The Speed of Learning in Noisy Games: Partial Reinforcement and the Sustainability of Cooperation

Bereby‐Meyer, Yoella; Roth, Alvin E.

doi:10.1257/aer.96.4.1029

Cited by 142 publications

(120 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The results reveal quick learning toward (36) analyzed a similar pattern in a different experimental paradigm and notice that the results reflect a payoff variability effect: High payoff variability moves choice behavior toward random choice, which is related to the observation that variance in outcomes slows learning (37). Bereby-Meyer and Roth (38) demonstrate the significance of the payoff variability effect in an experiment in which participants repeatedly played 10-period prisoner's dilemma games, i.e., in which they played 10 periods with a fixed other player, and then were matched with a new player to play 10 more periods, until they had played 200 periods, in 20 10-period games. In all conditions of the experiment, they saw what the other player had done after each period.…”

Section: Learning In Basic Choice Tasksmentioning

confidence: 71%

Maximization, learning, and economic behavior

Erev

Roth

2014

Proc. Natl. Acad. Sci. U.S.A.

142

102

View full text Add to dashboard Cite

The rationality assumption that underlies mainstream economic theory has proved to be a useful approximation, despite the fact that systematic violations to its predictions can be found. That is, the assumption of rational behavior is useful in understanding the ways in which many successful economic institutions function, although it is also true that actual human behavior falls systematically short of perfect rationality. We consider a possible explanation of this apparent inconsistency, suggesting that mechanisms that rest on the rationality assumption are likely to be successful when they create an environment in which the behavior they try to facilitate leads to the best payoff for all agents on average, and most of the time. Review of basic learning research suggests that, under these conditions, people quickly learn to maximize expected return. This review also shows that there are many situations in which experience does not increase maximization. In many cases, experience leads people to underweight rare events. In addition, the current paper suggests that it is convenient to distinguish between two behavioral approaches to improve economic analyses. The first, and more conventional approach among behavioral economists and psychologists interested in judgment and decision making, highlights violations of the rational model and proposes descriptive models that capture these violations. The second approach studies human learning to clarify the conditions under which people quickly learn to maximize expected return. The current review highlights one set of conditions of this type and shows how the understanding of these conditions can facilitate market design.decisions from experience | mechanism design | contingencies of reinforcements | experience-description gap | reinforcement learning E conomics has long used idealized models of perfectly rational individual behavior as useful approximations for explaining market institutions and other economic phenomena. Many of the successful economic inventions (e.g., trading, commodity markets, auctions, matching market institutions) seem to work in ways that are well accounted for by rational models. For example, trading is most likely to be successful when all sides benefit from the trade. Careful empirical studies show, however, important violations of the rationality assumption (e.g., refs. 1-3). Thus, it seems that the usefulness of the leading economic analyses may be enhanced by relaxing the rationality assumption and improving the descriptive value of the underlying models, in part by identifying where the perfect-rationality approximations work more or less well and the purposes for which they may be more or less useful.Much of what has come to be called behavioral economics can be described as an attempt to make economic analyses more accurate by modifying the rational model to incorporate psychological insights. The most basic rational model, the expected value rule, models people as assigning cash equivalents to possible outcomes, and then selecting th...

show abstract

Section: Learning In Basic Choice Tasksmentioning

confidence: 71%

Maximization, learning, and economic behavior

Erev

Roth

2014

Proc. Natl. Acad. Sci. U.S.A.

142

102

View full text Add to dashboard Cite

show abstract

“…Uma possibilidade de estudo futuro seria, na condição de acesso sempre à pontuação do "outro" (AS), disponibilizar os pontos de forma probabilística, como no estudo de BerebyMeyer e Roth (2006). O acesso sempre à soma da pontuação do "outro", mas com o aparecimento probabilístico da soma da pontuação do "outro" jogador poderia promover a diminuição da "cooperação" na estratégia TFT?…”

Section: Discussionunclassified

“…(1) Ainda conforme Bereby-Meyer e Roth (2006), na condição do pagamento probabilístico, o comportamento aprendido foi muito parecido com o comportamento aprendido na condição do pagamento determinístico, mas a aprendizagem no probabilístico foi mais lenta. Para os jogadores que não aprenderam a "cooperar" na condição probabilística, a diminuição da taxa de "cooperação" foi mais lenta em relação àqueles que não "cooperaram" nas condições de pagamento determinístico.…”

Section: Os Participantes Dos Três Experimentos Foram Submetidos Ao Junclassified

“…Segundo Bereby-Meyer e Roth (2006), nas condições determinísticas os jogadores aprenderam a "cooperar" cedo e a "delatar" mais tarde. Na condição probabilística houve muito menos aprendizagem em "cooperar" nas tentativas iniciais.…”

Section: Os Participantes Dos Três Experimentos Foram Submetidos Ao Junclassified

“…Nas duas primeiras condições, SEM ACESSO e ACESSO SEMPRE, à soma da pontuação do "outro" e as características determinísticas da estratégia TFT serão mantidas. Porém, na terceira condição, o fato do acesso ser randômico, haverá uma semelhança com as características probabilísticas das interações sociais, descritas nos experimentos conduzidos por Baker e Rachlin (2001) e Bereby-Meyer e Roth (2006), pois nem sempre os participantes terão acesso aos ganhos dos "outros", como conseqüência de tais interações. "Humanos freqüentemente emitem respostas que produzem resultados do seu e do desempenho do outro".…”

Section: Os Participantes Dos Três Experimentos Foram Submetidos Ao Junclassified

See 2 more Smart Citations

Efeitos do tipo de acesso à soma da pontuação do outro jogador na emissão de respostas "cooperativas" no jogo dilema do prisioneiro repetido

Faleiros¹

View full text Add to dashboard Cite

The Role of Subsidies in Coordination Games with Interconnected Risk

Gong

Heal

Krantz

et al. 2014

Behavioral Decision Making

View full text Add to dashboard Cite

Can subsidies promote Pareto‐optimum coordination? We found that partially subsidizing the cooperative actions for two out of six players in a laboratory coordination game usually produced better coordination and higher total social welfare with both deterministic and stochastic payoffs. Not only were the subsidized players more likely to cooperate (choose the Pareto‐optimum action), but the unsubsidized players increased their expectations on how likely others would cooperate, and they cooperated more frequently themselves. After removal of the subsidy, high levels of coordination continued in most groups with stochastic payoffs but declined in deterministic ones. This carry‐over disparity between the deterministic and stochastic settings was consistent with the economic theories that agents were more likely to keep the status quo option under uncertainty than without uncertainty. Hence, players with stochastic payoffs were more likely to keep the high coordination level (status quo) brought by the subsidy in the previous subsidy session. A post‐game survey also indicated that with stochastic payoffs, players focused on risk reduction. Temporary subsidies promoted lasting coordination because even after subsidy was removed, players still assumed that others players would prefer reduced risks from cooperation. With deterministic payoffs, however, the subsidy might crowd out other rationales for coordination, with many players indicating that the subsidy was the only reason for anyone to cooperate. Hence, the coordination level dropped when the subsidy was removed. Copyright © 2014 John Wiley & Sons, Ltd.

show abstract

The Speed of Learning in Noisy Games: Partial Reinforcement and the Sustainability of Cooperation

Cited by 142 publications

References 39 publications

Maximization, learning, and economic behavior

Maximization, learning, and economic behavior

Efeitos do tipo de acesso à soma da pontuação do outro jogador na emissão de respostas "cooperativas" no jogo dilema do prisioneiro repetido

The Role of Subsidies in Coordination Games with Interconnected Risk

Contact Info

Product

Resources

About