dp11-20.pdf (160.61 kB)
Download file

Probability Matching and Reinforcement Learning

Download (160.61 kB)
report
posted on 20.04.2011, 13:58 by Javier Rivas
Probability matching occurs when an action is chosen with a frequency equivalent to the probability of that action being the best choice. This sub-optimal behavior has been reported repeatedly by psychologist and experimental economist. We provide an evolutionary foundation for this phenomenon by showing that learning by reinforcement can lead to probability matching and, if learning occurs su ciently slowly, probability matching does not only occur in choice frequencies but also in choice probabilities. Our results are completed by proving that there exists no quasi-linear reinforcement learning speci cation such that behavior is optimal for all environments where counterfactuals are observed.

History

Publisher

Dept. of Economics, University of Leicester

Available date

20/04/2011

Publisher version

http://www.le.ac.uk/ec/research/discussion/Papers2011.html

Book series

Papers in Economics;11/20

Language

en

Usage metrics

Categories

Exports