Thursday, March 28, 2002

POP SCIENCE EXPLAINS IT ALL: Financial Times article pointing out that the game theory taught in business school reeks of faddishness:

Recent work by Kesten Green, a researcher at the Victoria University at Wellington, New Zealand, supports the belief that game theory is a poor way of trying to predict the outcome of complex situations.

Mr Green recruited a number of game theory enthusiasts and asked them to predict the outcome of six bargaining situations. The scenarios he described were thinly disguised adaptations of real-life conflicts: a wrangle between artists and government over financial support; a conflict between a pharmaceuticals company and the consumers of one of its drugs; a spat between sports team owners and players over broadcast revenue rights.

To set a benchmark, he also recruited students and asked them to predict the outcome of the same scenarios, based only on "unaided judgment", or common sense.

The game theorists proved no more proficient than the students. In some situations, game theory was less accurate than unaided judgment. Students asked to role-play the scenarios - "you be the union boss and I'll be the chief negotiator" - were the most accurate of the lot.

It's probably stating the obvious, but the article points out the vast gulf between game theory taught in mathematics departments and game theory taught in business schools. Is there any good science fiction based on game theory? I'd like to know; as a theory it's never been something that set my imagination on fire but I'm sure somebody's turned it into imaginary-world gold.

No comments: