Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
(a) (5 marks) Find the value of the constant c. Also find the PMF of X.
Rubric: 1 mark for the sample space, 1 mark for the expression of
B.
Step 2.
Rubric: 1 mark for the correct formula for conditional pmf, 1 mark
for substituting the values correctly into the formula and 1 mark for
the final answer.
In case the student does not use the same set theoretic notation
as above, please do not deduct marks. Please look at the logical
consistency of his/her answers, regardless of notation.
Second method:
Step 3. Write the formulae for PX (B) and the pmf and evaluate PX (B).
X
PX (B) = PX (x).
x∈B
(
1
8 if x = 0, π4 , . . . , 7π
4
PX (x) =
0 otherwise .
Therefore X 3
PX (B) = PX (x) = .
8
x∈B
Step 4. Therefore (
1
3 if x = 0, π4 , π2
PX|B (x) =
0 otherwise .
The pirate knows that there are 19 gold coins hidden in the North, 16
in the East, 18 in the South and 18 in the West. The probability of
finding treasure is the same in all these four directions.
NA NC
WC EA
o
WA EC
SC SA
Answer 3.
(a) The assumption that all the attempts are independent necessitates that
the attempts in each of the 8 directions be regarded as a separate Geo-
metric random variable. Since the probability in all clockwise directions
is the same, the answer can also be expressed as the sum of 2 Pascal
random variables.
Rubric: 2 grace marks are given in case the number of attempts to find
treasure is modelled as a random variable having finite range. In this
case 1.5 marks are for the pmf and 0.5 for the expectation.
In case the student indicates that the random variable can be described
using the Geometric or Pascal distribution, or any other discrete random
variable having infinite range, with enough justification for this answer,
without calculating the pmf or the expectation, s/he is awarded 3 marks.
In case a discrete random variable having infinite range is used to model,
the remaining 2 marks of the question are for calculating the pmf and
the expectation.
The answer: My assumption is that the pirate knows the number of gold
coins located in each of the 8 directions. Without this assumption, the
problem would be incorrect, because if the pirate continues to try and
find treasure in say the East Anticlockwise direction, after having found
the treasure, then the probability of finding treasure in that direction
is zero, which contradicts the independence assumption, because the
conditional probability of finding the treasure must remain the same as
earlier.
So let the probability of finding the treasure in the NA direction be q.
This means that the probability of finding treasure in the NC direction
must be 2q. It also follows that the probability of finding treasure in
EA, WA, SA is q and the probability of finding treasure in the EC, WC
and SC directions is 2q. Let the number of attempts in direction EA be
denoted XEA , and likewise for the rest of the directions. Let the total
number of attempts to find all the treasure be X. Then
X = XEA + XEC + XW A + XW C + XSA + XSC + XN A + XN C .
Then
E(X) = E(XEA ) + E(XEC ) + E(XW A ) + E(XW C )
+ E(XN A ) + E(XN C ) + E(XSA ) + E(XSC )
4 2
= +
q q
6
= .
q
Other solutions with different assumptions were graded on a case by
case basis. In each instance up to half a mark was deducted if the
independence assumption on the trials was not taken into account.
(b) Two approaches to solving this are described, along with rubric. Alter-
native answers were graded on a case by case basis.
First method
Step 1. Let X be the number of moves that the knight can make on any
of the 64 possible squares. The sample space consists of all 64
squares. The range of X is
SX = {2, 3, 4, 6, 8}.
Second method
Step 1. Let X be the number of moves that the knight can make on any
of the 64 possible squares. Then X can be expressed as the sum
of eight random variables
X = X1 + X2 + . . . + X8
where Xj is the number of possible moves that the knight can
make in direction j. The possible directions are: north-left,
north-right,...,west-left, west-right.
Rubric: 3 marks for expressing this idea.
Step 2. The expectation of each of these 8 random variables is the same
because of symmetry. So it is sufficient to fix one direction, say
north-right and calculate the expectation. This is a Bernoulli
random variable. The probability that there is a possible move
7×6
in the north-right direction is , because this move is not
64
possible in the extreme right column, nor in the top two rows.
Thus the required expectation is
7×6 1
E(X) = ×8=5 .
64 4
Rubric: 2 marks for the above argument.