____________________________________________________________________________________________________

Subject ECONOMICS

Paper No and Title Paper 1: Quantitative Methods-I (Mathematical methods)

Module No and Title Module 39: Game Theory I

Module Tag ECO_P1_M39

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

TABLE OF CONTENTS
1. Learning Outcomes
2. Introduction
3. What is a game?
4. Definitions
5. Illustrations
6. Saddle point for a two person zero sum game
7. Summary

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

1. Learning Outcomes

After studying this module, you shall be able to

 Know basic concepts of Theory of Games, its assumptions and its basic types.
 Learn the concepts of selecting the optimal strategies while playing a game
 Find the Nash Equilibrium based on pure strategies for two person games.
 Determine the saddle points for a two person zero sum game and hence the value
of the game.

2. Introduction
In today’s world we come across conflict and competition every now and then. A
competitive situation occurs if two or more individuals are making decisions in situation
that involves conflicting interests and where the outcome is controlled by the decisions of
all parties concerned. Restricting ourselves to purely fair situations, we deal with those
competitive situations where each participant acts in a rational manner and tries to
resolve the conflict of interests in his favor. In this context, this competitive situation is
referred to as game.
The term game is defined as an activity between two or more persons, each of whom
makes his own decision to defeat others, obeying a set of rules, at the end of which each
person gets some benefit or suffers a loss. The players of the game deal with the problem
of choosing a rational course of action where the player is confronted in a situation whose
outcome depends not only upon his own actions but also upon the action of others. Now
the question arises as to what strategy should be followed by a player to make decisions
while playing a game. The answer to this question lies in the logic of rational decisions
which is nothing but a quantitative technique called theory of games or game theory.
Game theory provides the general rules concerning the logic that underlines strategic
behavior of all types. It is a mathematical theory that deals with the general features of
competitive situations in a formal way. It places particular emphasis on the decision
making processes of the adversaries. In the year 1928, John Von Neumann developed the
theory of games and later got a book entitled “The Theory of Games and Economic
Behavior” published with Oskar Morgenstern. The approach of game theory is now
widely used in economics, Business Administration, Sociology, Psychology and Political
Science as well as in Military Training.
A game is assumed to possess the following properties:
 The number of competitors or players is finite.
ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)
MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

 Each player has available to him a finite set of possible courses of action.
 Each combination of courses of action by the player results in an outcome which
may be in the form of points, money or payments (positive, negative or zero) to
each competitor.

3. What is a game?

With reference to the literal meaning of game defined above, we now consider a
simple example of a game. Suppose there are two players say A and B who are sitting on
a table with a partition in between. Each has with him 50 beads each of color red and
blue. At the signal of the referee each player has to put a bead in the container kept in
front of him not visible to the other player. So each player has two choices open to him.
Each play is a combination of the choice played by both the players in one go. The results
of each of the possible four combinations is known in advance to both the players but the
uncertainty in the game arises because neither player knows which bead his opponent
will put next. At the end of 50 plays the game is over and the points won by each of the
players are totaled and the winner is declared. The resulting gains or losses called the
pay-offs of each player can be represented by means of the following table:

Game 1 Player A
B Red Blue
A
Red 2 4

Blue -3 -2

This is the pay-off matrix for player A therefore the positive values represent the gain of
A to be paid to him by B and the negative values represent the loss of A or equivalently
the gain of B to be paid to him by A. The rows here represent the choices for player A
and the columns represent the choices for player B. Here if A chooses to put a red bead
and B chooses to put a blue bead then we examine the top right hand corner of the matrix.
So from the above payoff matrix for player A it can be seen that if player A choses to put
a red bead into the container he will win 2 points if B also puts in red bead and 4 points if
B puts in a blue bead.

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

Now the question that comes to mind is “If the above pay-off matrices are provided to the
players before the start of the play, what should be the course of action followed by each
of the player so as to maximize their respective gains or minimize their respective
losses?”
We first introduce a few definitions which would help us in deciding the most appropriate
course of action of a game:

4. Definitions

Strategy: A strategy for a given player is a set of rules that specifies which of the
available courses of action should be adopted by him at each play. It precisely refers to
the total pattern of choices employed by any player. The set of rules may vary from one
player to the other.

Types of strategies:

 Pure strategy: A pure strategy is a specific choice of strategy among the possible
moves a player may choose and then sticking to his choice once and for all. Such
a strategy is followed in a deterministic situation. In case of pure strategy all
through, the players are playing with a single strategy.

 Mixed Strategy: If a player always keeps on guessing as to which courses of

action will be adopted by the opponents, then there arises the probabilistic
situation. In case of mixed strategy the players randomize their strategies to assign
a probability to each choice and play their choices according to those
probabilities. A pure strategy is a special case of a mixed strategy.

 Dominant Strategy: If there is one optimal choice of strategy for each player no
matter what the other player does, such a strategy is called a dominant strategy.

 Nash Equilibrium: A pair of strategies is Nash equilibrium if each player is

making the optimal choice, given the other person’s choice. Equivalently, Nash
equilibrium can be interpreted as a pair of expectations about each person’s
choice such that when the other person’s choice is revealed; neither individual
wants to change his strategy.

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

Types of games:

 Two person games: In two person games the players may have many possible
choices for each play of the game with them but the number of players remains
only two.

 n-person games: In case of more than two persons, the game is referred to as n-
person games.

 Zero sum game: A game in which the sum of the points or payments of all the
competitors is zero for every possible outcome of the game is called a zero sum
game. In such a game the sum of points won equals the sum of points lost. For
example two persons game is always zero sum game since one player loses what
the other player wins.

 Non-zero sum game: If the pay-offs from any sum of play of the game is either
positive or negative but not zero then such a game is said to be non-zero sum
game.

 Value of the Game: It refers to the average payoff per play of the game over an
extended period of time throughout the course of the game. We illustrate this
concept by means of a few examples.

5. Illustrations

1. We again consider Game 1 discussed earlier.

Game 1 Player A
B Red Blue
A
Red 2 4

Blue -3 -2

Here it can be seen that starting with player A, he would choose red bead for each play of
the game and in that case player B cannot win. So playing intelligently, player B will
respond each time by choosing a red bead in order to minimize his own loss. In such a
game it is easy to determine individual player’s strategies. So the solution can be given in
ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)
MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

the tuple form as (Red, Red) where the first component of the tuple represents the optimal
strategy of A and the second component of the tuple represents the optimal strategy
chosen by player B. In such games each player is making one choice and sticking to it.
So this is the case of pure strategy.
But all games may not have such simple solutions. For instance consider the following
pay-off for Player A:

1. Game 2

B Red Blue
A
Red 2 -4

Blue -1 3

In the above case the player A would not chose red bead in each play because if player B
chooses to put a blue bead B wins 4 points himself. Similarly Player B would not choose
blue bead hoping to win 4 points because player A could counter the situation with a blue
bead and win 3 points himself. In such a situation, the players would play each of their
choices a part of the time only. This would be a case of mixed strategy.
We now consider another game:
2. Suppose there are two people playing a game where Player A will write one of the
two words from “left” and “right” on a piece of paper and Player B will write one
of the words from “up” and “down” simultaneously. But none of them can see
what the other player has written. Once they have written their choices, the papers
will be examined and they will each get the payoff according to the payoff matrix
given below:
Game 3
B Up Down
A
Left 2,3 1,2

Right 3,2 2,1

In this matrix the payoff to A is the first entry in the box and payoff to B is the second
entry. Now player A has two strategies: either he can choose left or right. Clearly for
Player A it is always better to select right since his payoffs from this choice (3 or 2)
are always greater than the corresponding entries in the row representing left (2 or 1).

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

Similarly it is always better for Player B to select up since 3 and 2 dominate 2 and 1.
So the optimum or the equilibrium strategy expected in this game is that A plays right
and B plays up. This is the case of dominant strategy as A will get a higher payoff if
he plays right irrespective of whatever choice B makes. Similarly whichever choice A
makes, B will get a higher payoff if he plays up. So these choices dominate the
alternatives and the equilibrium exists in dominant strategies. So in this game, A will
receive an equilibrium payoff of 3 and B will receive an equilibrium payoff of 2.
As seen, Game 3 has a very simple solution as it was the case of dominant strategy
equilibrium, but this doesn’t happen that often. Let us consider another game
3. Game 4

B Up Down
A
Left 3,2 1,1

Right 1,1 2,3

In Game 4, when B chooses up then A’s payoffs are 3 if he goes with left and 1 if he
goes with right. On the other hand, if B chooses down then A’s payoffs are 1 if he
goes with left and 2 if he goes with right. This means that when B chooses up, A
would want to go with left and when B chooses down, A would want to go with right.
Thus A’s strategy would depend on what he thinks B would do. Similarly B’s
strategy would depend on what he thinks A would do. But none of them knows what
the other player would do. However each player may have some expectation about
what the other player’s choice may be.
We now intend to find a pair of strategies for which A’s choice is optimal, given B’s
choice and B’s choice is optimal given A’s choice. Such a pair of strategies is called
the Nash equilibrium where each player is making the optimal choice, given the
other player’s choice. Here the strategy (Left, Up) is a Nash equilibrium. This is
because if A chooses left then the best choice for B is up since the payoff to B from
choosing up is 2 and from down is 1. And if B chooses up, then the best choice for A
is left since the payoff to A from choosing left is 3 and from right is 1. Similarly, it
can be seen that (Right, Down) is also a Nash Equilibrium.
Example1. Each of the two players A and B shows either one or two fingers
simultaneously. The four possible outcomes along with the payoffs are as follows:
i. Both the players show one finger for which player B will get two rupees
from player A.

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

ii. Player A shows one and player B shows two fingers for which player A
will get three rupees from player B.
iii. Player B shows one and player A shows two fingers for which player A
will get three rupees from player B.
iv. Both the players show two fingers for which A will get four rupees from
player B.

Form the payoff matrix for the game and find the value of the game.
Solution: We now represent the game in the form of the payoff matrix for player A

B
A One finger Two fingers
One finger -2 3
Two fingers 3 4

Now playing intelligently player B would play his first column on each play
and player A would respond by playing his second row on each play of the game.
So the Nash equilibrium for the game is (Two, One). As a result B wins two
rupees on each play of the game. So the value of the game is 3 indicating that A is
the winner.

Example 2: Determine the optimum strategies for the two players A and B and
find the value of the game for the following payoff matrix:
B
A I II III IV
I 3 -1 4 2
II -1 -3 -7 0

III 4 -6 2 -9

Solution: Playing cautiously each player would assume the worst and act
accordingly. If player A plays with strategy I then B would play with strategy II
and would win 1 point otherwise he would face a loss of 3, 4 and 2 points if he
plays with strategy I, III or IV respectively. If player A plays with strategy II then
worst would happen to him if B would play with strategy III which would make
him win 7 points. If player A plays with strategy III then the worst he can expect
is losing 9 points which would be won by B if he plays with strategy IV. Thus
ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)
MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

player A can make the best of the situation by expecting the maximum among
these minimal payoffs i.e. highest of -1,-7 and -9. So player A would play with
strategy I
Similarly, from the point of view of player B, if he plays with Strategy I
then the maximum he can lose is 4 points if A plays with strategy III. Playing
with strategy II player B would never face any loss whatever may be the strategy
of player A but in this case A would play with strategy I in order to minimize his
loss. If player B plays with strategy III he can at the most face a loss of 4 points if
player A plays with strategy I. Lastly while playing with strategy IV, player B can
get a loss of maximum 2 points if A plays with strategy I. Thus player B can make
the best of the situation by expecting the minimum among these maximal payoffs
i.e. least of 4,-1, 4 and 2. So player B would play with strategy II. So the Nash
equilibrium exists in this case and is given by (I, II) with I as the optimal strategy
of A and II as the optimal strategy of B. Hence in each play, B will win 1 point
and X will lose 1 point. The value of the game is therefore -1.
It should be noted here that the above two game problems are problems of pure
strategy for which Nash equilibrium is shown to exist. . However it can be observed
that for a game, a Nash Equilibrium of the type defined above may or may not exist. For
such problems Nash Euilibrium based on mixed strategies can be found which would be
seen later. Moreover, even if the Nash Equilibrium exists, it may or may not be unique.
For example, Game 4 had two Nash Equilibria namely (Left, Up) and (Right
Down).

6. SADDLE POINT FOR A TWO PERSON ZERO SUM GAME

Let us consider a two person zero sum game with a given pay off matrix. As the
interest of one player opposes the interest of other, therefore if player A wants to
win, player B would try to minimize the gains of A.
Thus if player A adopts a particular strategy player B would like to play back with
that strategy for which the payoff to A is minimum for that particular strategy.
Also the both the players are aware of this common mindset of each other.
In simple words, the objective of player A would be to select that strategy for
which the minimum assured return is maximum. So he would try to identify the
maximum of the minimal pay off and play the corresponding maximum strategy.
Similarly the optimal course of action for player B would be to go with that particular
strategy for which his maximum loss will be a minimum. So he needs to play with
minimax strategy.
This selection of maximin and minimax strategies of two players is based upon
the maximin-minimax criterion which ensures the best of the worst results.

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

The maximin payoff is called maximin value and minimax payoff is the minimax
value when both of these values are equal then the common value is called the
value of the game.
The position in the payoff matrix where the maximum of row minima coincides
with the minimum of column maxima is called the saddle point of the game and
the pay off value corresponding to the saddle point gives the value of the game.
Hence at the saddle point maximin and minimax values of the game are equal.
Remarks

• Saddle point of a game need not be unique but value of game is unique.
• In general, maximin value  value of the game  minimax value.
• At saddle point in a payoff matrix maximin value = minimax value
• Any two saddle points in the game will have the same value

METHOD OF DETERMINING A SADDLE POINT

 Select the minimum element of each row of the payoff matrix and underline them.
 Select the maximum element of each column of the payoff matrix and strike them
off.
 If there appears an element in the payoff matrix which is striked off and
underlined both, the position of that element is a saddle point of the matrix and
the strategies corresponding to that position are the optimal strategies for the two
players.

Now let us consider a game for which the payoff for player A is given by -
B B1 B2 B3
A
A1 4 6 4
A2 3 -1 0
A3 4 7 1

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

We now find the saddle point of this game

B B1 B2 B3
A
A1 4 6 4
A2 3 -1 0
A3 4 7 1

Since the first and the third entry of first row are both striked off and underlined.
Therefore, there exist two saddle points for this game given by (A1, B1) and (A1,
B3) and the value of game is 4 units. Hence player A wins the game by 4 units.
Hence for the game with pure strategies, the saddle points can always be found by
this method. Here it can be observed that Nash Equilibrium also exists for this
game, given by (A1, B3) and (A1, B1).
It is important to mention here by that for two person zero sum games, saddle
points and pure strategy Nash equilibrium are the same.

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I
____________________________________________________________________________________________________

7. Summary

 Game theory deals with the general analysis of strategic interaction.

 Game theory can be used to study economic behavior, political
negotiations, parlor games etc.
 The earliest game theory was first developed by John von Neumann in
1928
 He later got a book entitled “The theory of games and Economic behavior
“published with Oskar Morgenstern, the book that established game theory
as a field.
 A game consists of a set of players, a set of strategies corresponding to
each player and the payoffs associated with each of their strategies.
 A game is represented by means of a payoff matrix which presents in a
tabular form the gain or loss earned by players on selecting a particular
strategy.
 Analyzing the various strategies, each player selects an optimal strategy in
order to maximize his gain or equivalently minimize his losses. The final
payoff of the game is called the value of the game.
 The pair of optimal strategies constitutes Nash equilibrium.
 For a 2-person zero sum game , saddle point can be found by minimax
criterion
 Saddle point and pure strategy Nash equilibrium are the same for a two
person zero sum game.

ECONOMICS PAPER No. 1: Quantitative Methods-I (Mathematical methods)

MODULE No. 39: Game Theory I