"Inside  every small problem is a large problem struggling to get out."

## Thread and Full Text View

Respond to the question: Iterated assurance games?

 01/28/2000 12:51 PM by pedro; More on the chicken game
Consider the following chicken game:

 C N C 3 , 3 1, 4 N 4 , 1 0 , 0

If this game is only played once there are two pure strategy NE: (C,N) and (N,C) and
mixed strategy NE: (1/2C+1/2N, 1/2C+1/2N).
If this game is played infinitely there may be much more possible outcomes. For example
(C,C) may be an outcome in equilibrium. How? Imaging that the players play (C,C) from
the beginning, if one deviates, he/she is punished by reverting to the static NE that is worse
for him/her. Then, in analyzing if deviating or not, the players should do the following
calculation:
U(C)=3 and U(N)=4(1-d)+d
where d is the discount factor.
U(C)>U(N) if d>1/3.
Then, if d is greater or equal to 1/3, playing (C,C) in every period is an equilibrium outcome.

A good reference for repeated games theory is "Game Theory", Ch.5, Fudenberg and
Tirole. [Manage messages]

 01/27/2000 02:30 PM by Pedro Dal Bó; Chicken game
I think that the chicken game is quite different from the prisoner dilemma. In the latter both agents would like to cooperate but the incentives to cheat make both of them deviate and end in an outcome that is not Pareto [View full text and thread]

 01/27/2000 06:50 AM by name withheld; iterated assurance games
Since there are variations on the one-shot prisoner's dilemma, I am wondering if there are variations on the iterated prisoner's dilemma game. Particularly chicken and the assurance game. Has there been any papers written on these [View full text and thread]