I read about this one -- IMHO they made a meta-error. Because now you just have a prisoner's dilemma problem regarding how the agents WITHIN that "southampton group" should act -- ie whether or not to follow the scheme the "programmers set" as it were
BTW talking of the _Evolution of Cooperation_ book, Its a great pity Axlerod more or less went crazy later in life and joined the U.N. !!!!
Posted by JPMay at October 23, 2004 07:14 AMJP,
well, you *could* say that ... I'd say that the agents were always supposed to cooperate anyway, that's the beginning assumption of the two prisoners.
What is interesting though is that up until now, everyone assumed that it was always exactly two prisoners. And what happened between those exact two prisoners? Of course, in the game, there were multiple sets of two prisoners, and nobody thought about collusion across sets, only within each set, as if that was the entire universe. A classic example of being stuck in the box of ones assumptions!
But in practice, what the Southampton team looked at is more the "real problem." When you think about those two prisoners, the reason they cooperate is because the Don back at their home with their wifes and their children will rub out anyone who doesn't cooperate...
So there are more than 2 players in the game, in general. And the simplification of 2 players that was made for explanatory purposes back when Axelrod was sane has kind of become an accepted tenet of the science, at least as the game would have it.