers. We complete our proof here.
Problem 16 (Republic of Srpska 2005).
Prove that in an acute triangle the following inequality holds:
_
sin sin cos +
_
sin sin cos +
_
sin sin cos
3

6
4
.
Solution.
Here we go again with the trigonometric inequality. Thus, we may need some geometry formulas
between the triangle ABC and its circumcircle (O).
Sine Rule
sin A
a
=
sin B
b
=
sin C
c
Cosine Rule
cos A =
b
2
+c
2
a
2
2bc
cos B =
a
2
+c
2
b
2
2ac
cos C =
a
2
+b
2
c
2
2ab
and
Area(ABC) =
abc
4R
where R is the radius of the circumcircle of ABC
Furthermore, there is one geometric inequality that is well-known to us
a
2
+b
2
+c
2
9R
2
or it is interpreted as follows
cos(2A) + cos(2B) + cos(2C)
3
2
(1)
The proof of this inequality can be obtain using vector method, let G be the centroid of the triangle
ABC, one has

0 =

GA +

GB +

GC
3

OG =

OA +

OB +

OC
9OG
2
=
_

OA +

OB +

OC
_
2
= 3R
2
+ 2
_

OA.

OB +

OB.

OC +

OC.

OA
_
= 3R
2
+ 2R
2
(cos(2A) + cos(2B) + cos(2C))
with notice that OG
2
0 and

OA.

OB =

OA

OB

cos(

OA,

OB) and therefore, we assert the

inequality (1).
8
As a consequence, many other trigonometric inequalities can be derived from (1) such as
cos
2
(A) + cos
2
(B) + cos
2
(C)
3
4
(2)
and
sin
2
(A) + sin
2
(B) + sin
2
(C)
9
4
(3)
Problem 17 (USAMO 2013, Day 2, Problem 1).
Find all real numbers x, y, z 1 satisfying
min(

x +xyz,

y +xyz,

z +xyz) =

x 1 +
_
y 1 +

z 1.
Solution.
Without loss of generality, we may assume that z = min{x, y, z}. Therefore, we need to nd x
y z 1 such that

z +xyz =

x 1 +
_
y 1 +

z 1
Substitute a =

x 1 ; b =

y 1 and c =

z 1, the above equationt is equivalent to

_
c
2
+ 1 + (a
2
+ 1)(b
2
+ 1)(c
2
+ 1) = a +b +c
c
2
+ 1 + (a
2
+ 1)(b
2
+ 1)(c
2
+ 1) = (a +b +c)
2
c
2
+ (abc)
2
+
a,b,c

cyc
(ab)
2
+
a,b,c

cyc
(c)
2
+ 2 =
a,b,c

cyc
c
2
+ 2(ab +bc +ca)
c
2
+ (abc)
2
+ (ab)
2
+ (bc)
2
+ (ca)
2
+ 2 = 2(ab +bc +ca)
2
c
2
(a
2
+ 1)(b
2
+ 1) + 1 + (ab 1)
2
= 2c(a +b)
We will show that LHS RHS. In fact, one applies Cauchy Schwarz and AM-GM successively as
follows
c
2
(a
2
+ 1)(b
2
+ 1) + 1 + (ab 1)
2
c
2
(a
2
+ 1)(b
2
+ 1) + 1
c
2
(a +b)
2
+ 1
2c(a +b)
The equality occurs if and only if
_

_
a =
1
b
c(a +b) = 1
=
_
_
_
ab = 1
c(a +b) = 1
If we let a = k be any arbitrary positive real number, we have
_

_
b =
1
k
c
_
k +
1
k
_
= 1
=
_

_
b =
1
k
c =
k
k
2
+ 1
Thus, the set of solutions is of the form
(x, y, z) =
_
k
2
+ 1,
1
k
4
+ 1,
k
2
(k
2
+ 1)
2
+ 1
_
for any k > 0.
and its permutations.
9
Problem 18 (In the World of Mathematics, Volume 18 (2012), Issue 4 (416)).
Let 0 < a, b, c < 1 be such that 8abc 1. Prove that
(ab 1)(bc 1)(ca 1) < ln a ln b ln c
Solution.
We may rewrite the inequality to make it look a bit more formal as follows
(1 ab)(1 bc)(1 ca) > (ln a)(ln b)(ln c)
Now, the idea is to use the condition 8abc 1 eectively. We try the following estimation
1 x
x
ln x for all x (0, 1)
However, this does not work well simply because we have not utilise the given condition. However,
we are getting close to what we want, the above estimation is far away from ln x, with a little
modication, it allows us to have the better approximation as shown below
1 x
2
2x
ln x for all x (0, 1)
This approximation is better because it is closer to ln x as
1 x
x

1 x
2
2x
for all x (0, 1)
To prove the result above, we consider the following function on (0, 1)
f(x) =
1 x
2
2x
+ ln(x)
Taking the derivative of this yields
f

(x) =
1 2x
3
+ 2x
2x
2
Lemma
2x
3
+ 1 > 2x for all x R
+
We only need to use AM-GM to prove this lemma
2x
3
+
x
2
2x
2
2x
2
+ 1 2

2x
Adding these inequalities, we have
2x
3
+ 1
_
2

2
1
2
_
x > 2x
Thus, we proved the lemma successfully. Now, back to the original inequality, we only need to show
that
(1 ab)(1 bc)(1 ca) (1 a
2
)(1 b
2
)(1 c
2
)
However, with a little help from (1 ab)
2
(1 a
2
)(1 b)
2
, we manage to prove this inequality
completely.
10
Problem 19 (USAMO 2013, Day 1, Problem 1).
In triangle ABC, points P, Q, R lie on sides BC, CA, AB respectively. Let
A
,
B
,
C
denote the
circumcircles of triangles AQR, BRP, CPQ, respectively. Given the fact that segment AP intersects

A
,
B
,
C
again at X, Y, Z respectively, prove that Y X/XZ = BP/PC.
Solution.
We consider the similar triangles AZQ ACP and ARY APB, one has
ZQ
PC
=
AQ
AP
;
RY
BP
=
AR
AP
Therefore, by dividing side by side of these fractions, one obtains
PB
PC
=
AQ
AR

RY
ZQ
We wish that
AQ
AR

RY
ZQ
=
XY
XZ
We establish the following lemma
Lemma

XQZ =

XRY
The proof of this lemma can be proceeded as follows:
We notice that

Y RB =

ZQA and

ARX =

CQX. Therefore,

XRY = 180
0

ARX

Y RB

XRY = 180
0

CQX

ZQA

XRY =

XQZ
11
The extensive use of Sine Rule is applied on triangle RY X, QZX and ARQ, one has
XY
RY
=
sin XRY
sin Y XR
;
XZ
ZQ
=
sin XQZ
sin ZXQ
;
AQ
AR
=
sin ZXQ
sin Y XR
Therefore, we nally obtain
XY
XZ
=
PB
PC
Problem 20 (USAJMO 2013, Day 2, Problem 5).
Quadrilateral XABY is inscribed in the semicircle with diameter XY . Segments AY and BX
meet at P. Point Z is the foot of the perpendicular from P to line XY . Point C lies on such that
line XC is perpendicular to line AZ. Let Q be the intersection of segments AY and XC. Prove that
BY
XP
+
CY
XQ
=
AY
AX
.
Solution.
Firstly, we investigate more to see what happens to the identity if we use similar triangles here.
Thus, it is easy to show that PAX PBY and QAX QCY . Therefore, we can obtain the
following equal ratios:
AX
PX
=
BY
PY
;
AX
QX
=
CY
QY
Thus, the identity to prove is equivalent to
BY
XP
+
CY
XQ
=
AY
AX
AX
XP
BY +
AX
XQ
CY = AY
BY
PY
BY +
CY
QY
CY = AY
BY
2
PY
+
CY
2
QY
= AY
12
Therefore, the identity now looks a little neater than before. However, we may construct more points
to aid the process of nding solution. Let M and N be points on AY such that BM AY and
CN AY . Consequently, it follows that BY
2
= Y M.Y P and Y C
2
= Y N.Y Q. Thus, It suces to
show that
Y M +Y N = AY
Notice that this identity occurs if and only if BC is parallel to AY which leads us to the important
lemma
Lemma
BC is parallel to AY
The proof of the lemma is as follows:
We notice that AP is the perpendicular of

ZAB, thus

Y AB =

ZAY . Moreover, since AZ is
parallel to CY, one has

CY A =

ZAY and therefore,

Y AB =

CY A or BC is parallel to AY
and this completes the proof of the original identity.
Problem 21 (IMC 2013, day 2, problem 1).
Let z be a complex number with |z + 1|> 2. Prove that |z
3
+ 1|> 1.
Solution (Viet Quoc Hoang).
We denote the following quantities (note that this kind of substitution has apperared in many other
problems so it becomes a critical technique here)
_
S = z +z
P = z.z.
The rst thing to see is S and P are both real numbers (P is positive in addition to that). From the
given condition, we have
|z + 1| > 2
(z + 1)(z + 1) > 4
P +S > 3
Moreover, we need to show that
|z
3
+ 1| > 1
(z
3
+ 1)(z
3
+ 1) > 1
P
3
+S
3
3PS > 0
(P +S)(P
2
PS +S
2
) > 3PS
However, since P +S > 3 > 0 and P
2
PS+S
2
PS (note that if PS < 0, we obtain the conclusion
immediately). Thus, we have just done proving the problem.
Solution (MathUniverse).
Let z + 1 = a +bi. Then we need to prove: a
2
+b
2
= t > 4 implies:
(a
2
+b
2
)[(a
2
b
2
3a + 3)
2
+b
2
(2a 3)
2
] > 1
Or after using Lagranges identity:
[a(t + 3) 3t]
2
+b
2
(t 3)
2
> 1
13
If a 0 or b
2
1, then inequality obviously holds. Therefore, suppose a 0 and b
2
1. If we
denote b
2
= x 1, we need to prove:
f(t) = [(t + 3)

t x 3t]
2
+x(t 3)
2
> 1
f

(t) = 2
_
(t + 3)

t x 3t

t x +
t + 3
2

t x
3
_
+ 2x(t 3)
Since:
(1) :

t x +
t + 3
2

t x
3
3t + 1
2

t x
3 >
3t + 1
t
3 > 0
Also, using Am-Gm:
(2) :
(t + 3)

t 1
t
3 =
(t + 3)(t 1)
t

t 1
3
(t + 3)(t 1)
t
_
t 1
4
+ 1
_ 3 = 1
4
t
0
We conclude that f

(t) > 0, so we only need to prove f(4) > 0 what is:

(7

4 x 12)
2
+x > 1 9(16x 15)
2
0
.
Solution (Mathlinker).
Let z = re
i
. We are given |z + 1|> 2, which is equivalent (after squaring) to r
2
+ 2r cos > 3. By
solving the quadratic x
2
+ 2x cos 3 = 0, it becomes r >

cos
2
+ 3 cos > 0.
What do we need to prove? That |z
3
+ 1|> 1. After squaring it becomes r
6
+ 2r
3
cos(3) > 0. Since
r > 0, we can divide by r
3
and it becomes simply r
3
> 2 cos(3) = 6 cos 8 cos
3
.
By plugging the minimal value for r, this problem is equivalent to the inequality (

cos
2
+ 3
cos )
3
6 cos 8 cos
3
. Letting t = cos and expanding this, it simplies to:

t
2
+ 3(4t
2
+ 3) t(15 4t
2
)
This is almost only a function of t
2
! Note that when 1 t 0, the RHS is non-positive and the
inequality is evident. When 0 t 1, we can square both sides and let t
2
= a:
(a + 3)(4a + 3)
2
a(15 4a)
2
It seems like a cubic inequality, but the main terms cancel each other and we remain with the quadratic
(8a 3)
2
0, a trivial inequality.
So when = arccos
_
3
8
, r =

cos
2
+ 3 cos =
_
3
2
, we get:
|z + 1|= 2, |z
3
+ 1|= 1
14