Wednesday, January 07, 2009

CORS : Iterated Prisoner's Dilemma

As I placed and won my CORS module bid, using an insane showhand bid of 6000+ points, for probably the last time in my undergraduate times, it suddenly dawned upon me that CORS bidding was a game of iterated prisoner's dilemma.

There are many 'strategies' for CORS bidding, but probably the strategy with the most effect is to deny all opponents of additional information, i.e., to not submit any bids during the open bidding phase. Conversely, by submitting an accurate bid (a bid that is valued according to your evaluation of the worth of the module), your bid provides information to potential competitors.

Of course, if everyone is selfish and does not contribute accurate bids, then the open phase is defeated, and everyone's utility is hampered.

Expressing the above in a game table,
Now, since the results of bidding are only known after the bidding has concluded, and that students have other semesters left, this provides some kind of force against concealing your bid. The rationale is that if everyone finds out that they have been exploited due to revealing their bids, then they too could retaliate by not revealing their bids the next time round. Hence, everyone loses, and overall utility is reduced. Everyone realizes that their utility is best optimized by maintaining the status quo.

The situation, however, is disturbed by people who have no fear of retaliation, i.e. final year students. Hence, they tend to do funny things to the system, i.e. random bids, showhand bids, or just hogging multiple candidate modules before dropping them.

Problem is, if the system is indeed an iterated prisoner's dilemma, then knowing that final year students would do X in their final year, one intuitively realizes that during their 3rd year, the force of the contract is nonexistent, since "good" behavior in year 3 does not influence others to behave well in their final years. Extend this argument, and one realizes that the system is unstable.

No comments: