Abstract
OpenSpiel is a collection of environments and algorithms for research ingeneral reinforcement learning and search/planning in games. OpenSpiel supportsn-player (single- and multi- agent) zero-sum, cooperative and general-sum,one-shot and sequential, strictly turn-taking and simultaneous-move, perfectand imperfect information games, as well as traditional multiagent environmentssuch as (partially- and fully- observable) grid worlds and social dilemmas.OpenSpiel also includes tools to analyze learning dynamics and other commonevaluation metrics. This document serves both as an overview of the code baseand an introduction to the terminology, core concepts, and algorithms acrossthe fields of reinforcement learning, computational game theory, and search.