February 25, 2005
Lifelike Characters Who Ultimatum Bargain
Colin Camerer of Caltech gave the 7th Annual Pinkel Endowed Lecture on Mind/Brain Paradigms at Penn today, describing his work in behavioral game theory and neuroeconomics.
Behavioral game theory is an interest of mine, and an interesting new line of inquiry, theory, and experiment, but my excuse for posting about the lecture is an application of BGT that Camerer mentioned: the development of “lifelike” simulated opponents for training environments, to provide realistic practice in strategic thinking. (I assume it might have been too confusing to mention “computer games” in a talk about game theory.) While this was just a “future work” line on the last slide, it’s a worthwhile idea and might prove useful for BGT as well as computer games. Work along these lines would not just be good for an eBaybot or in-game equivalent; it could improve all sorts of decisionmaking based on value and preference. With good parameterized models for how people make strategic choices, we even could move to present culturally differentiated agents, making important decisions much as experiments in different cultures have shown that people do.