IPDToolkit: An R package for simulation and Bayesian analysis of iterated prisoner’s dilemma game-play under third-party arbitration

Abstract

Recently, researchers have begun studying the role that third-party arbitration may play in the evolution of cooperation. Using the iterated prisoner’s dilemma (IPD), they show that arbitration can mitigate the negative effects of perception errors on the stability of cooperative strategies. Open questions, both theoretical and empirical, however, remain. To promote research on the role of third-party arbitration, we introduce an R package, IPDToolkit, which facilitates both simulation of synthetic data and Bayesian analysis of empirical data. To address theoretical questions, IPDToolkit provides a Monte Carlo simulation engine that can be used to generate play between arbitrary strategies in the IPD with arbitration and assess expected pay-offs. To address empirical questions, IPDToolkit provides customizable, Bayesian finite-mixture models that can be used to identify the strategies responsible for generating empirical game-play data. We present a complete workflow using IPDToolkit to teach end-users its functionality.

Keywords

• Prisoner’s dilemma
• Arbitration
• Game theory
• Behavioral economics
• Bayesian analysis