Current mood:

crazy

game theorists unite (untie?)
In my daily browsing of /. I came across a story that shocked me at first. Wired News is running a story about the 20th-anniversary running of an Axelrod style Iterated Prisoner's Dilemma competition.
Apparently Tit-For-Tat was toppled this year by some strategies from the University of Southampton. Which..well is a little surprising, since TFT is a pretty good strategy. What was really amazing at first was that their "strategy" is an ESS (Evolutionarily Stable Strategy).

Maybe I'm just losing touch with my inner game theorist, but it did surprise me, at least until I read how it happened.

Basically, what happened is that they submitted 60 strategies which would all recognize each other and upon doing so, immediately assume "master and slave" roles with each other, and then just sabatoge anyone else.

Surprise turned to disgust. Maybe I'm just a purist who thinks that the whole point of the Prisoner's Dilemma thought game is that the prisoners aren't allowed to communicate, and this is essentially the same as them "tapping in Morse code on the prison wall."

Game theory is an excellent field of mathematics, and was one of the most enjoyable courses I took as an undergraduate. I'm led to ask the community now, have you ever participated in an IPD tournament? What was the idea of your strategy? (And how did you do?)

I ran a similar tournament a couple years ago, for the game of Chicken. (Picture two cars driving at each other. Both swerve, little gained for each. If you swerve, and your opponent doesn't then you lose a bit, your opponent gains a bit. If nobody swerves, both lose a lot.)

A lot of fun, and these things lead to some pretty interesting strategies.