A1254
Title: Rating of players by Laplace approximation and dynamic modeling
Authors: Robert Hsuan-Fu Hua - Florida State University (United States) [presenting]
Abstract: The Elo rating system is a simple and widely used method for calculating players' skills from paired comparison data. Many have extended it in various ways. Yet the question of updating players variances remains to be further explored. The issue of variance update is addressed by using the laplace approximation for posterior distribution, together with a random walk model for the dynamics of players strengths, and a lower bound on players variances. The random walk model is motivated by the Glicko system, but nonidentically distributed increments are assumed to take care of player heterogeneity. Experiments on men's professional matches showed that the prediction accuracy slightly improves when the variance update is performed. It also showed that new players strengths may be better captured with the variance update.
