Elo Ratings for Large Tournaments of Software Agents in Asymmetric Games

TL;DR

This paper proposes a revised Elo rating system tailored for large tournaments of AI agents in asymmetric games, addressing differences from human players and complex game setups.

Contribution

It introduces modifications to the Elo system to better evaluate AI agents in asymmetric and complex game environments, with guidelines for tournament implementation.

Findings

Revised Elo system accommodates AI training and game asymmetry

Guidelines improve fairness and accuracy in AI tournament ratings

Addresses challenges of large-scale AI evaluation in complex games

Abstract

The Elo rating system has been used world wide for individual sports and team sports, as exemplified by the European Go Federation (EGF), International Chess Federation (FIDE), International Federation of Association Football (FIFA), and many others. To evaluate the performance of artificial intelligence agents, it is natural to evaluate them on the same Elo scale as humans, such as the rating of 5185 attributed to AlphaGo Zero. There are several fundamental differences between humans and AI that suggest modifications to the system, which in turn require revisiting Elo's fundamental rationale. AI is typically trained on many more games than humans play, and we have little a-priori information on newly created AI agents. Further, AI is being extended into games which are asymmetric between the players, and which could even have large complex boards with different setup in every game, such as commercial paper strategy games. We present a revised rating system, and guidelines for tournaments, to reflect these differences.