Documente Academic
Documente Profesional
Documente Cultură
Simultaneous-Move Games I
1 Simultaneous-Move Games I
Strategic / Normal Form
Pure versus Mixed Strategies
Prisoner’s Dilemma
Dominant Strategies
Dominance Solvability (Iterated Elimination of Dominated Strategies)
Best-Response Analysis
Rationalizability
Nash Equilibrium
Extension (More Players)
Simultaneous-Move Game
Simultaneous-Move Game
In simultaneous-move games, players have to make strategy choices simultane-
ously, without knowing the strategies that have been chosen by other player(s).
2 Rock-Paper-Scissors
Although there is no information what other players actually choose, the strategic
choices available to each player are common knowledge.
Players must think not only about their own best strategic choice, but also the best
strategic choice of other players.
Normal / Strategic Form
Column Player
(2nd payoff)
Strategy C1 Strategy C2
Strategy R1 a ,b c ,d
Row Player
(1st payoff)
Strategy R2 e ,f g ,h
Pure versus Mixed Strategies
Pure Strategy
A player pursues a pure strategy if she always chooses the same strategic action out
of all the action choices available.
Mixed Strategy
A player pursues a mixed strategy if she randomizes in some manner among the strate-
gic action choices available to her.
- Sometimes pitch a curveball, sometimes a slider ("mix it up"and "keep them guessing").
In the classic movie Butch Cassidy and the Sundance Kid (1969), two criminals get
caught suspected of committing a train robbery. They are kept and interrogated in
separate cells. The Cops do not have enough evidence to prove robbery, but they have
enough to nail them on weapons charges.
1 If they both confess, they get sentenced to jail for 5 years for robbery.
2 If neither confesses, they get sentenced to jail for 2 years on weapons charges.
3 If only of them confesses, the rat gets lenient treatment of 1 year in jail, the other gets 10
years. Sundance Kid
(Robert Redford)
Confess Deny
Confess -5 , -5 -1 , -10
Butch Cassidy
(Paul Newman)
Deny -10 , -1 -2 , -2
Dominant Strategies
A dominant strategy is the best choice regardless of what the player thinks her opponent might do.
Sundance Kid
(Robert Redford)
Confess Deny
Confess -5 , -5 -1 , -10
Butch Cassidy
(Paul Newman)
Deny -10 , -1 -2 , -2
Confess is a dominant strategy for both criminals. Their payoffs are (−5, −5).
Both could be better off by choosing to deny which would result in payoffs
(−2, −2).
Stampede
When a fire breaks out in a building, people often suffer injuries/death from
trampling rather than the smoke or fire.
Player 2
Rush Walk
Rush -5 , -5 0 , -8
Player 1
Walk -8 , 0 -1 , -1
A moderately dangerous physical situation turns fatal due to the human response,
which nevertheless, is rational from a self-preservation perspective.
Solutions
Dominance Solvability
A game is dominance solvable if successive elimination of dominated strategies results
in a unique outcome.
Player 2
Left Center Right
Top 2 ,3 0 ,2 0 ,1
Player 1 Middle 3 ,0 3 ,1 1 ,2
Bottom 4 ,0 2 ,2 0 ,1
Player 2
Left Center Right
Top 2 ,3 0 ,2 0 ,1
Player 1 Middle 3 ,0 3 ,1 1 ,2
Bottom 4 ,0 2 ,2 0 ,1
Player 2
Left Center Right
Top 2 ,3 0 ,2 0 ,1
Player 1 Middle 3 ,0 3 ,1 1 ,2
Bottom 4 ,0 2 ,2 0 ,1
Player 2
Left Center Right
Top 2 ,3 0 ,2 0 ,1
Player 1 Middle 3 ,0 3 ,1 1 ,2
Bottom 4 ,0 2 ,2 0 ,1
Player 2
Left Center Right
Top 2 ,3 0 ,2 0 ,1
Player 1 Middle 3 ,0 3 ,1 1 ,2
Bottom 4 ,0 2 ,2 0 ,1
Eliminating strategies for other players can never cause a strictly dominated strategy for
Player i to cease being strictly dominated.
Player 2
1 Eliminate Middle (strictly dominated by Bottom) Left Right
Bottom 4 ,2 3 ,1
1 Eliminate Right (strictly dominated by Left)
Weakly dominated strategies can cease to be weakly dominated based on the order of
elimination.
Player 2
1 Eliminate Middle (strictly dominated by Top) Left Right
Player 1 Middle 1 ,1 0 ,0
1 Eliminate Bottom (strictly dominated by Top)
2 Bottom 0 ,0 1 ,1
Eliminate Right (weakly dominated by Left)
Best Response
Player A’s best-response to player B’s strategy B1 is the strategy of player A out of all
available strategic choices that gives player A the highest payoff against strategy B1 .
Take a player’s perspective: For each of the choices of the other player, what is
the best choice for this player?
Underline the best-responses for all players and all strategy choices.
A best-response might not be unique. A player can have a number of strategies that are all best responses if they all earn the same
payoff, and earn a greater payoff than other strategies. In such cases, underline all of the best responses.
Player 2
Left Center Right
Top 2 ,3 0 ,2 0 ,1
Player 1 Middle 3 ,0 3 ,1 1 ,2
Bottom 4 ,0 2 ,2 0 ,1
Rationalizability
Never-Best Response
A Never-Best-Response is a strategy that is non-optimal against any strategy choice
of the other player.
Rationalizability
The set of strategies that survives iterated elimination of never-best-responses is called
rationalizable.
Top 2 ,3 0 ,2 0 ,1
Player 1 Middle 3 ,0 3 ,1 1 ,2
Bottom 4 ,0 2 ,2 0 ,1
Top is a never-best-response for player 1, i.e. it is non-optimal no matter what player 2 does.
Iteratively Elimination Never-Best Responses
Player 2 Player 2
Left Center Right Left Center Right
Top 2 ,3 0 ,2 0 ,1 Top 2 ,3 0 ,2 0 ,1
Bottom 4 ,0 2 ,2 0 ,1 Bottom 4 ,0 2 ,2 0 ,1
Player 2 Player 2
Left Center Right Left Center Right
Top 2 ,3 0 ,2 0 ,1 Top 2 ,3 0 ,2 0 ,1
Bottom 4 ,0 2 ,2 0 ,1 Bottom 4 ,0 2 ,2 0 ,1
Finally, Center is a never-best-response for player 2 and (Middle, Right) is the unique
outcome from iteratively eliminating never-best-responses in this game.
P - Beauty Contest (Guessing Game)
2
The winner of the game is the player whose choice is closest to 3 of the overall average.
2
Any number in the interval [68, 100] is a never-best-response and can be eliminated, since 3
The remaining range of numbers is now [0, 67] and any number in the range [46, 67] is a
2
never-best-response and can be eliminated, since 3 of the average is at or below 44 23 even if
every other player picks 67.
Born in Bluefield (West Virginia), mathematician and economist John Forbes Nash
(1928 - ) developed the equilibrium concept named after himself in his doctoral disser-
tation at Princeton in 1949. For this contribution and his mathematical bargaining theory
for cooperative games, he shared the 1994 Nobel Memorial Prize in Economics (with
Reinhard Selten and John Harsanyi). Nash was troubled for many years by mental ill-
ness, portrayed in the Academy Award winning movie, A Beautiful Mind (2001), which
is based on Sylvia Nasar’s biography of John F. Nash with the same title (Simon &
Schuster, 1998).
John F. Nash
Unfortunately, the game theory examples depicted in the movie A Beautiful Mind are misleading
and incorrect!
Nash Equilibrium
A Nash Equilibrium is a list of strategies, one for each player, with the property that
each is a best response to the remaining strategies by the other players. (i.e. maximizes
the corresponding player’s payoff among available options).
Sundance Kid
Confess Deny
Confess -5 , -5 -1 , -10
Butch Cassidy
Deny -10 , -1 -2 , -2
The unique Nash equilibrium in the Prisoner’s Dilemma is for both criminals to
confess.
Cell-by-Cell Method
For each cell check whether a player has a profitable deviation to another strategy choice given the other players’
equilibrium choice.
Strategy-by-Strategy
Simpler than cell-by-cell inspection, faster and is less likely to result in mistakes: For a given strategy of the other
player, mark the best-response; Nash equilibria are characterized by best responses of all players.
Player 2
Find all Nash equilibria. V W X Y Z
A 9 ,9 7 ,1 5 ,6 3 ,4 1 ,1
B 7 ,8 5 ,2 3 ,6 1 ,4 3 ,3
Player 1 C 5 ,6 3 ,3 1 ,8 9 ,7 1 ,5
D 3 ,9 1 ,9 9 ,4 7 ,9 5 ,9
E 1 ,2 9 ,8 7 ,7 5 ,6 3 ,7
Common Error I
One of the most common error in the analysis of equilibria is confusing equilibrium
strategies with equilibrium payoffs.
Player 2
X Y
A 3 ,3 1 ,1
Player 1
B 2 ,4 5 ,2
The equilibrium is NOT (3, 3), which are the payoffs players earn in equilibrium.
Common Error II
A slightly changed version of the previous game illustrates why this distinction is
important.
Player 2
X Y
A 3 ,3 1 ,1
Player 1
B 2 ,4 3 ,3
The game still has a unique Nash-equilibrium, which is (A,X), and the payoffs in
equilibrium are still (3,3).
The payoffs (3,3) do not uniquely identify the equilibrium outcome since (3,3) is
also obtainable if the players play (B,Y).
Nash Equilibrium II
Player 2
Left Center Right
Top 2 ,3 0 ,2 0 ,1
Player 1 Middle 3 ,0 3 ,1 1 ,2
Bottom 4 ,0 2 ,2 0 ,1
Player 2
Left Center-Left Center-Right Right
Top 6 ,3 3 ,5 1 ,1 4 ,0
Upper-Mid 1 ,1 3 ,2 2 ,0 4 ,3
Player 1
Lower-Mid 5 ,2 4 ,3 3 ,2 5 ,2
Bottom 7 ,6 2 ,5 0 ,3 8 ,7
Player 2
Left Center-Left Center-Right Right
Top 6 ,3 3 ,5 1 ,1 4 ,0
Upper-Mid 1 ,1 3 ,2 2 ,0 4 ,3
Player 1
Lower-Mid 5 ,2 4 ,3 3 ,2 5 ,2
Bottom 7 ,6 2 ,5 0 ,3 8 ,7
Player 2
Left Center-Left Center-Right Right
Top 6 ,3 3 ,5 1 ,1 4 ,0
Upper-Mid 1 ,1 3 ,2 2 ,0 4 ,3
Player 1
Lower-Mid 5 ,2 4 ,3 3 ,2 5 ,2
Bottom 7 ,6 2 ,5 0 ,3 8 ,7
Player 2
Left Center-Left Center-Right Right
Top 6 ,3 3 ,5 1 ,1 4 ,0
Upper-Mid 1 ,1 3 ,2 2 ,0 4 ,3
Player 1
Lower-Mid 5 ,2 4 ,3 3 ,2 5 ,2
Bottom 7 ,6 2 ,5 0 ,3 8 ,7
Simultaneous-move games are often played with more than just two players.
2 The networks battle for share of viewers (0 − 100). Higher shares are preferred
due to higher advertising revenues.
CBC and CTV choose within the tables and Global chooses across the tables.
Sitcom 38 , 31 , 33 33 , 35 , 32 Sitcom 36 , 30 , 34 35 , 35 , 30
CBC CBC
Game Show 32 , 34 , 34 30 , 36 , 34 Game Show 33 , 34 , 33 33 , 35 , 32
The unique Nash equilibrium has CBC and Global choosing the sitcom, while
CTV chooses the game show.