You are on page 1of 5

NATIONAL UNIVERSITY OF SINGAPORE

FACULTY OF SCIENCE

SEMESTER 2 EXAMINATION 2010 2011


GEK1544

The Mathematics of Games

May 2011 Time allowed : 2 hours

INSTRUCTIONS TO CANDIDATES
1.

This examination paper contains a total of FOUR (4) questions and comprises FIVE (5) printed pages (including this page).

2.

Answer ALL questions.

3.

Candidates are each allowed to bring in ONE (1) hand-written, double-sided


help-sheet no larger than A 4 size.

4.

Candidates may use calculators. However, they should lay out systematically
the various steps in the calculations.

PAGE 2

Answer ALL the four questions.

Question 1

[ 25 marks ]

In natural craps, the following expectations are known :


X(pass) 0.0143 ,
X(dont pass) + 0.0143 .
(You are not required to derive the expressions.)
(i) Suppose that a casino in which the no result craps roll for Dont Pass
bettors is total of 3 (instead of the usual total of 12 ). That is, when the
initial roll has total equal to 3, it results in a loss to a Pass bettor, and neither
win nor loss for a Dont Pass bettor . Compute the expectation
X( dont pass , total of 3 is not a win) .
Justify your answer.
(ii)

Explain the notion of free odds in craps.

(iii) Under the condition in (i) , compute the expectation


X( dont pass plus free odds , total of 3 is not a win ) = 0. ???? .
(Your answer should be accurate up to 4 decimal places.) Justify your answer.

Question 2 starts on page three .

PAGE 3
Question 2

[ 25 marks ]

There are three identical fair dice. On each die, which is considered as a cube,
numbers 1 and 6 are arranged on the opposite faces, so are numbers 2 and 5, and
numbers 3 and 4.
(i) When the three dice are thrown and settle on a table, let the numbers on
the top faces be x, y, and z , where x, y and z are integers between 1 and 6,
inclusively. Explain why the sum of the numbers which appear on the three faces
opposite to the three top faces is given by
21 (x + y + z) .

(ii)

Using (i) , show that


P3 ( 10) = P3 (> 10) ,

where P3 ( 10) is the probability of obtaining three numbers (on the top faces)
from the three dice so that the sum is lesser than or equal to 10. Likewise, P3 (> 10)
is the probability of obtaining three numbers (on the top faces) from the three dice
so that the sum is bigger than 10.
(iii) Add one more die of the same type. Now we have four identical fair dice.
Explain why it is NOT possible to obtain an integer n, so that
P4 ( n) = P4 (> n) ,
where P4 ( n) is the probability of obtaining four numbers (on the top faces) from
the four dice so that the sum is lesser than or equal to n . Similarly, P4 (> n) is
the probability of obtaining four numbers (on the top faces) from the four dice so
that the sum is bigger than n .

Question 3 starts on page four .

PAGE 4
Question 3

[ 25 marks ]

Consider a zero-sum game with two players A and B . A has strategies A1 , A2


and A3 , while player B has strategies B1 , B2 and B3 . The payoffs for A are
shown in the table below .

(i)

B1

B2

B3

A1

A2

A3

Show that there are no dominating strategies for A and B .

(ii) Show that there are two Maxi - Mini strategies for A , namely, A2 and A3 .
Show that B1 is the only Maxi - Mini strategy for B .
(iii) Suppose that A chooses the Maxi - Mini strategy A2 . Then B will play B2
strategy to gain 4 . But if A anticipates that B plays B2 strategy, A will play
A1 strategy to gain 4 . The argument can be continued and B does not want to
wander away from the Maxi - Mini strategy B1 too far. Thus consider the diagram
B1

B2

A1

A2

Let p1 be the probability for A to play strategy A1 , and q1 the probability for B to
play strategy B1 . Find the optimal value of p1 for A to play the mixed strategies
between A1 and A2 .
(iv) On the other hand, suppose that A chooses the Maxi - Mini strategy A3 .
Then B will play B3 strategy to gain 4 . But if A anticipates that B plays B3
strategy, A will play A2 strategy to gain 4 . Thus consider the diagram
B1

B3

A2

A3

Let p2 be the probability for A to play strategy A3 , and q2 the probability for B to
play strategy B1 . Find the optimal value of p2 for A to play the mixed strategies
between A2 and A3 .
(iv) Should A choose A2 as the Maxi - Mini strategy to play the mixed strategies
as described in (iii) , or choose A3 as the Maxi - Mini strategy to play the mixed
strategies as described in (iv) ? Justify your answers.
Question 4 starts on page five .

PAGE 5
Question 4

[ 25 marks ]

In Game Theory, consider two companies, one called Good Company, and the other
one Bad Company. There are two job seekers, one called Good Employee, and the
other Bad Employee. Good Company is willing to pay a higher wage (W ) than
Bad Company. That is,
Wgood > Wbad .
(4.1)
However, there is no real way to tell in advance which job seeker will be of the
good or bad type, except the signal on the cost of education (C) the job seeker
paid . Good Employee knows that she/ he deserves to be paid more for her/ his
productivity (P ) , so she/ he invested more on education than Bad Employee. That
is,
Cgood > Cbad > 0 .
(4.2)
Assume that productivity P is given by
Pgood = k Cgood

and

Pbad = k Cbad ,

(4.3)

where k is a fixed positive number . Set


Payoff for a company (successful hiring)
= productivity of the hired employee wage to be paid .
Payoff for a company (unsuccessful hiring) = 0 .
Payoff for a person (successfully employed)
= wage cost invested in education .
Payoff for a person (not employed by either company) = 0 .
Suppose that each company makes only one offer, and each job seeker applies to
only one company. Moreover, assume that
Wgood = h Cgood

and

Wbad k Cbad > 0 .

(4.4)

Here h is a fixed positive number . Provide a relation on k and h so that the


strategies
Good Company seeks to hire Good Employee (only) ; Bad Company seeks to hire
Good Employee (only) ; Good Employee applies to Good Company (only) , and
is ready to work once hired ; Bad Employee applies to Bad Company (only) , and is
ready to work once hired
(4.5)
form a Nash equilibrium . Justify your answers [ specifically, check that the answer
you provide really makes the strategies in (4.5) into a Nash equilibrium ] .

END OF PAPER.