You are on page 1of 9

EE6417: Incentive-Centered Design

Lecture 06

Bharadwaj Satchidanandan

Department of Electrical Engineering


Indian Institute of Technology Madras

Bharadwaj Satchidanandan EE6417: Incentive-Centered Design Lecture 06 1/9


Example

Example: The General’s Dilemma

Player 2
A R
A (1, 1) (−1, 2)
Player 1
R (2, −1) (0, 0)

The pure strategy profile (R, R) is a Nash equilibrium.

Bharadwaj Satchidanandan EE6417: Incentive-Centered Design Lecture 06 2/9


Example

Example: Stag-and-Hare
There are two hunters who can either choose to hunt a stag or a hare. The stag can
feed them for a longer time, but cannot be hunted down unless both hunters hunt
together. While the hare is a smaller meal, it can be hunted down by each hunter by
himself. The decision problem faced by the hunters can be described by a game with the
following payoff matrix.

Hunter 2
S H
S (2, 2) (0, 1)
Hunter 1
H (1, 0) (1, 1)

Both (S, S) and (H, H) are Nash equilibria of the game.


This game is prototypical of many coordination problems that show up in a society.

Bharadwaj Satchidanandan EE6417: Incentive-Centered Design Lecture 06 3/9


Example

Example: Rock-Paper-Scissors

Player 2
R P S
R (0, 0) (−1, 1) (1, −1)
Player 1 P (1, −1) (0, 0) (−1, 1)
S (−1, 1) (1, −1) (0, 0)

There is no pure strategy Nash equilibrium.


   
 1/3 1/3 
The mixed strategy profile 1/3 , 1/3 is a Nash equilibrium.
1/3 1/3
How do we know that?

Bharadwaj Satchidanandan EE6417: Incentive-Centered Design Lecture 06 4/9


Checking if a given Strategy Profile is a Nash Equilibrium

It is easy to check if a given pure strategy profile s = (s1 , . . . , sn ) is a Nash


equilibrium or not: For each player i, fix all other players’ strategies at s−i , and
scan over Si to see if any strategy si′ yields Player i a higher payoff than si . If not,
then s is a pure strategy Nash equilibrium.
How to check if a given mixed strategy x = (x1 , . . . , xn ) is a Nash equilibrium?
In principle, one could proceed analogously: For each player i, fix all other players’
strategies at x−i and scan over ∆(Si ) to see if there is any mixed strategy x′i that
yields a higher payoff to player i than xi . The problem, of course, is that ∆(Si ) is
an infinite set, and so conducting the check would take an infinite amount of time.
This begs the question, “How to check if a given mixed strategy profile is a Nash
equilibrium or not in a finite amount of time?”

Bharadwaj Satchidanandan EE6417: Incentive-Centered Design Lecture 06 5/9


Checking if a given Strategy Profile is a Nash Equilibrium
The following proposition suggests a finite time algorithm to check if a given
mixed strategy profile is a Nash equilibrium or not.

Notation: Given a pure strategy s of a player, denote by 1s the mixed strategy


that puts probability 1 on the pure strategy s and probability 0 on all other pure
strategies. Simply put, 1s denotes the pure strategy s.

Proposition
A mixed strategy profile (x1 , . . . , xn ) is a Nash equilibrium if and only if for every
i ∈ {1, . . . , n} and every si , si′ ∈ Si ,

µi (1si , x−i ) = µi (1si′ , x−i ). (1)

Proof.
Exercise.

Bharadwaj Satchidanandan EE6417: Incentive-Centered Design Lecture 06 6/9


Example

Example: Matching Pennies

Player 2
H T
H (1, −1) (−1, 1)
Player 1
T (−1, 1) (1, −1)

There are no pure-strategy Nash equilibria.


   
0.5 0.5
The mixed strategy profile , is a Nash equilibrium.
0.5 0.5

Bharadwaj Satchidanandan EE6417: Incentive-Centered Design Lecture 06 7/9


A Slice of History (After Nash)
Nash established his result in the context of Normal-form games,
which are rather general models of games.
▶ Via strategic normalization, one can convert any extensive form game
into a Normal form game.
▶ Borel himself recognized this and explained the idea in a few short
sentences at the very beginning of his paper.
▶ Von Neumann elucidated the idea in greater detail in his 1928 paper.
Until Nash, the most pressing issue in Game theory was the definition
and existence of equilibrium. Once Nash settled that question, certain
secondary concerns came to the fore, a longstanding one of which was
the adequacy of the normal form.
Certain limitations of the normal form were recognized. Certain Nash
equilibria of the normal form are irrational in some sense when
reexamined in the extensive form. This observation led to refined
notions of Nash equilibrium.

Bharadwaj Satchidanandan EE6417: Incentive-Centered Design Lecture 06 8/9


A Slice of History (After Nash)
Notions of stability of a Nash equilibrium were also developed, which led to
another basis for refined notions of Nash equilibrium.
As a result, the ensuing decades witnessed a proliferation of many refined
notions of Nash equilibrium:
▶ Essential equilibrium (Wu Wen-Tsun and Jiang Jia-He, 1962),
▶ Subgame Perfect equilibrium (Selten, 1965),
▶ Regular equilibrium (Harsanyi, 1973),
▶ Trembling Hand Perfect equilibrium (Selten, 1975),
▶ Proper equilibrium (Myerson, 1978),
▶ Sequential equilibrium (Kreps and Wilson, 1982),
▶ Etc.
These constitute the theory of equilibrium selection. When there are
multiple Nash equilibria, on what basis does the Game theorist choose one
NE over another as a model for real-world behavior? This is the question
that the theory of equilibrium selection attempts to answer. We don’t study
that in this course.

Bharadwaj Satchidanandan EE6417: Incentive-Centered Design Lecture 06 9/9

You might also like