10-03-2025

Theory of Games

 The term ‘game’ refers to a situation of conflict and competition in

which two or more competitors (or participants) are involved in the
decision-making process in anticipation of certain outcomes over a
period of time.
 The competitors are referred to as players.
 A player may be an individual, group of individuals, or an organization.

Game Theory  Examples of competitive interactions:

 Pricing of products, where sale of any product is determined not only by
Sessions 15 & 16 its price but also by the price set by competitors for a similar product
 The success of any TV channel program largely depends on the
competitors present in the same time slot and the program they are
telecasting.

Theory of games Theory of Games

 Theory of games provides a series of mathematical models that  Number of players

may be useful in explaining interactive decision-making concepts,  Two-person game: a game involving two players
where two or more competitors are involved under conditions of  N-person game: game involving several players
conflict and competition.
 Sum of gains and losses
 Such models provide an opportunity to a competitor to evaluate
not only his personal decision alternatives (courses of action), but  Zero-sum game: when the sum of all the gains is exactly equal to the
sum of all losses
also the evaluation of the competitor’s possible choices in order to
win the game.  Non-zero-sum game: when the sum of all gains is not equal to the sum
of all the losses
 A game is defined in terms of:
 Number of Players
 Sum of gains and losses
 Strategy

Theory of Games Types of strategies followed by players

a. Pure Strategy A particular strategy that a player chooses to play

 Strategy The strategy for a player is the list of all possible actions again and again regardless of other player’s strategy, is referred as
(moves, decision alternatives or courses of action) that are likely to pure strategy. The objective of the players is to maximize their gains
be adopted by him for every payoff (outcome). or minimize their losses.
 It is assumed that the players are aware of the rules of the game b. Mixed Strategy A set of strategies that a player chooses on a
governing their decision alternatives (or strategies). particular move of the game with some fixed probability are called
 The outcome resulting from a particular strategy is also known to the mixed strategies.
players in advance and is expressed in terms of numerical values (e.g.  Thus, there is a probabilistic situation and objective of the each player is
money, per cent of market share or utility). to maximize expected gain or to minimize expected loss by making the
 The particular strategy that optimizes a player’s gains or losses, choice among pure strategies with fixed probabilities.
without knowing the competitor’s strategies, is called optimal
strategy. The expected outcome, when players use their optimal
strategy, is called value of the game.

1
 10-03-2025

Two-person zero sum game illustration Two-person zero sum game illustration
 Suppose there are two firms, A and B, selling a competing product, and are
now competing to gain a larger market share. Under these circumstances,
 Table below shows the payoffs from various combination of strategies of A and B
the share gained by one player must be equal to the share lost by the
other player (sum of gains and losses equal zero).

 Let us consider that each firm has a choice of three strategies to gain
market share: low advertising, high advertising, and quality improvement.
We also assume that currently they are sharing the market equally and
further that each of the firms can employ only one of the strategies at a  Under the (𝑎1 , 𝑏1) strategy combination, A gains a market share of 12 percent,
time. There are 9 possible combinations of moves (3 x 3) as shown below: while under the (𝑎1 , 𝑏2) strategy combination A loses 8 percent market share. The
pay-off matrix is drawn from A’s point of view—a positive pay-off indicates that
the firm A has gained the market share at the expense of firm B while negative
pay-off imply B’s gain at A’s expense.

 What is the best strategy for A and B, considering that both have the payoff
information associated with the various strategies, but do not know which strategy
the other player is going to adopt?

Two-person zero sum game illustration Saddle Point

 Conservatively, the players would expect the worst and proceed with their  In the example, it is clear that the maximin strategy of A (𝑎2) and the
actions. minimax strategy of B (𝑏3), both lead to the same pay-off. By adopting the
 Firm A would choose the maximum of the minimal pay-offs (maximin strategy). maximin strategy, A can stop B from lowering its (A’s) gain in market share
Since the minimal pay-offs corresponding to 𝑎1 , 𝑎2, 𝑎3 are respectively -8, 3, and - below 3 percent. Similarly, by adopting the minimax strategy, B can
10, firm A would select strategy 𝑎2. prevent A from taking any more share than 3 percent. The situation,
 Similarly, B would adopt a cautious approach. When B adopts 𝑏1, it can expect therefore, is one of equilibrium. The point of equilibrium is known as the
firm A to employ 𝑎1 which will ensure firm A the maximum advantage. In a similar saddle point.
way, if B adopts 𝑏2 and 𝑏3, in both cases A would adopt 𝑎2. To minimize the
advantage accruing to A, the form would select a strategy that would yield the
least advantage to A. Hence B would select 𝑏3. Thus, B would adopt a minimax
strategy.
 The resultant pay-off of A is 3. This is known as the value of the game (V), which
represents the final pay-off to the winning player.
 Since the pay-off matrix is drawn from A’s point of view, if the value of the game
is positive, as in the present case, it is favorable to A while if the game value is
negative, the game is favorable to B.

Example 1 Solution to Example 1

 Two leading firms NT Ltd. and SR Ltd. has been selling a similar shirting product
for several years. Recently, the marketing director of NT Ltd. asked his team SR’s Strategy
members to identify the right advertising strategy that the company should
No adv Med. Heavy Row
adopt for the product. The system group of NT developed the following data for
varying degrees of advertising: (𝑏1) Adv. (𝑏2) adv. (𝑏3) minimum

a. No advertising, medium advertising, and heavy advertising for both firms will result in NT’s No adv. (𝑎1) 50 40 28 28
equal market share. Strategy Med. adv. (𝑎2) 70 50 45 45
b. NT with no advertising: 40 percent of the market with medium advertising by SR and 28 Heavy adv. (𝑎3) 75 52.5 50 50*
percent of the market with heavy advertising by SR
c. NT using medium advertising: 70 percent of the market with no advertising by Sr and Col. maximum 75 52.5 50*
45 market with heavy advertising by SR
d. NT using heavy advertising: 75 percent of the market with no advertising by SR and
52.5 percent of the market share with medium advertising by SR.

 Based on the above information, identify the appropriate strategies for NR and
SR.

2
 10-03-2025

Example 2: Multiple saddle points Mixed strategies

Firm B’s Strategy

𝑏1 𝑏2 𝑏3 𝑏4 Row  It is possible that there is no saddle point of a game and so it is not possible
to find its solution in terms of the pure strategies – the maximin and the
minima minimax.
Firm A’s 𝑎1 4 –16 14 –15 –16
 Games without saddle points are not strictly determined. The solution to
Strategy such problems calls for employing mixed strategies.
𝑎2 –6 7 –4 –6 –6*
 A mixed strategy represents a combination of two or more strategies that
𝑎3 6 –2 0 –6 –6*
are selected one at a time, according to predetermined probabilities.
Column 6 7 14 –6*  Thus, in employing a mixed strategy, a player decides to mix his choices
maxima among alternatives in a certain ratio.

Example 2: Mixed strategies Solution to example 2

 The following is the pay-off matrix of a game being played by A and B.  The problem does not have a saddle point. If A plays 𝑎1 , then B
Determine the optimal strategies for the players and the value of the would play 𝑏2 while for 𝑎2 played by A, B would choose to play 𝑏1 .
game.
 So, if B knows what choice A will make then B can ensure that he
gains by choosing a strategy opposite to the one desired by A.
 Thus, it is of utmost importance for A to make it difficult for B to guess
as to what choice he is going to make.
 Similarly, B would like to make it very difficult for A to assess the
strategy he is likely to adopt.
 It would pay each one of them to play either of the respective
strategies open to each with certain probability.

Solution to example 2 Solution to example 2

 Suppose A plays strategy 𝑎1 with probability x and plays strategy with

probability 1 – x. If B plays strategy 𝑏1, then A’s expected pay-off can be  A would do best to adopt the strategies 𝑎1 and 𝑎2 , choosing in a random
manner, in the proportion 2:3 (i.e. 2/5 and 3/5).
determined in reference to the figures given in the first column of the pay-
off matrix as follows:  The expected pay-off for A using this mixed strategy equals

 Expected pay-off (given B plays 𝑏1) = 8x - 6 (1 – x)  8 (2/5) – 6(3/5) = –2/5

 –7 (2/5) + 4 (3/5) = –2/5

 Similarly, if B plays strategy 𝑏2 , the expected pay-off of A can be
determined as follows:  We can determine the mixed strategy for B in a similar manner as for A. Thus, if B
plays strategy 𝑏1 with probability y and strategy 𝑏2 with probability 1 – y, we
 Expected pay-off (given that B plays 𝑏2) = –7x + 4(1 – x)
have
 The value of x is determined in a way such that the expected pay-off for A
 Expected pay-off (given A plays 𝑎1 ) = 8y – 7(1 – y)
is the same irrespective of the strategy adopted by B. This value can be
obtained by equating these two equations.  And expected pay-off (given A plays 𝑎2 ) = –6y + 4(1 – y)
 Thus, 8x - 6(1 – x) = –7x + 4(1 – x) => x = 10/25 = 2/5  Or y = 11/25

3
 10-03-2025

Solution to example 2 General expression of a two-player

zero-sum game
 Thus, B should play strategies 𝑏1 and 𝑏2 in the ratio 11:14 in a random  In general, for a zero-sum two-persons game in which each of the
manner. players, say A and B, have strategies 𝑎1 and 𝑎2 and 𝑏1 and 𝑏2 ,
respectively, and the pay-offs are as given below, then, if x is the
 B’s gain in the long run will be opposite of A’s. Hence, B shall gain 2/5 per probability with which A chooses strategy 𝑎1 and if y is the
play in the long run. probability that B plays strategy 𝑏1 , we have,
 Thus, A and B should both use mixed strategies as given here and the value
of game equals –2/5.

General expression of a two-player Dominance Rule

zero-sum game
𝑎22 −𝑎21  In a game, sometimes a strategy available to a player might be found to be
 𝑥 = (𝑎 preferable to some other strategy/ strategies. Such a strategy is said to
11 +𝑎22)−(𝑎12 +𝑎21 ) dominate the other one(s). The concept of domination is very usefully
employed in simplifying the games and thus helps in finding solutions to the
𝑎22 −𝑎12 games.
 𝑦 = (𝑎
11 +𝑎22 )−(𝑎12 +𝑎21 )

𝑎11 𝑎22 −𝑎12 𝑎21

𝑉= (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )

 In the above pay-off matrix, every element of the second row exceeds the
corresponding element of the third row. Clearly, A shall never prefer to play 𝑎3 ,
because, in comparison to this strategy, it shall be better off by adopting 𝑎2 ,
regardless of what strategy is adopted by B. Thus, 𝑎3 is dominated by 𝑎2 and
hence, it can be deleted.

Dominance Rule Example 3

 So, the new pay-off matrix becomes:

 Find the optimal strategies for A and B in the following game. Also
obtain the value of the game.

 From the reduced matrix, we observe that the first column values are larger
than their counterparts in the third column. Since, B would like to minimize
the pay-offs to A, it would always prefer to choose 𝑏3 to 𝑏1. Thus, strategy 𝑏1
is dominated by strategy 𝑏3. Now, deleting 𝑏1, we have,

Further, deleting of 𝑎1 yields only 𝑎2 open to firm A. Since firm B loses more by choosing
𝑏2, it would adopt 𝑏3. Thus, the strategies are 𝑎2 and 𝑏3, and a value 3 as the solution to this
reduced 1 x 1 game.

4
 10-03-2025

Solution to example 3 Exercise 1

 Removing the dominated strategy (𝑎3), the reduced matrix becomes:  Two packaged food manufacturers ‘McFood’ and ‘BigBite’ are competing
for increasing market shares. The pay-off matrix shown in the following
table, describes the increase in market share for ‘Mcfood’ and decrease in
market share for ‘BigBite’. Determine the optimal strategies for both the
firms and also the value of the game.

 Again, 𝑏1 is dominated by 𝑏2. Thus, the revised matrix becomes

Solution to Exercise 1

 Reduced form of the game

 x = 0.5; y = 0.1; V = 1.5

 Thus, optimal strategy for McFood is (0. 0.5, 0.5, 0), for BigBite is (0.1, 0.9, 0.
0), and the value of the game V = 1.5