An Extension of Counterfactual Regret Minimization for Multiplayer Card Games

IPSJ

Description

Counterfactual Regret Minimization (CFR) [1] is one of the state-of-the-art methods for solving large imperfect-information games. It shows great performance in solving 1-to-1 poker games. But there is still little re-search about how to apply it to multi-player poker games. In this paper, we will apply CFR to an extension of poker which is played by 4 players (2-to-2), and compare its performance with random policy.

Counterfactual Regret Minimization (CFR) [1] is one of the state-of-the-art methods for solving large imperfect-information games. It shows great performance in solving 1-to-1 poker games. But there is still little re-search about how to apply it to multi-player poker games. In this paper, we will apply CFR to an extension of poker which is played by 4 players (2-to-2), and compare its performance with random policy.

Journal

Keywords

Details 詳細情報について

Report a problem

Back to top