Mixed Strategies

ECON2112

Mixed Strategies
Introduction

Example

heads
tails

heads
1, 1
1, 1

tails

1, 1
1, 1

This game does not have a Nash equilibrium in pure strategies.

Both players randomizing and playing 12 heads + 12 tails is a Nash

equilibrium.

Before that, we need to assume that players have preferences

defined over the set (S ) that are representable by means of a
von-Newman Morgenstern utility function.

Mixed Strategies
Introduction

Example

heads
tails

heads
1, 1
1, 1

tails

1, 1
1, 1

This game does not have a Nash equilibrium in pure strategies.

Both players randomizing and playing 12 heads + 12 tails is a Nash

equilibrium.

Before that, we need to assume that players have preferences

defined over the set (S ) that are representable by means of a
von-Newman Morgenstern utility function.

Mixed Strategies
Introduction

Example

heads
tails

heads
1, 1
1, 1

tails

1, 1
1, 1

This game does not have a Nash equilibrium in pure strategies.

Both players randomizing and playing 12 heads + 12 tails is a Nash

equilibrium.

Before that, we need to assume that players have preferences

defined over the set (S ) that are representable by means of a
von-Newman Morgenstern utility function.

Mixed Strategies
Introduction

General Setting. We will have:

A set of players N = {1, . . . , n},

For each i N a finite set of pure strategies Si = {si1 , . . . , si i }.

The set of pure strategy profiles is S = S1 Sn .

Players have preferences defined (S ). (Not simply over S as in

the previous lecture.) We need that because players are now
allowed to randomize between pure strategies.

We will assume that such preferences are representable by

means of a von-Newman Morgenstern utility function.

That means that we can represent preferences over (S )

assigning numerical values to each pure strategy profile.

Mixed Strategies
Introduction

General Setting. We will have:

A set of players N = {1, . . . , n},

For each i N a finite set of pure strategies Si = {si1 , . . . , si i }.

The set of pure strategy profiles is S = S1 Sn .

Players have preferences defined (S ). (Not simply over S as in

the previous lecture.) We need that because players are now
allowed to randomize between pure strategies.

We will assume that such preferences are representable by

means of a von-Newman Morgenstern utility function.

That means that we can represent preferences over (S )

assigning numerical values to each pure strategy profile.

Mixed Strategies
Introduction

General Setting. We will have:

A set of players N = {1, . . . , n},

For each i N a finite set of pure strategies Si = {si1 , . . . , si i }.

The set of pure strategy profiles is S = S1 Sn .

Players have preferences defined (S ). (Not simply over S as in

the previous lecture.) We need that because players are now
allowed to randomize between pure strategies.

We will assume that such preferences are representable by

means of a von-Newman Morgenstern utility function.

That means that we can represent preferences over (S )

assigning numerical values to each pure strategy profile.

Mixed Strategies
Introduction

General Setting. We will have:

A set of players N = {1, . . . , n},

For each i N a finite set of pure strategies Si = {si1 , . . . , si i }.

The set of pure strategy profiles is S = S1 Sn .

Players have preferences defined (S ). (Not simply over S as in

the previous lecture.) We need that because players are now
allowed to randomize between pure strategies.

We will assume that such preferences are representable by

means of a von-Newman Morgenstern utility function.

That means that we can represent preferences over (S )

assigning numerical values to each pure strategy profile.

Mixed Strategies
Introduction

General Setting. We will have:

A set of players N = {1, . . . , n},

For each i N a finite set of pure strategies Si = {si1 , . . . , si i }.

The set of pure strategy profiles is S = S1 Sn .

Players have preferences defined (S ). (Not simply over S as in

the previous lecture.) We need that because players are now
allowed to randomize between pure strategies.

We will assume that such preferences are representable by

means of a von-Newman Morgenstern utility function.

That means that we can represent preferences over (S )

assigning numerical values to each pure strategy profile.

Mixed Strategies
Introduction

General Setting. We will have:

A set of players N = {1, . . . , n},

For each i N a finite set of pure strategies Si = {si1 , . . . , si i }.

The set of pure strategy profiles is S = S1 Sn .

Players have preferences defined (S ). (Not simply over S as in

the previous lecture.) We need that because players are now
allowed to randomize between pure strategies.

We will assume that such preferences are representable by

means of a von-Newman Morgenstern utility function.

That means that we can represent preferences over (S )

assigning numerical values to each pure strategy profile.

Mixed Strategies
Introduction

Example (Matching Pennies (1))

T
B

L
1, 1
1, 1

1, 1
1, 1

S1 = {T , B }, S2 = {L, R }.

S = S1 S2 . i.e. S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ) (not over S).

We assume that such preferences are representable by means of

a von-Newman Morgenstern utility function.

Therefore, for each player, we assign numerical values:

u1 (T , L) = 1, u1 (T , R ) = 1, u1 (B , L) = 1, u1 (B , R ) = 1 and
u2 (T , L) = 1, u2 (T , R ) = 1, u2 (B , L) = 1, u2 (B , R ) = 1.

Mixed Strategies
Introduction

Example (Matching Pennies (1))

T
B

L
1, 1
1, 1

1, 1
1, 1

S1 = {T , B }, S2 = {L, R }.

S = S1 S2 . i.e. S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ) (not over S).

We assume that such preferences are representable by means of

a von-Newman Morgenstern utility function.

Therefore, for each player, we assign numerical values:

u1 (T , L) = 1, u1 (T , R ) = 1, u1 (B , L) = 1, u1 (B , R ) = 1 and
u2 (T , L) = 1, u2 (T , R ) = 1, u2 (B , L) = 1, u2 (B , R ) = 1.

Mixed Strategies
Introduction

Example (Matching Pennies (1))

T
B

L
1, 1
1, 1

1, 1
1, 1

S1 = {T , B }, S2 = {L, R }.

S = S1 S2 . i.e. S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ) (not over S).

We assume that such preferences are representable by means of

a von-Newman Morgenstern utility function.

Therefore, for each player, we assign numerical values:

u1 (T , L) = 1, u1 (T , R ) = 1, u1 (B , L) = 1, u1 (B , R ) = 1 and
u2 (T , L) = 1, u2 (T , R ) = 1, u2 (B , L) = 1, u2 (B , R ) = 1.

Mixed Strategies
Introduction

Example (Matching Pennies (1))

T
B

L
1, 1
1, 1

1, 1
1, 1

S1 = {T , B }, S2 = {L, R }.

S = S1 S2 . i.e. S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ) (not over S).

We assume that such preferences are representable by means of

a von-Newman Morgenstern utility function.

Therefore, for each player, we assign numerical values:

u1 (T , L) = 1, u1 (T , R ) = 1, u1 (B , L) = 1, u1 (B , R ) = 1 and
u2 (T , L) = 1, u2 (T , R ) = 1, u2 (B , L) = 1, u2 (B , R ) = 1.

Mixed Strategies
Introduction

Example (Matching Pennies (1))

T
B

L
1, 1
1, 1

1, 1
1, 1

S1 = {T , B }, S2 = {L, R }.

S = S1 S2 . i.e. S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ) (not over S).

We assume that such preferences are representable by means of

a von-Newman Morgenstern utility function.

Therefore, for each player, we assign numerical values:

u1 (T , L) = 1, u1 (T , R ) = 1, u1 (B , L) = 1, u1 (B , R ) = 1 and
u2 (T , L) = 1, u2 (T , R ) = 1, u2 (B , L) = 1, u2 (B , R ) = 1.

Mixed Strategies
Introduction

Example (Matching Pennies (2))

T
B

L
1, 1
1, 1

1, 1
1, 1

Now, given a particular probability distribution over S, we can

calculate expected utilities.

Suppose player 2 plays 2 = 21 L + 12 R.

Then player 1s expected utility from playing T is

U1 (T , 2 ) = 1 1 + 1 (1) = 0.
2
2

And player 1s expected utility from playing B is

U1 (B , 2 ) = 1 (1) + 1 1 = 0.
2
2

Mixed Strategies
Introduction

Example (Matching Pennies (2))

T
B

L
1, 1
1, 1

1, 1
1, 1

Now, given a particular probability distribution over S, we can

calculate expected utilities.

Suppose player 2 plays 2 = 21 L + 12 R.

Then player 1s expected utility from playing T is

U1 (T , 2 ) = 1 1 + 1 (1) = 0.
2
2

And player 1s expected utility from playing B is

U1 (B , 2 ) = 1 (1) + 1 1 = 0.
2
2

Mixed Strategies
Introduction

Example (Matching Pennies (2))

T
B

L
1, 1
1, 1

1, 1
1, 1

Now, given a particular probability distribution over S, we can

calculate expected utilities.

Suppose player 2 plays 2 = 21 L + 12 R.

Then player 1s expected utility from playing T is

U1 (T , 2 ) = 1 1 + 1 (1) = 0.
2
2

And player 1s expected utility from playing B is

U1 (B , 2 ) = 1 (1) + 1 1 = 0.
2
2

Mixed Strategies
Introduction

Example (Matching Pennies (2))

T
B

L
1, 1
1, 1

1, 1
1, 1

Now, given a particular probability distribution over S, we can

calculate expected utilities.

Suppose player 2 plays 2 = 21 L + 12 R.

Then player 1s expected utility from playing T is

U1 (T , 2 ) = 1 1 + 1 (1) = 0.
2
2

And player 1s expected utility from playing B is

U1 (B , 2 ) = 1 (1) + 1 1 = 0.
2
2

Mixed Strategies
Introduction

Example (Matching Pennies (3))

T
B

L
1, 1
1, 1

1, 1
1, 1

Therefore, if player 2 plays 2 = 12 L + 12 R, player 1 is indifferent

between T , B.

Player 1, will also be indifferent between T , B and T + (1 )B

for any value of [0, 1].

We can repeat the same argument changing the roles of player 1

and player 2.

Consequently, ( 21 T + 12 B , 21 L + 12 R ) is a Nash equilibrium.

Mixed Strategies
Introduction

Example (Matching Pennies (3))

T
B

L
1, 1
1, 1

1, 1
1, 1

Therefore, if player 2 plays 2 = 12 L + 12 R, player 1 is indifferent

between T , B.

Player 1, will also be indifferent between T , B and T + (1 )B

for any value of [0, 1].

We can repeat the same argument changing the roles of player 1

and player 2.

Consequently, ( 21 T + 12 B , 21 L + 12 R ) is a Nash equilibrium.

Mixed Strategies
Introduction

Example (Matching Pennies (3))

T
B

L
1, 1
1, 1

1, 1
1, 1

Therefore, if player 2 plays 2 = 12 L + 12 R, player 1 is indifferent

between T , B.

Player 1, will also be indifferent between T , B and T + (1 )B

for any value of [0, 1].

We can repeat the same argument changing the roles of player 1

and player 2.

Consequently, ( 21 T + 12 B , 21 L + 12 R ) is a Nash equilibrium.

Mixed Strategies
Introduction

Example (Matching Pennies (3))

T
B

L
1, 1
1, 1

1, 1
1, 1

Therefore, if player 2 plays 2 = 12 L + 12 R, player 1 is indifferent

between T , B.

Player 1, will also be indifferent between T , B and T + (1 )B

for any value of [0, 1].

We can repeat the same argument changing the roles of player 1

and player 2.

Consequently, ( 21 T + 12 B , 21 L + 12 R ) is a Nash equilibrium.

Mixed Strategies
Introduction

Example (The Battle of the Sexes (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

The set of pure strategy profiles is

S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ).

Player i has to choose a strategy from (Si ).

The game has two Nash equilibria in pure strategies (T , L) and

(B , R ).

What about Nash equilibria in mixed strategies?

Mixed Strategies
Introduction

Example (The Battle of the Sexes (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

The set of pure strategy profiles is

S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ).

Player i has to choose a strategy from (Si ).

The game has two Nash equilibria in pure strategies (T , L) and

(B , R ).

What about Nash equilibria in mixed strategies?

Mixed Strategies
Introduction

Example (The Battle of the Sexes (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

The set of pure strategy profiles is

S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ).

Player i has to choose a strategy from (Si ).

The game has two Nash equilibria in pure strategies (T , L) and

(B , R ).

What about Nash equilibria in mixed strategies?

Mixed Strategies
Introduction

Example (The Battle of the Sexes (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

The set of pure strategy profiles is

S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ).

Player i has to choose a strategy from (Si ).

The game has two Nash equilibria in pure strategies (T , L) and

(B , R ).

What about Nash equilibria in mixed strategies?

Mixed Strategies
Introduction

Example (The Battle of the Sexes (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

The set of pure strategy profiles is

S = {(T , L), (T , R ), (B , L), (B , R )}.

Preferences are defined over (S ).

Player i has to choose a strategy from (Si ).

The game has two Nash equilibria in pure strategies (T , L) and

(B , R ).

What about Nash equilibria in mixed strategies?

Mixed Strategies
Introduction

Example (The Battle of the Sexes (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

Suppose player 2 plays 2 = 14 L + 34 R .

We can compute player 1s expected utilities from T :

U1 (T , 2 ) = 1 3 + 3 0 = 3 .
4
4
4

And player 1s expected utility from B:

U1 (B , 2 ) = 1 0 + 3 1 = 3 .
4
4
4

Mixed Strategies
Introduction

Example (The Battle of the Sexes (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

Suppose player 2 plays 2 = 14 L + 34 R .

We can compute player 1s expected utilities from T :

U1 (T , 2 ) = 1 3 + 3 0 = 3 .
4
4
4

And player 1s expected utility from B:

U1 (B , 2 ) = 1 0 + 3 1 = 3 .
4
4
4

Mixed Strategies
Introduction

Example (The Battle of the Sexes (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

Suppose player 2 plays 2 = 14 L + 34 R .

We can compute player 1s expected utilities from T :

U1 (T , 2 ) = 1 3 + 3 0 = 3 .
4
4
4

And player 1s expected utility from B:

U1 (B , 2 ) = 1 0 + 3 1 = 3 .
4
4
4

Mixed Strategies
Introduction

Example (The Battle of the Sexes (2))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

Player 1 is indifferent between T and B. Hence, he is also

indifferent between T , B, and randomizing between the two.

If player 1 plays 1 = 34 T + 41 B then player 2 is indifferent

between L, R and randomizing between the two.

Therefore, = (1 , 2 ) = ( 43 T + 14 B , 14 L + 34 R ) is a Nash
equilibrium.

Mixed Strategies
Introduction

Example (The Battle of the Sexes (3))

1
4

3
4

3
4

3, 1

0, 0

1
4

0, 0

1, 3

Under the strategy profile ( 34 T + 41 B , 14 L + 43 R ):

3
(T , L) occurs with probability 16
,
9
(T , R ) occurs with probability 16
,
1
(B , L) occurs with probability 16
,
3
(B , R ) occurs with probability 16
.

Mixed Strategies
Normal Form Games

Definition (Finite Normal Form Game)

A finite normal form game G = (N , {Si }i , {ui }i ) consists of

a set of players N,

for each i N a set of pure strategies Si , and

for each i N a utility function ui : S R.

(Where we assume that the ui s are bernullian utility functions.)

Mixed Strategies
Mixed Strategy Set.

Consider a normal form game G = (N , {Si }i , {ui }i ):

Si is player is set of pure strategies.

i = (Si ) is player is set of mixed strategies

Example

Suppose that player 1s set of pure strategies is S1 = {T , B }.

Then player 1s set of mixed strategies is:
1 = (S1 ) = {T + (1 )B : [0, 1]}.
For instance:
1 T + 1 B,
2
2
3 T + 1 B,
4
4
1 T + 4 B,
5
5
0T + 1B = B
are mixed strategies.

Mixed Strategies
Nash Equilibrium

Consider a game G = (N , {Si }i , {ui }i ), and denote the set of mixed

strategy profiles as = 1 n , where i = (Si ).

Definition (Nash Equilibrium)

A strategy profile = (1 , . . . , n ) is a Nash equilibrium if for
each i N
Ui (i , i ) Ui (i , i ) for every i i .

Mixed Strategies
Carrier or Support of a Mixed Strategy

Definition (Carrier or Support)

The carrier or support of a mixed strategy i , which we denote as
C (i ), is the set of pure strategies that receive strictly positive
s
probability from i . That is, C (i ) = {si Si : i i > 0}.

Example

Let S1 = {T , M , B } be the set of pure strategies of player 1.

Consider the mixed strategy 1 = 31 T + 32 M.

The Carrier of 1 is C (1 ) = {T , M }.

Mixed Strategies
Carrier or Support of a Mixed Strategy

Definition (Carrier or Support)

The carrier or support of a mixed strategy i , which we denote as
C (i ), is the set of pure strategies that receive strictly positive
s
probability from i . That is, C (i ) = {si Si : i i > 0}.

Example

Let S1 = {T , M , B } be the set of pure strategies of player 1.

Consider the mixed strategy 1 = 31 T + 32 M.

The Carrier of 1 is C (1 ) = {T , M }.

Mixed Strategies
Carrier or Support of a Mixed Strategy

Definition (Carrier or Support)

The carrier or support of a mixed strategy i , which we denote as
C (i ), is the set of pure strategies that receive strictly positive
s
probability from i . That is, C (i ) = {si Si : i i > 0}.

Example

Let S1 = {T , M , B } be the set of pure strategies of player 1.

Consider the mixed strategy 1 = 31 T + 32 M.

The Carrier of 1 is C (1 ) = {T , M }.

Mixed Strategies
Carrier or Support of a Mixed Strategy

Definition (Carrier or Support)

The carrier or support of a mixed strategy i , which we denote as
C (i ), is the set of pure strategies that receive strictly positive
s
probability from i . That is, C (i ) = {si Si : i i > 0}.

Example

Let S1 = {T , M , B } be the set of pure strategies of player 1.

Consider the mixed strategy 1 = 31 T + 32 M.

The Carrier of 1 is C (1 ) = {T , M }.

Mixed Strategies
Mixed Strategies in a Nash Equilibrium

Let be a Nash Equilibrium.

Notice that no player can unilaterally deviate and obtain a higher

payoff.

Suppose that for player i, the carrier of i contains more than one
pure strategy

Then if players other than i play according to i , player i is

indifferent between any two elements of C (i )

In other words, if a player plays a mixed strategy in equilibrium, then

he is indifferent among all his pure strategies that receive positive
probability.

Mixed Strategies
Mixed Strategies in a Nash Equilibrium

Example

T
B

L
3, 1
0, 0

R
0, 0
1, 3

Under the Nash equilibrium ( 34 T + 14 B , 14 L + 34 R ) player 1 is indifferent

between T and B and player 2 is indifferent between L and R.

This implies that:

An equilibrium in dominant strategies is always in pure strategies

A strict equilibrium is always in pure strategies.

Mixed Strategies
Dominated Strategies

Definition (Strictly Dominated Strategy)

A strategy i is strictly dominated by i if for every i i
Ui (i , i ) < Ui (i , i ).
However, we only need to check against pure strategy profiles of the
opponents. This implies:

Definition (Strictly Dominated Strategy)

A strategy i is strictly dominated by i if for every si Si
Ui (si , i ) < Ui (si , i ).

Mixed Strategies
Weakly Dominated Strategy

Definition (Weakly Dominated Strategy)

A strategy i is weakly dominated by i if for every i i
Ui (i , i ) Ui (i , i )
and there exists at least one i such that
Ui (i , i ) < Ui (i , i ).

Mixed Strategies
Weakly Dominated Strategy

Again, we only need to check against pure strategy profiles of the

opponents.

Definition (Weakly Dominated Strategy)

A strategy i is weakly dominated by i if for every si Si
Ui (si , i ) Ui (si , i )
such that
and there exists at least one s
i

Ui (s
i , i ) < Ui (si , i ).

Mixed Strategies
Dominated strategies. Properties.

If a pure strategy si is dominated then every mixed strategy that

gives si a strictly positive probability is also dominated.

Example

T
B

L
1, 1
0, 0

B is dominated.

1 = 12 T + 12 B is also dominated.

R
0, 0
0, 0

If player 2 plays L, 1 gives a payoff equal to 12 and T gives a

payoff equal to 1.
If player 2 plays R, both 1 and T give a payoff equal to 0.

Mixed Strategies
Dominated strategies. Properties.

More generally, if a mixed strategy i is dominated then every

mixed strategy whose carrier coincides with or contains C (i ) is
also dominated.

Mixed Strategies
Dominated strategies. Properties.

A mixed strategy that does not give strictly positive probability to

any dominated pure strategy may be dominated.

Example

T
M
B

L
10,10
0,0
6,6

R
0,0
10,10
6,6

T is not dominated.

M is not dominated.

But 21 T + 12 M is strictly dominated by B

Therefore, the following strategies are dominated:

T + (1 )M, for every (0, 1)
T + M + (1 )B, for every > 0 and > 0 such that
+ < 1.

Mixed Strategies
Dominated strategies. Properties.

A strategy (pure of mixed) i may only be dominated by a mixed

strategy.

Example

T
M
B

L
10,10
0,0
4,4

B is not dominated by T .

B is not dominated by M.

But B is dominated by 21 T + 12 M.

R
0,0
10,10
4,4

Mixed Strategies
Admissible Equilibrium

Definition (Undominated Nash Equilibrium)

The strategy profile is an undominated equilibrium, or admissible
equilibrium, if is a Nash equilibrium where no player uses a
dominated strategy.

Mixed Strategies
Best Reply Correspondence

Definition (Pure Best Reply)

We say that player is pure strategy si is a pure best reply against i
if for all si
Ui (i , si ) Ui (i , si ).
The set of player is best replies against i is denoted PBRi (i ), or
simply PBRi ().

Mixed Strategies
Best Reply Correspondence

Definition (Best Reply)

We say that player is pure strategy i is a best reply against i if for
all i
Ui (i , i ) Ui (i , i ).

We denote the set of player is best replies against i as

BRi (i ), or simply BRi ()

Notice that BRi (i ) = BRi () = (PBRi ()).

Mixed Strategies
Best Reply Correspondence. Nash Equilibrium

Definition (Nash Equilibrium)

The strategy profile is a Nash equilibrium if every player is playing a
best reply against . That is, i BRi ( ) for every i N.

Mixed Strategies
Best Reply Correspondence. Example

Example (The Battle of the Sexes. (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

Suppose that player 2 plays L with probability y and R with

probability 1 y.

We have that:
U1 (T , y ) = 3y
U1 (B , y ) = 1 y .

Mixed Strategies
Best Reply Correspondence. Example

Example (The Battle of the Sexes. (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

Suppose that player 2 plays L with probability y and R with

probability 1 y.

We have that:
U1 (T , y ) = 3y
U1 (B , y ) = 1 y .

Mixed Strategies
Best Reply Correspondence. Example

Example (The Battle of the Sexes. (1))

T
B

L
3, 1
0, 0

R
0, 0
1, 3

Suppose that player 2 plays L with probability y and R with

probability 1 y.

We have that:
U1 (T , y ) = 3y
U1 (B , y ) = 1 y .

Mixed Strategies
Best Reply Correspondence. Example.

Example (The Battle of the Sexes. (2))

y

1y

3, 1

0, 0

0, 0

1, 3

U1 (T , y ) =3y
U1 (B , y ) =1 y .
T %1 B if and only if y 41 .

Mixed Strategies
Best Reply Correspondence. Example.

Example (The Battle of the Sexes. (3))

T %1 B if and only if y 41 .
f x represents the probability that player 1 plays T , we have that:

if y > 14
x = 1
BR1 (y ) x [0, 1] if y = 14

x =0
if y < 14 .

Mixed Strategies
Best Reply Correspondence. Example.

Example (The Battle of the Sexes. (4))

y

1
BR1 (y )

1
4
x
1

Mixed Strategies
Best Reply Correspondence. Example.

Example (The Battle of the Sexes. (5))

L

3, 1

0, 0

(1 x ) B

0, 0

1, 3

We can also calculate BR2 :

U2 (L, x ) =x
U2 (R , x ) =3(1 x ).
L %2 R if and only if x 34 .

Mixed Strategies
Best Reply Correspondence. Example.

Example (The Battle of the Sexes. (6))

L %2 R if and only if x 43 .
(Remember that y represents the probability that 2 plays L)

if x >
y = 1
BR2 (x ) y [0, 1] if x =

y =0
if x <

3
4
3
4
3
4

Mixed Strategies
Best Reply Correspondence. Example.

Example (The Battle of the Sexes. (7))

y

1
BR1 (y )

BR2 (x )

1
4
3
4

x
1

Mixed Strategies
Best Reply Correspondence. Example.

Example (The Battle of the Sexes. (8))

The set of Nash equilibria corresponds to all the points where BR1 (y )
and BR2 (x ) intersect.

(T , L) corresponds to x = 1 and y = 1.

(B , R ) corresponds to x = 0 and y = 0.

( 34 T + 14 B , 14 L + 34 R ) corresponds to x =

3
4

and y = 41 .