Wired has a news story about a "new approach" to prisoners' dilemma, but despite the acclaim in the article the "strategy" is insignificant -- that it works is an artifact of poorly phrased rules, not of a useful advance in philosophy.

Proving that a new approach can secure victory in a classic strategy game, a team from England's Southampton University has won the 20th-anniversary Iterated Prisoner's Dilemma competition, toppling the long-term winner from its throne. ...

Before Southampton came along, a strategy called Tit for Tat had a consistent record of winning the game. Under that strategy, a player's first move is always to cooperate with other players. Afterward, the player echoes whatever the other players do. The strategy is similar to the one nuclear powers adopted during the Cold War, each promising not to use its weaponry so long as the other side refrained from doing so as well. ...

Teams could submit multiple strategies, or players, and the Southampton team submitted 60 programs. These, Jennings explained, were all slight variations on a theme and were designed to execute a known series of five to 10 moves by which they could recognize each other. Once two Southampton players recognized each other, they were designed to immediately assume "master and slave" roles -- one would sacrifice itself so the other could win repeatedly.

If the program recognized that another player was not a Southampton entry, it would immediately defect to act as a spoiler for the non-Southampton player. The result is that Southampton had the top three performers -- but also a load of utter failures at the bottom of the table who sacrificed themselves for the good of the team.

It's a gimmick, plain and simple, and it works for the same reason parts of the USSR looked prosperous to Western visitors: it's easy to push one guy to the top if you're willing to sacrifice 59 others. The organizer of the competition, Graham Kendall, explains why Southampton's results are meaningless (from my perspective).
Kendall noted that there was nothing in the competition rules to preclude such a strategy, though he admitted that the ability to submit multiple players means it's difficult to tell whether this strategy would really beat Tit for Tat in the original version. But he believes it would be impossible to prevent collusion between entrants.
This collusion strategy obviously wouldn't beat Tit-for-Tat if multiple entries weren't allowed. The only reason the collusion worked is because there's no cost associated with sacrificing one player to move another ahead in the rankings; allowing multiple entries that are scored separately makes it trivially easy to break the game. What would be far more interesting would be to allow as many entries as desired, but to then rank each team by the average score of all its entries. In such a system, Southampton's strategy would be a clear loser, just as communism always is.

(HT: GeekPress.)



Email blogmasterofnoneATgmailDOTcom for text link and key word rates.

Site Info