Probability matching and reinforcement learning

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

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 psychologists and experimental economists. We provide an evolutionary foundation for this phenomenon by showing that learning by reinforcement can lead to probability matching and, if the learning occurs sufficiently 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 specification such that the behavior is optimal for all environments where counterfactuals are observed.
Original languageEnglish
Pages (from-to)17-21
Number of pages5
JournalJournal of Mathematical Economics
Volume49
Issue number1
DOIs
Publication statusPublished - Jan 2013

Fingerprint

Reinforcement learning
Reinforcement Learning
Reinforcement
Specification
Specifications
Learning

Cite this

Probability matching and reinforcement learning. / Rivas, Javier.

In: Journal of Mathematical Economics, Vol. 49, No. 1, 01.2013, p. 17-21.

Research output: Contribution to journalArticle

@article{14abcb70833441be8872caaf5ce03a3b,
title = "Probability matching and reinforcement learning",
abstract = "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 psychologists and experimental economists. We provide an evolutionary foundation for this phenomenon by showing that learning by reinforcement can lead to probability matching and, if the learning occurs sufficiently 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 specification such that the behavior is optimal for all environments where counterfactuals are observed.",
author = "Javier Rivas",
year = "2013",
month = "1",
doi = "10.1016/j.jmateco.2012.09.004",
language = "English",
volume = "49",
pages = "17--21",
journal = "Journal of Mathematical Economics",
issn = "0304-4068",
publisher = "Elsevier",
number = "1",

}

TY - JOUR

T1 - Probability matching and reinforcement learning

AU - Rivas, Javier

PY - 2013/1

Y1 - 2013/1

N2 - 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 psychologists and experimental economists. We provide an evolutionary foundation for this phenomenon by showing that learning by reinforcement can lead to probability matching and, if the learning occurs sufficiently 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 specification such that the behavior is optimal for all environments where counterfactuals are observed.

AB - 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 psychologists and experimental economists. We provide an evolutionary foundation for this phenomenon by showing that learning by reinforcement can lead to probability matching and, if the learning occurs sufficiently 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 specification such that the behavior is optimal for all environments where counterfactuals are observed.

UR - http://www.scopus.com/inward/record.url?scp=84872268565&partnerID=8YFLogxK

UR - http://dx.doi.org/10.1016/j.jmateco.2012.09.004

U2 - 10.1016/j.jmateco.2012.09.004

DO - 10.1016/j.jmateco.2012.09.004

M3 - Article

VL - 49

SP - 17

EP - 21

JO - Journal of Mathematical Economics

JF - Journal of Mathematical Economics

SN - 0304-4068

IS - 1

ER -