Game Theory Calls Cooperation Into Question

An anonymous reader sends this excerpt from Quanta Magazine: The physicist Freeman Dyson and the computer scientist William Press, both highly accomplished in their fields, have found a new solution to a famous, decades-old game theory scenario called the prisoner's dilemma, in which players must decide whether to cheat or cooperate with a partner. The prisoner's dilemma has long been used to help explain how cooperation might endure in nature. After all, natural selection is ruled by the survival of the fittest, so one might expect that selfish strategies benefiting the individual would be most likely to persist. But careful study of the prisoner's dilemma revealed that organisms could act entirely in their own self-interest and still create a cooperative community.

Press and Dyson's new solution to the problem, however, threw that rosy perspective into question (abstract). It suggested the best strategies were selfish ones that led to extortion, not cooperation.

[Theoretical biologist Joshua] Plotkin found the duo's math remarkable in its elegance. But the outcome troubled him. Nature includes numerous examples of cooperative behavior. For example, vampire bats donate some of their blood meal to community members that fail to find prey. Some species of birds and social insects routinely help raise another's brood. Even bacteria can cooperate, sticking to each other so that some may survive poison. If extortion reigns, what drives these and other acts of selflessness?"

Comments

Direct link to the original paper

[mc6809e]

Iterated Prisoner's Dilemma contains strategies that dominate any evolutionary opponent

The Selfish Gene

[flyhigher]

The selfish gene theory popularized by Richard Dawkins states that evolution works on genes, not on individuals. Any gene which gives rise to behavior that will cause more copies of that gene to survive, will increase its percentage in the gene pool at large.

http://en.wikipedia.org/wiki/G...

Evolution isn't about personal survival

[taustin]

It's about procreation and the survival of the genetic line. Individual survival is irrelevant, especially once one has procreated. (Though even those who don't contribute to the survival of the genetic line of their family - the person who has a sibling willing to sacrifice themselves to save the family enhances the chances the family will procreate.)

This kind of confusion is what happens when people try to do research outside of their expertise. If you want to understand biology, ask a biologist, not a physicist or a computer geek. (Though a lot of biologists make the same mistake, of course.)

Re:Tit-for-tat is not ESS. It is well known.

[david_thornley]

The advantages of tit-for-tat are that it will stay cooperative against cooperative opponents, maximizing the total gain, and that it will not lose to its other player by more than one defection. It isn't necessarily the best strategy, but it has some provable advantages.

Therefore, if this strategy, whatever the heck it is, plays against tit-for-tat, it will come out ahead by a small amount. No extortion is possible against tit-for-tat, since it has a very short memory. Any serious attempt to hurt it hurts the opponent almost as much. The outcome of the game will be determined by the opponent, but it isn't clear to me that this is good for the opponent.

In a series of games, with players changing algorithms, tit-for-tat is not particularly susceptible to extortion, since it fundamentally yields the opponent one extra defection/cooperation win. Any attempt to extort it into more than one will fail in a competitive environment, since, if tit-for-tat is trashed the opponent is trashed almost as much. Tit-for-tat against itself, or any other strategy that won't defect first, will get straight cooperation rewards, while defect-first strategies have to accept some mutual defections, lowering the total score.

So, while I'm willing to concede that the mathematics is correct (it's been a long time since I read a mathematical paper, so I haven't checked it out fully yet), it doesn't look like it's going to make much of a difference in final score.