While I am not sure if this game offers a great example for a prisoners dilemma I am sure its logic can offer us some help. When faced with an iterated prisoners dilemma the best strategy is "tit for tat". Cooperate from the start and when the other bertrays you, fight him until he gives in and is willing to cooperate again.
But as I said this is not a prisoners dilemma. In a prisoners dilemma one should have the option between defecting and cooperating. Obviously cooperating here is not very fruitful since you have to kill eachother off one way or the other.

But I catch your drift.

The prisoner's dilemma as I learned on my Microeconomy class, states exactly the opposite. It states that if you don't know how the other person is gonna react, you should always choose the option that is best for YOU un average, and it doesn't matter if it helps to the community or not.

But the point is, this is an iterative (i.e., repeated) scenario. So yeah, you can rip someone off, but for the rest of your history that guy will remember that you aren't to be trusted.

Meatbomb is right. Read Axelrod: Evolution of cooperation. It shows that Tit for tat is the best strategy in an iterative scenario. But again, this is not a prisoners dilemma.

It's not tit for tat. It's assuming everyone does whats best for them so you need to do whats best for you. That the outcome often ends up being the same in the particular example is incidental.

It's not tit for tat. It's assuming everyone does whats best for them so you need to do whats best for you. That the outcome often ends up being the same in the particular example is incidental.

I have a Game Theoretic Methods final tomorrow, so this is my kind of forum discussion :). I don't know if an iterative tit-for -tat really describes this either- equilibrium in tit for tat strategies usually hinges on the fact that people exponentially discount future payoffs that will come in later time periods. So there is an equilibrium between getting a bigger payoff in the current time frame and a lower payoff in the periods thereafter, because the opponent retaliates. This game has only one pay off period in a given game, and in iterated games, you are not facing the same opponent, so I don't think there is a simple game theory explanation. although I find the original comment the most useful- there is incentive to cooperate. I think the reason this works is that you get a positive ranking payoff for 2nd and usually third place finishes. So my strategy has been to cooperate when I am losing- offer the clearly dominant player no contest if he will spare me until the end. This usually gets me second place I have found. When I start off in a good game and I come out in front, I simply play to win. Cooperating and taking second every game sure beats getting killed off early in some games and winning others.

But at the same time some people would call the "wait it out" strategy annoyingly opportunist.

However I agree with the meatbomb and overlast on cooperation being the key strategy. The more of an asshole you are to others, the more they'll want to beat on you in future games. firm but fair is the way to go. Be nice to me and I'll be nice to you.

The bigger question is what to do when facing an opponent you've never seen. Cooperate or attack mercilessly? Tit for tat says you should give the benefit of the doubt, but hmmm ...

I find I'm at a real disadvantage if I'm playing against people who already know each other.

BUMP!