Normal Form

The Normal Form
Abhinash Borah
February 1, 2023
Road Map
Non-cooperative game theory has two building blocks

Normal or strategic form games
Extensive form games
We will begin by defining the normal form. Then as its first and most
obvious application we will use it to analyze strategic problems with
complete information occuring in a static environment, often referred to
as static games of complete information
Please distinguish between the terms “normal form games” and “static
games of complete information”
Abhinash Borah Normal Form February 1, 2023 2 / 73

Road Map
A 2 × 2 classification of strategic environments:
Is the environment static or dynamic?
Is the environment one of complete or incomplete information?
Four classes of strategic environments:
1 static games of complete information
2 dynamic games of complete information
3 static games of incomplete information
4 dynamic games of incomplete information

Normal Form Games
Normal Form or Strategic Game
First defined by von Neumann and Morgenstern in their classic book Games
and Economic Behavior (1944). But the roots of the concept can be traced
back to von Neumann’s 1928 paper: “Zur Theories der Gesellschaftsspiele”

Normal Form Games
Normal Form or Strategic Game
First defined by von Neumann and Morgenstern in their classic book Games
and Economic Behavior (1944). But the roots of the concept can be traced
back to von Neumann’s 1928 paper: “Zur Theories der Gesellschaftsspiele”
A normal-form or strategic game consists of 3 elements:
1 A set of players
2 A set of strategies for each player
3 A payoff or utility function for each player which specifies what that
player’s payoff is for any profile of strategies adopted by the players
A strategy for a player is to be understood as a complete plan of action

Normal Form Games
Example: Cournot Duopoly

2 firms have to simultaneously decide how much quantity of a homogenous
product to produce in a market with demand function P = 100 − Q. Both
firms have identical cost of production given by c(q) = q2 .
Can you represent this strategic problem as a normal form game?

Normal Form Games


1 Who are the players?

Normal Form Games


Firms 1, 2

Normal Form Games


Firms 1, 2
2 What is the set of strategies for each player?

Normal Form Games


Firms 1, 2
For each firm any non-negative level of quantity qi , i = 1, 2

Normal Form Games


Firms 1, 2
3 What is the payoff function for each player?

Normal Form Games


Firms 1, 2
Each firm’s profit
v1 (q1 , q2 ) =

Normal Form Games


Firms 1, 2
Each firm’s profit
v1 (q1 , q2 ) = (100 − (q1 + q2 ))q1 − q21

v2 (q1 , q2 ) = (100 − (q1 + q2 ))q2 − q22
Normal Form Games
Example: Stackelberg Duopoly

Firm 1 decides first and then firm 2 decides (after observing 1’s choice) how
much quantity of a homogenous product to produce in a market with demand
function P = 100 − Q. Both firms have identical cost of production: c(q) = q2 .
In terms of our four way classification of strategic environments, which

category does this strategic problem fall under?

Normal Form Games


Firms 1, 2

Normal Form Games


Firms 1, 2
2 What is the set of strategies for each player? (if you can answer this, you
would have figured out von Neumann’s first great contribution to GT)

Normal Form Games


Firms 1, 2
2 What is the set of strategies for each player? (if you can answer this, you
would have figured out von Neumann’s first great contribution to GT)

Normal Form Games

What are players’ payoffs for any strategy profile (q1 , q2 :R+ →R+ )?
Each firm’s objective is to maximize profits given by
vi (qi , qj ) = (100 − (qi + qj ))qi − q2i , i, j = 1, 2.
Accordingly firms’ payoffs as a function of any strategy profile (q1 ,
q2 :R+ →R+ ) is given by:
U1 (q1 , q2 : R+ → R+ ) = [100 − (q1 + q2 (q1 ))]q1 − q21
= v1 (q1 , q2 (q1 ))
U2 (q1 , q2 : R+ → R+ ) = [100 − (q1 + q2 (q1 ))]q2 (q1 ) − [q2 (q1 )]2
= v2 (q1 , q2 (q1 ))
Normal Form Games
Normal-Form or Strategic Game
A little more formally . . .
DEFINITION
A normal-form or strategic game consists of 3 elements:
1 A set of players involved in the strategic situation, which we denote by

N = {1, ..., n}
2 A set of pure strategies for each player i ∈ N, which we denote by Si
3 A payoff function for each player i ∈ N,

vi : S := S1 × S2 × . . . × Sn → R,
which, for any strategy profile s := (s1 , . . . , sn ) ∈ S, gives player i’s
payoff or utility, vi (s). (More formally, vi represents i’s preferences)

Normal Form Games
Normal-Form Game: Interpretation
The requirement that the payoff of each player i,

vi : S = S1 × S2 × . . . × Sn → R,
be defined over S, rather than Si , is the feature that distinguishes a
normal-form game from a decision problem
Because of the strategic interactions involved, each player cares about

not just her own strategy but also about those of others.

Normal Form Games
Normal-Form Game: Interpretation
Strategic Independence
Players choose their strategies “simultaneously” and independently

Each player is unaware, when choosing her strategy, of the strategy
choices being made by the other players
There is no information, except for the details of the game, on which a

player can base her expectation of the other players’ strategies

Normal Form Games
Finite Games
DEFINITION
A normal form game is called finite if each player has a finite number of
strategies

Normal Form Games
Matrix representation of finite 2 player games
Prisoner’s Dilemma: Two suspects charged with a crime are questioned

in separate cells. If they both confess, each will be sentenced to 3 years
in prison. If only one of them confesses, he will be freed and used as a
witness against the other, who will receive a sentence of 4 years. If
neither confess both will be charged of a minor crime and sentenced to 1
year in prison. Assume each person’s payoff from spending x years in
prison is 4 − x

Normal Form Games

prison is 4 − x
Don’t Confess (N) Confess (Y)
Don’t Confess (N) 3, 3 0, 4
Confess (Y) 4, 0 1, 1
What is the set of efficient outcomes?

Normal Form Games
Battle of the Sexes/Coordination Game (Luce and Raiffa, 1957): Bob

and Ann want to see each other but have no way of communicating.
Each can independently decide to go to a soccer game or a concert. Both
would rather spend the evening together than apart, but Bob would rather
they be together at the soccer game while Ann would rather they be
together at the concert

Normal Form Games
Battle of the Sexes/Coordination Game (Luce and Raiffa, 1957): Bob

and Ann want to see each other but have no way of communicating.
Each can independently decide to go to a soccer game or a concert. Both
would rather spend the evening together than apart, but Bob would rather
they be together at the soccer game while Ann would rather they be
together at the concert
Soccer Concert
Soccer 1, 2 0, 0
Concert 0, 0 2, 1

Normal Form Games
Provisioning a Public Good: There are 3 individuals who can either

vote in favor (F) or against (A) provisioning a public good, which costs 3
and provides each individual a benefit of 2. The public good is
provisioned if at least one of the individuals vote F. Individuals who vote
A pay 0 whereas individuals who vote F share the cost with all other
individuals who vote F. Each individual’s payoff is the difference
between her benefit and cost.
Is it efficient to provision the public good?

Normal Form Games
Provisioning a Public Good

Favor (F) Against (A)
Favor (F) 1, 1, 1 .5, 2, .5
Against (A) 2, .5, .5 2, 2, −1

Player 3 chooses Favor (F)
Favor (F) .5, .5, 2 −1, 2, 2
Against (A) 2, −1, 2 0, 0, 0

Player 3 chooses Against (A)
Normal Form Games
Tree representation of dynamic strategic environments
It is often convenient to represent dynamic strategic environments by means

of trees.
Example (BoS revisited): Suppose Ann has the opportunity to first

choose whether to go to the concert or soccer and then inform Bob of her
choice. Thereafter, Bob makes his choice between soccer and concert.
Represent this strategic environment as a normal form game.

Normal Form Games

!
! "
"
+(*
# $
!
)(*
% &
'() *(+

Normal Form Games

!
! "
"
+(*
# $
!
)(*
% &
'() *(+
Set of player 1’ pure strategies: S1 = {(A, E), (A, F), (B, E), (B, F)}
Set of player 2’s strategies: S2 = {C, D}

Normal Form Games
Example (Allocation procedure): Two individuals, 1 and 2, use the

following procedure to share two units of a desirable indivisible good.
Individual 1 first proposes how many units, 0, 1 or 2, person 2 gets. Then
person 2, after seeing 1’s proposal, either accepts (A) or rejects (R) the
proposal. In the event of rejection, both individuals receive zero units.
Each individual cares only about the number of objects she obtains.
Represent this strategic environment as a normal form game.

Normal Form Games
Mixed Strategy
A player may choose to randomize over her pure strategies

Think of a goalkeeper during a penalty shootout. She may randomize
between diving left, diving right or not diving
Correspondingly, the beliefs that a player’s opponents hold about her

strategy need not be deterministic but can be probabilistic
A mixed strategy of a player is a probability distribution over her pure

strategies
Critical assumption: Each player’s randomization under a mixed strategy

is independent of those of her opponents

Normal Form Games
Example: Mixed Strategy
L M R
U 4, 3 5, 1 6, 2
M 2, 1 8, 4 3, 6
D 3, 0 9, 6 2, 8
Example

1 1
σ1 = (σ1 (U), σ1 (M), σ1 (D)) = 0, ,
2 2

1 1 1
σ2 = (σ2 (L), σ2 (M), σ2 (R)) = , ,
3 3 3

Normal Form Games
Example: Calculating Payoffs under Mixed Strategies
L ( 13 ) M ( 13 ) R ( 31 )
U (0) 4, 3 5, 1 6, 2
M ( 12 ) 2, 1 8, 4 3, 6
D ( 12 ) 3, 0 9, 6 2, 8
Calculate player 1 and 2’s payoffs under the mixed strategies:

1 1
σ1 = (σ1 (U), σ1 (M), σ1 (D)) = 0, ,
2 2

1 1 1
σ2 = (σ2 (L), σ2 (M), σ2 (R)) = , ,
3 3 3
Normal Form Games
Example: Calculating Payoffs under Mixed Strategies
L ( 31 ) M ( 13 ) R ( 13 )
U (0) 4, 3 5, 1 6, 2
(0. 13 = 0) (0. 31 = 0) (0. 13 = 0)
M ( 21 ) 2, 1 8, 4 3, 6
( 12 . 31= 6 ) ( 2 . 3 = 6 ) ( 2 . 3 = 61 )
1 1 1 1 1 1
D ( 12 ) 3, 0 9, 6 2, 8
( 12 . 31 = 61 ) ( 12 . 31 = 16 ) ( 12 . 13 = 61 )
1’s expected payoff = 16 (2 + 8 + 3 + 3 + 9 + 2) = 27

6
2’s expected payoff = 16 (1 + 4 + 6 + 0 + 6 + 8) = 25

6
Normal Form Games
Analyzing games
So far we have developed a language to map strategic environments into

the formal structure of normal form games
We have not said anything about what the outcomes of such strategic
interactions modeled using the normal form may be. That is the question
we turn to next
Of course, to systematically answer this question, we will have to make

assumptions, some of which may be quite demanding

Normal Form Games
ASSUMPTION: Rationality
ASSUMPTION
Each player chooses her strategy si ∈ Si to maximize her payoff consistent
with her beliefs about what is going on in the game

Dominance
Example
Prisoner’s Dilemma
Confess (Y) 4, 0 1, 1

Dominance
Dominance and Dominant Strategy Equilibrium
A strategy s0i of player i is strictly dominated by another strategy si if

for any possible combination of other players’ strategies, player i’s
payoff from si is strictly greater than that from s0i
si ∈ Si is a strictly dominant strategy for i if every other strategy of i is

strictly dominated by it

Dominance
Notation
We often denote a strategy profile s = (s1 , . . . , sn ) ∈ S by s = (si , s−i ),

where:
s−i = (s1 , ..., si−1 , si+1 , ..., sn ) ∈ S−i = S1 × ... × Si−1 × Si+1 × ... × Sn

Dominance
DEFINITION
Strategy s0i ∈ Si is strictly dominated by another strategy si ∈ Si if
vi (si , s−i ) > vi (s0i , s−i ), for all s−i ∈ S−i
DEFINITION
si ∈ Si is a strictly dominant strategy for i if
vi (si , s−i ) > vi (s0i , s−i ), for all s0i ∈ Si , s0i 6= si , and all s−i ∈ S−i

Dominance
DEFINITION
The strategy profile s∗ = (s∗1 , ..., s∗n ) ∈ S is a strictly dominant strategy
equilibrium if s∗i ∈ Si is a strictly dominant strategy for i, for all i ∈ N.
Example: (confess, confess) is a strictly dominant strategy equilibrium in the

Prisoner’s Dilemma game
PROPOSITION
If a strictly dominant strategy equilibrium exists in any normal-form game,
then it is unique.

Dominance
Does the battle of sexes game have a strictly dominant strategy

equilibrium?
Soccer Concert
Soccer 2, 1 0, 0
Concert 0, 0 1, 2

Dominance
Does the Cournot duopoly game with market demand given by

P = 100 − Q and each firm’s cost function given by c(q) = q2 have a
strictly dominant strategy equilibrium?

Dominance
Weakly dominant strategy
A strategy s0i of player i is weakly dominated by another strategy si if for any

possible combination of other players’ strategies, player i’s payoff from si is
no less than that from s0i and for some combination of others’ strategies, it is
strictly greater.
DEFINITION
Strategy s0i ∈ Si is weakly dominated by another strategy si ∈ Si if
vi (si , s−i ) ≥ vi (s0i , s−i ), for all s−i ∈ S−i ; and

vi (si , s−i ) > vi (s0i , s−i ), for some s−i ∈ S−i

Dominance
An observation
Note that a strategy of a player may be dominated by a mixed strategy

even though it is not dominated by any of her pure strategies. We do not
draw on mixed strategy based dominance arguments too often in this
course, but this is something you should be aware of

Dominance
ASSUMPTION: Intelligence
ASSUMPTION
All players know everything about the game: the actions, the outcomes and
the preferences of all the players

Dominance
ASSUMPTION: Common Knowledge
DEFINITION
An event is common knowledge if everyone knows it; everyone knows that
everyone knows it; everyone knows that everyone knows that everyone knows
it; and so on ad infinitum
ASSUMPTION
The fact that players are rational and intelligent is common knowledge among
all the players of the game

Dominance
Iterated Elimination of Strictly Dominated Strategies
The scope of rationality and common knowledge of rationality
Iterated Elimination of Strictly Dominated Strategies (IESDS)

Dominance
Example: IESDS
L M R
U 4, 3 5, 1 6, 2
M 2, 1 8, 4 3, 6
D 3, 0 9, 6 2, 8

Dominance
Example: IESDS
L M R
U 4, 3 5, 1 6, 2
M 2, 1 8, 4 3, 6
D 3, 0 9, 6 2, 8
Player 2 is rational and will never play M

Player 1 knows that player 2 is rational and will never play M. So 1 will
never play M or D
Player 2 knows that player 1 knows that she (player 2) is rational and
will never play M; so she knows that player 1 will never play M or D,
which leaves player 1 with U. So player 2 will play L. Therefore, (U, L)
is the unique strategy profile that survives IESDS
Dominance
Example: IESDS in pick the integer game
What is the set of strategy profiles that survive IESDS?

Dominance
Example: IESDS in Cournot Duopoly

What is the set of strategy profiles that survive IESDS?

Dominance
Example: IESDS and Battle of Sexes
What is the set of strategy profiles that survive IESDS in the battle of
sexes game?
Soccer Concert
Soccer 2, 1 0, 0
Concert 0, 0 1, 2

Dominance
IESDS: Comments
The set of strategy profiles that survive IESDS in any normal-form game
is non-empty but need not be singleton
If for a normal form game, s∗ ∈ S is a strictly dominant strategy

equilibrium, then s∗ uniquely survives IESDS

Dominance
Best Response
DEFINITION
The strategy si ∈ Si is player i’s best response to her opponents’ strategies
s−i ∈ S−i if
vi (si , s−i ) ≥ vi (s0i , s−i ), for all s0i ∈ Si
DEFINITION
The best response correspondence of player i selects for each s−i ∈ S−i , a
subset Bi (s−i ) ⊆ Si consisting of all the strategies si ∈ Si that are a best
response to s−i .

Dominance
Example: Best Response Correspondence
What is player 1’s best response correspondence in the following normal-form

game?
L M R
U 3, 3 5, 1 6, 2
M 4, 1 8, 4 3, 6
D 4, 0 9, 6 6, 8

Dominance
Example: Best Response Correspondence
Consider the Cournot duopoly game with market demand given by

P = 100 − Q and each firm’s cost function given by c(q) = q2 . What is
player i’s best response correspondence, i = 1, 2?

Dominance
Nash Equilibrium: Motivation
Consider once again the battle of sexes game:
Soccer Concert
Soccer 2, 1 0, 0
Concert 0, 0 1, 2
We saw that all four strategy profiles survive IESDS. Consider the outcome
(soccer, concert). What beliefs on the part of players make this outcome
consistent with rationality?

Dominance
Consider once again the battle of sexes game:
Soccer Concert
Soccer 2, 1 0, 0
Concert 0, 0 1, 2
We saw that all four strategy profiles survive IESDS. Consider the outcome
(soccer, concert). What beliefs on the part of players make this outcome
consistent with rationality? Player 1 must believe that player 2 is going to the
soccer game; and 2 must believe that 1 is going for the concert. However,
there is something troubling about the reasoning that sustains this outcome.
The players are taking actions that are best responses to their beliefs but their
beliefs about each other are wrong!

Dominance
The concept of a Nash equilibrium ties together beliefs and actions. It is a

system of beliefs and a profile of actions under which:
1 Each player is playing a best response to her beliefs about her opponents’
strategies
2 The beliefs of the players about their opponents are correct
IMPORTANT: When we formally specify a definition of a Nash equilibrium,

the fact that the definition subsumes both these ideas is sometimes lost on the
reader, especially to those who are seeing this concept for the first time. It is
critical that you do not loose track of this point when we introduce this
concept formally, which we do next

Dominance
Pure Strategy Nash Equilibrium (PSNE): Definition
DEFINITION
A pure strategy Nash equilibrium (PSNE) of a normal-form game is a strategy
profile s∗ = (s∗1 , . . . , s∗n ) ∈ S with the property that for every player i, we have
vi (s∗i , s∗−i ) ≥ vi (si , s∗−i ) for all si ∈ Si

Dominance
An Alternative Formulation of PSNE
In a Nash equilibrium strategy profile, players are playing mutual best

responses. We may re-write the definition in a manner that highlights this
point
DEFINITION
A PSNE of a normal-form game is a strategy profile s∗ = (s∗1 , . . . , s∗n ) ∈ S
with the property that s∗i ∈ Bi (s∗−i ) for all players i
Because players are playing mutual best responses under a Nash equilibrium,
it may also be thought of as a strategy profile from which no player has a
unilateral incentive to deviate.

Dominance
Nash Equilibrium: Interpretation
The notion of a Nash equilibrium captures a steady state of the play of a

normal-form game in which each player holds the correct belief about
the other player’s behavior and acts rationally
This steady state is precisely the one from which no player has an
incentive to unilaterally deviate
The notion does not attempt to examine the process by which a steady
state is reached
A Nash equilibrium may be interpreted as a self-enforcing agreement
A Nash equilibrium may be interpreted as a stable social convention

Dominance
Example: PSNE in Cournot Duopoly

product to produce in a market with demand function P = 100 − Q, when
each firm has cost of production given by c(q) = q2 . Identify the set of PSNE

Dominance
Example: PSNE in Cournot Duopoly

product to produce in a market with demand function P = 100 − Q, when
each firm has cost of production given by c(q) = q2 . Identify the set of PSNE
What is Bi (qj ) (i, j = 1, 2)?
Solve: maxqi (100 − (qi + qj ))qi − q2i
First Order Condition

qj
qi = 25 − 4
Best response correspondences

q2 q1
q1 = 25 − 4 and q2 = 25 − 4
Nash Equilibrium
q∗1 = q∗2 = 20
Dominance
Example: PSNE in Prisoner’s Dilemma

prison is 4 − x

Dominance

prison is 4 − x
Confess (Y) 4, 0 1, 1

Dominance

prison is 4 − x
Confess (Y) 4, 0 1, 1

Dominance
Example—PSNE in Finite 3 Player Game
Provisioning a Public Good: There are 3 individuals who can either

vote in favor (F) or against (A) provisioning a public good, which costs
$3 and provides each individual a benefit of $2. The public good is
provisioned if at least one of the individuals vote F. Individuals who vote
A pay $0 whereas individuals who vote F share the cost with all other
individuals who vote F. Each individual’s payoff is the difference
between her benefit and cost.
Is it efficient to provision the public good?

Dominance

Favor (F) 1, 1, 1 .5, 2, .5
Against (A) 2, .5, .5 2, 2, −1

Favor (F) .5, .5, 2 −1, 2, 2
Against (A) 2, −1, 2 0, 0, 0

Dominance

Favor (F) 1, 1, 1 .5, 2, .5
Against (A) 2, .5, .5 2, 2, −1

Favor (F) .5, .5, 2 −1, 2, 2
Against (A) 2, −1, 2 0, 0, 0

Dominance
Example: PSNE in Battle of Sexes—Multiple PSNE
Coordination Game/ “Battle of the Sexes” (Luce and Raiffa 1957)
Soccer Concert
Soccer 2, 1 0, 0
Concert 0, 0 1, 2

Dominance
Example: PSNE in Battle of Sexes—Multiple PSNE
Coordination Game/ “Battle of the Sexes” (Luce and Raiffa 1957)
Soccer Concert
Soccer 2, 1 0, 0
Concert 0, 0 1, 2

Dominance
Example: PSNE in Matching Pennies–No PSNE
Matching Pennies
Head Tail
Head 1, −1 −1, 1
Tail −1, 1 1, −1

Dominance
Example: PSNE in Matching Pennies—No PSNE
Matching Pennies
Head Tail
Head 1, −1 −1, 1
Tail −1, 1 1, −1

Dominance
Example: PSNE in Tragedy of Commons (Hardin)
There are n countries, each of which creates a certain amount of air pollution
in the course of undertaking economic activity. In particular, if country i
produces ki units of air pollution, then its GDP is ln(ki ). The stock of clean air
in the world is a public good and there is a total amount of clean air equal to
Pn each country i = 1, ..., n has
K. The total amount of clean air left after
produced ki units of pollution is K − i=1 ki . Apart from its GDP, each
country also cares about the amount of clean air as it affects the quality of life
of its citizens. Country i’s payoff function is given by:
n
X
vi (ki , k−i ) = ln(ki ) + ln(K − kj )
j=1
Model this strategic situation as a normal-form game and identify its set of
PSNE. Comment on their efficiency.

Dominance
Example: PSNE in Bertrand Duopoly
There are two profit maximizing firms producing a homogenous product in a

market with demand function P = 100 − Q. Each firm has a cost of
production given by c(q) = 10q. The two firms have to simultaneously decide
what price to set for the product. The firm setting the lower price will get to
service the entire demand in the market prevailing at that price. In case the
two prices are the same, the market will split equally between the two.
Model this strategic situation as a normal form game and identify its set of
PSNE. Compare the outcome with that under the corresponding Cournot
duopoly.

Dominance
Example: PSNE in a Model of Electoral Competition
A unit mass of citizens is uniformly distributed on the interval [−0.5, 0.5]. A

citizen’s position on this interval reflects her political ideology with the
citizen positioned at −0.5 (respectively, 0.5) being the most left-leaning
(respectively, right-leaning). There are two politicians who have to decide
where to position themselves on the ideological interval. A citizen votes for
the politician who is closest to her position. There is no cost associated with
voting and so we assume that all citizens vote. The outcome of the election is
determined by majority rule. The politicians prefer winning the election to a
tie and prefer a tie to losing.
Model this strategic situation as a normal form game and identify its set of
PSNE.

Dominance
Mixed strategy Nash equilibrium
We now develop a notion of a Nash equilibrium in mixed strategies
In stating the definition, we restrict attention to two player games. The

definition can easily be extended to n player games, for n > 2

Dominance
Calculating Payoffs under Mixed Strategies
Let Σi denote the set of all mixed strategies of player i = 1, 2
Player i’s payoff, i = 1, 2, under mixed strategy profile σ = (σ1 , σ2 ) is

given by:
X
Evi (σ1 , σ2 ) = (σ1 (s1 ) × σ2 (s2 ))vi (s1 , s2 )
(s1 ,s2 )∈S

Dominance
Mixed Strategy Nash Equilibrium (MSNE)—Definition
DEFINITION
A mixed strategy Nash equilibrium (MSNE) of a two-player normal form
game is a profile (σ1∗ , σ2∗ ) with the property that:
Ev1 (σ1∗ , σ2∗ ) ≥ Ev1 (σ1 , σ2∗ ), for all σ1 ∈ Σ1

Ev2 (σ1∗ , σ2∗ ) ≥ Ev2 (σ1∗ , σ2 ), for all σ2 ∈ Σ2
Comment: Degenerate MSNE is a PSNE

Dominance
Mixed Strategy Nash Equilibrium—Important Property
PROPOSITION
(σ1∗ , σ2∗ ) is a mixed strategy Nash equilibrium of a strategic game if and only
if for each player i = 1, 2, every si ∈ Si for which σi∗ (si ) > 0 is a best
response to σ−i ∗

Dominance
Example: MSNE in Matching Pennies
Heads (H) Tails (T)
Heads (H) 1, −1 −1, 1
Tails (T) −1, 1 1, −1

What are the MSNE of this game?
Consider mixed strategies (σ1 (H), σ1 (T)) = (p, 1 − p) and (σ2 (H), σ2 (T)) =
(q, 1 − q). If these are to be a (non-degenerate) MSNE, then
q.1 + (1 − q)(−1) = q(−1) + (1 − q).1, and

p(−1) + (1 − p).1 = p.1 + (1 − p)(−1)
Solving these equations simultaneously give p = 1

2 and q = 12 . Hence σ1∗ =
( 12 , 21 ) and σ2∗ = ( 12 , 12 ) is a MSNE.
Dominance
Example: MSNE in Battle of Sexes
Soccer (S) Concert (C)
Soccer (S) 2, 1 0, 0
Concert (C) 0, 0 1, 2
Are there any other MSNE beyond the two degenerate ones associated with
the two PSNE?
If (σ1 (S), σ1 (C)) = (p, 1 − p) and (σ2 (S), σ2 (C)) = (q, 1 − q) are to be a
(non-degenerate) MSNE, then
q.2 + (1 − q).0 = q.0 + (1 − q).1, and

p.1 + (1 − p).0 = p.0 + (1 − p).2
That is, p = 2
3 and q = 13 . Hence, σ1∗ = ( 32 , 13 ) and σ2∗ = ( 13 , 23 ) is a MSNE.
Dominance
Example: MSNE in Rock-Paper-Scissors
Rock (R) Paper (P) Scissor (S)
Rock (R) 0, 0 −1, 1 1, −1
Paper (P) 1, −1 0, 0 −1, 1
Scissor (S) −1, 1 1, −1 0, 0
Claim: In any MSNE, both players must put positive probability on all three
strategies

Dominance
Rock (R) 0, 0 −1, 1 1, −1
Paper (P) 1, −1 0, 0 −1, 1
Scissor (S) −1, 1 1, −1 0, 0
Therefore, (σ1 , σ2 ) = ((σ1 (R), σ1 (P), σ1 (S)), (σ2 (R), σ2 (P), σ2 (S))) is a
MSNE iff the payoff of player i = 1, 2 from each of her three pure strategies
against σj , j 6= i, are equal.
Evi (R, σj ) = σj (P).(−1) + σj (S).1
Evi (P, σj ) = σj (R).1 + σj (S).(−1)
Evi (S, σj ) = σj (R).(−1) + σj (P).1
Dominance
Rock (R) 0, 0 −1, 1 1, −1
Paper (P) 1, −1 0, 0 −1, 1
Scissor (S) −1, 1 1, −1 0, 0
Solving these equations establish that (σ1∗ , σ2∗ ) = (( 31 , 31 , 31 ), ( 13 , 13 , 13 )) is the

unique MSNE.

Dominance
Mixed Strategy Nash Equilibrium: Further Interpretation
Mixed strategies as objects of choice
Each player in equilibrium finds it optimal to commit to choosing an

action (pure strategy) based on the realization of a private randomization
device
After all players have so committed, the devices are operated and each
player’s action is chosen based on the realization of the device
No player has an incentive to renege on the commitment—Why?

Dominance
Mixed Strategy Nash Equilibrium—Existence
THEOREM
Nash Existence Result: Every finite normal-form game has a mixed strategy
Nash equilibrium
In 1994, John Nash, along with John Harsanyi and Reinhard Selten, were
awarded the Nobel prize in Economics “for their pioneering analysis of
equilibria in the theory of non-cooperative games.” Please look at:
http://www.nobelprize.org/nobel_prizes/
economic-sciences/laureates/1994/press.html

Normal Form

Uploaded by

Copyright:

Available Formats

Normal Form

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Normal Form

Uploaded by

Copyright:

Available Formats

The Normal Form

Non-cooperative game theory has two building blocks

Abhinash Borah Normal Form February 1, 2023 2 / 73

A 2 × 2 classification of strategic environments:

Is the environment static or dynamic?

Is the environment one of complete or incomplete information?

Four classes of strategic environments:

1 static games of complete information

2 dynamic games of complete information

3 static games of incomplete information

4 dynamic games of incomplete information

Abhinash Borah Normal Form February 1, 2023 3 / 73

Normal Form or Strategic Game

Abhinash Borah Normal Form February 1, 2023 4 / 73

Normal Form or Strategic Game

A normal-form or strategic game consists of 3 elements:

2 A set of strategies for each player

A strategy for a player is to be understood as a complete plan of action

Abhinash Borah Normal Form February 1, 2023 4 / 73

Example: Cournot Duopoly

Can you represent this strategic problem as a normal form game?

Abhinash Borah Normal Form February 1, 2023 5 / 73

Example: Cournot Duopoly

Can you represent this strategic problem as a normal form game?

Abhinash Borah Normal Form February 1, 2023 5 / 73

Example: Cournot Duopoly

Can you represent this strategic problem as a normal form game?

Abhinash Borah Normal Form February 1, 2023 5 / 73

Example: Cournot Duopoly

Can you represent this strategic problem as a normal form game?

Abhinash Borah Normal Form February 1, 2023 5 / 73

Example: Cournot Duopoly

Can you represent this strategic problem as a normal form game?

Abhinash Borah Normal Form February 1, 2023 5 / 73

Example: Cournot Duopoly

Can you represent this strategic problem as a normal form game?

Abhinash Borah Normal Form February 1, 2023 5 / 73

Example: Cournot Duopoly

Can you represent this strategic problem as a normal form game?

Abhinash Borah Normal Form February 1, 2023 5 / 73

Example: Cournot Duopoly

Can you represent this strategic problem as a normal form game?

v1 (q1 , q2 ) = (100 − (q1 + q2 ))q1 − q21

Example: Stackelberg Duopoly

In terms of our four way classification of strategic environments, which

Abhinash Borah Normal Form February 1, 2023 6 / 73

Example: Stackelberg Duopoly

In terms of our four way classification of strategic environments, which

Abhinash Borah Normal Form February 1, 2023 6 / 73

Example: Stackelberg Duopoly

In terms of our four way classification of strategic environments, which

Abhinash Borah Normal Form February 1, 2023 6 / 73

Example: Stackelberg Duopoly

In terms of our four way classification of strategic environments, which

3 What is the payoff function for each player?

Example: Stackelberg Duopoly

Normal-Form or Strategic Game

A little more formally . . .

1 A set of players involved in the strategic situation, which we denote by