Sunday, July 05, 2015

Peyton fest

I write to thank the organisers of Peyton Young’s 70th birthday symposium, held in Oxford last week. The great thing about birthday symposia is that they are not rational in the economic sense (they do not make money) so they feel like taking a stand against the bean-counting mentality. Through celebrating the work of one of the research community’s well-known figures, we celebrate the community itself, and the intrinsic interest of the topic. Here’s a couple of thoughts on rationality. Vince Crawford’s talk made a point that explains one reason why game theory usually focuses on rational agents rather than boundedly-rational ones: I am tempted to call it the Anna Karenina principle: that rational agents are all alike, but every irrational agent is irrational in its own way. Hence, the study of irrational agents becomes diffuse, with a difficulty to reach agreement on what kind of irrationality to allow. (e.g.: I care about what values of ε one could compute ε-Nash equilibria in polynomial time, and one could argue endlessly about what constant value of ε might be considered acceptable. This is why a PTAS would be so nice.) Prior to hearing that point being made, when asked to explain why economists focus on rational agents, I’d note the basic requirement of economic mechanisms that they should function well in the presence of rational agents, and in general it’s too much to ask that they should perform well in the presence of irrational ones. The other item: in Peyton’s speech at the end, he considered possible future directions for game theory, and proposed that adaptability to changing conditions may become the new rationality: an agent should be judged on how well it handles change rather than whether it best-responds to a fixed environment.

group photo

and the programme