Chapter Three

Game Theory
 Introduction
• Game theory is concerned with the general analysis of
strategic actions.

• Used to analyze the strategic behavior of agents.

• For instance, recall that each firm in duopoly market

takes in to account it rival’s action (reaction) to make
decision on level of price and output that maximize
profit.

• Such strategic interactions between firms can be

studied using game theory.

2
 What is game?
• A game is a situation in which players (participants) make
decision by taking into account other’s actions and
responses.

• The outcome of the game depends on the decision of each

players.

• For example, Cournot duopolist decides on the level of

output by taking into account the reaction of its competitor
and hence the level of profit obtained depends on decision
made by the two firms.

• Such decision-making represents a game.

3
 Components of a game
• Three basic elements of a game:

1. Player: may be individuals, firms (as in oligopoly markets), entire

nation (as in military conflict) or organization.
• Players are assumed to be rational.

2. Strategy: represents each course of action open to a player in a

game.
• A strategy is said to be optimal if it maximizes the expected return
of the strategic choice.

3. Payoffs: the final return (outcome) to the players of a game to

chosen combination of strategy.
• Payoff matrix summarizes the payoff of all strategies.

4
 Types of game
• Games can be broadly classified as: cooperative
game and non-cooperative game.

• Cooperative game: agents can have a binding

contract that allows them to make joint decision.
E.g. Collusive oligopoly

• Non-cooperative game: a binding contract can’t

be arrived and enforced.
E.g. Cournot duopoly model.
5
 Dominant Strategy and Nash Equilibrium

• Consider two firms (firm A and B) taking choice between “to

advertize” and “not to advertize” their products.

• The pay off matrix shows the possible amount of profit for
each firm under each strategy.

• Example 1:
Firm B's strategies
Advertize Not advertize
strategies
Firm A’s

Advertize 1,2 0,1

Not advertize 2,1 1,0

6
Dominant strategy and Nash equilibrium

• Two players: A and B

• Two strategies: Advertize and not to advertize

• Payoff: the first number in each of the four cells

represents the payoff for player A and the second
number represents the payoffs for player B.

• For instance: if firm A chooses to advertize and also

firm B advertises then A will get 1 and B will get 2.
7
 Solving the game:
• If firm A ‘Add’ then B will also ‘Add’ because (2>1)
and If A chooses ‘Not to add’ then B will choose to
‘Add’ because (1>0).

• If firm B ‘Add’ then firm A chooses ‘Not to add’

because (2>1) and firm B chooses to ‘Not to add’ A
will choose ‘not to add’ because (1>0)

• Thus, whatever firm A chooses B’s best strategy is to

‘Add’ and whatever B chooses firm A’s best strategy
is’ not to add’.
8
 Dominant Strategy
• It is an optimal strategy of a player no matter the
strategy chosen by the other player.

• From the game given above: A’s dominant strategy is

‘Not to add’ while B’s dominant strategy is to ‘Add’.

• Whenever player has dominant strategy, we call the

outcome of the game an equilibrium in dominant
strategy (dominant strategy equilibrium).

• In the example above firms will be in equilibrium

when firm A chooses ‘Not to add’ and B chooses to
‘Add’.
9
 Dominant Strategy
• Example 2:
Firm B's strategies
Advertize Not advertize
Advertize 1,0 2,1
Firm A’s strategies
Not advertize 0,1 1,2

• Exercise: consider two firms (A and B) making a choice between

decreasing and increasing price.

Firm B's strategies

Decrease Increase
Firm A’s strategies Decrease 10,5 15,0
Increase 6,8 10,2

• Find the dominant strategy for each firm

• Find the dominant strategy equilibrium

10
 Nash Equilibrium
• Games may not always have dominant strategy for
each player (and hence no dominant strategy
equilibrium).

• In such cases, equilibrium can be reached by finding

the best strategic responses of a player to optimal
choice of the other firm.

• Suppose there are two firms selling competing

products and hence needs to decide whether to
‘Advertize’ or ‘not advertize’.

11
 Nash Equilibrium
Firm B strategies

Advertize Not advertize

Firm A’s strategies Advertize 2,1 0,0
Not advertize 0,0 1,2

• Both firm A and B do not have dominant strategy.

• A’s best strategy depends on whatever B is doing and vice versa.

• Thus, strategies (Add, Add) and (Not Add, Not Add) are the
equilibrium of the game.

• Such kind of equilibrium is called Nash equilibrium.

• A’s choice is optimal given B’s choice and B’s choice is optimal
given A’s choice
12
 Nash Equilibrium
• Neither of the players knows what the other person
will do when he/she has to make choice of strategy.

• But they have some expectation about what the other

player will choose.

• Nash Equilibrium is a generalization of cournot

equilibrium.

• Dominant strategy equilibrium is Nash equilibrium,

but all Nash equilibria are not a dominant strategy.
13
 Problems with Nash equilibrium
• Uniqueness: Some games may have more than one Nash
equilibrium.

• Existence: some games might not have Nash equilibrium.

• Consider the following game:

Player B
Head Tail
Player A Head 1,-1 -1,1
Tail -1,1 1,-1

• No Nash equilibrium
14
 Mixed Strategies
• Some games may not have Nash equilibrium in a pure strategy.

• However with mixed strategy all games (with finite strategies) will
have at least one Nash equilibrium.

• So far we focused on pure strategy where players choose their

strategy once and for all.

• But sometimes players may randomize their choice of action and

play these strategies which leads to mixed strategy.

• In mixed strategy players assign probability to each choice and

will play their choice according to these probabilities.

15
 Mixed Strategies
• Suppose there are two strategies available for players (firm
A and B) …. “Advertize” and “not to advertize”

• Then, firm A may choose to ‘Add’ its products 50 percent of

the time and chooses ‘Not to add’ 50 percent of the time.

• Also firm B may choose to ‘Add’ 25 percent of the time and

‘Not to add’ 75 percent of the time.

• Thus, Nash equilibrium in mixed strategy occurs when each

agent chooses the optimal frequency with which to play his
strategies given the frequency choice of the other agents.

16
 Mixed Strategies
Firm B strategies
Advertize Not advertize
(25%) (75%)
Firm A’s strategies Advertize (50%) 0,0 0,-1
Not advertize (50%) 1,0 -1,3

• Payoff for firm A

• If ‘Add’ 0.5 (0.25) (0) + 0.5 (0.75) (0) = 0
• If ‘Not add’ 0.5 (0.25)(1) + 0.5(0.75) (-1) = -0.25

• Payoff for firm B

• If ‘Add’ 0.25 (0.5) (0) + 0.25 (0.5) (0) = 0
• If ‘Not add’ 0.75 (0.5)(-1) + 0.75(0.5) (3) = 0.75

• Therefore; Firm A chooses to ‘Add’ (because 0 > -0.25) and B

chooses ‘Not to add’ (because 0.75 > 0).

17
 The Prisoner’s Dilemma
• Suppose there are two individuals who commits a crime together
and are being questioned (investigated) in a separate rooms.

• The prisoners’ may choose between confessing and denying the

crime.

• If each prisoner confesses, they will be in prison for 3 months

each.

• If both deny, they will be in prison for 1 month each.

• If one of them confesses, he will be free and the other will be in

prison for 6 months.

18
 The prisoner’s Dilemma
Prisoner B
Confess Deny
Prisoner A Confess -3,-3 0,-6
Deny -6,0 -1,-1

• Nash equilibrium is (confess, confess) which is also a

dominant strategy Equilibrium.

• Efficiency?
• Player would have been better off by denying the crime
together (-1,-1) ……. the outcome is Pareto inefficient

19
 Cartel Cheating
• Prisoners’ dilemma game is useful in analyzing the behavior of
oligopoly.

• It can be used to indicate the situation under which firms would be

tempted to cheat on a cartel agreement by cutting price from agreed
level or increasing the output produced more than its quota.

• Suppose that there are two firms’ production oil, firm- A and firm- B
form cartel. Each firm has two possible strategies: to stick with the
cartel agreement or to cheat.

• Firm B
Stick to agreement cheat
Stick to agreement (4,4) (1,5)
Firm A
cheat (5,1) (3,3)

20
 Repeated Game
• In the previous example, the prisoners played the game once (a one
shot game).

• As a result they end up with inefficient equilibrium (confess,

confess).

• In repeated game, each player has the opportunity to establish a

reputation for cooperation and there by encourage the other players
to do the same.

• The game could be repeated for:

o Definite, known period of time

o Indefinite number of times

21
 Case 1: Definite number of times
• If the game is repeated for a fixed number of time (known
to the players) say for 5 rounds, then the defect strategy is
the best strategy in all rounds.

• This is because in the last round there is no future play and

hence ‘defect’ is the best strategy.

• So why one should cooperate next to the last round and so

on.

• If we can’t enforce cooperation in the last round, there will

be no way to enforce cooperation in the next to the last
round and so on.

22
 Case 2: Indefinite number of times
• If the game is going to be repeated indefinite period of time, a player
can reward or punish his/her opponent in the next (future round).

• If your opponent defected on the previous round, you defect in the

next round (punish) and vice versa.

• This is called the tit-for-tat strategy.

• In this strategy whatever your opponent did the last round, you do
it on this round.

• The strategy is rewarding as well as punishing.

• Thus, when the game is played for indefinite number of time, the
win-win strategy is the best solution (deny, deny).

23
 Sequential Games
• Until now we have been thinking about games in
which both players acts simultaneously.

• But in many situations one player gets to more

first, and the other player responds.
• Example; Stackelberg’s model

• Suppose A get to choose top or bottom.

• Then B gets observe the first player’s choice and

then chooses left or right.
24
 Sequential Games
Player B
Left Right
Player A Top 1,9 1,9
Bottom 0,0 2,1

• Two Nash equilibriums; (top, left) and (bottom, right).

• However, the payoff matrix hides the fact that one player gets to
know what the other player has chosen before he makes his choice.

• If the game is represented in extensive form one of the equilibra is

not reasonable.

• Extensive form: a way to represent the game that shows the time
pattern of the choices.

25
 Sequential Games

Left 1,9

Top
Right 1,9

A
Left 0,0
Bottom
Right 2,1

• Extensive form of representing a game

26
 Sequential Games
• Solve the game using the back-ward induction method.

• If A chooses Top, it doesn’t matter what B does, the payoff is (1, 9).

• If A chooses Bottom, the sensible thing for B is to choose right and

the payoff is (2, 1).

• Now think about player A’s initial choice.

• If the chooses top, the outcome will be (1, 9) and he will get 1.

• If he chooses bottom then he will get 2. Thus the sensible thing for
him is to choose bottom

• So, (bottom, right) will be the Nash equilibrium.

27
 Sequential Games
• B’s response……..

• From B’s point of view this is rather unfortunate since he ends up

with a payoff 1 rather than 9. What might he do about it?

• He can threaten to play left (L) if player a plays bottom (B).

• If player A thought that B would actually carry out this threat, he

would be well advised to play top.

• Because top gives him1, while bottom- if B carries out his threat-
will only give him 0.

• In this case, however, the threat is not credible (or it is an empty

threat).

28
 Game of entry deterrence
• Suppose that we have a monopolist who is facing a
threat of entry by another firm.

• The entrant decides whether or not to come into the

market, and then the incumbent decides whether or not
to fight (cutting its price in response).
1,9
Fight
Stay out
Don’t fight 1,9
E
Enter Fight 0,0

Don’t fight 2,1

29
 Game of entry deterrence
• Equilibrium outcome is (Enter, Don’t fight)

• But what if the incumbent can purchase some extra

production capacity that will allow it to produce more
output at same marginal cost he currently has.

• Let us assume that if it purchases that extra capacity

and if he chooses to fight that it will make a profit of
2.

• Thus the game tree will be changed as follows:

30
 Game of entry deterrence

Fight 1,9

stay out
Don’t Fight 1,9
E
Enter Fight 0,2

Don’t Fight 2,1

31
 Game of entry deterrence
• The increased capacity makes the threat of fighting
credible.

• If the potential entrant come into the market, the

incumbent will get 2 if it fights and 1 if it doesn’t.

• Thus, the incumbent will choose to fight.

• The entrant will therefore get a payoff zero if it enters

and 1 if it stays out.

• Therefore, it will stay out.

32