Evaluation of the Elo Rating System Based on the Results of the 44th Olympics in Chennai

Abstract

The characteristics of a professional chess player is described by the so-called Elo rating which is supposed to reflect the strength of the player, and is updated based on his or her performance, i.e. their wins, losses and draws with other players. A win or draw with higher-rated player means the increase in the rating of the given player, while a loss or draw with lower-rated player decreases his or her rating position. These gains and losses in the rating system are calculated according to a specific algorithm. The Elo rating is supposed to predict the expected score of a player having a given advantage or disadvantage over his or her opponent. The question is, however, how good these predictions are? Therefore, we want to investigate the accuracy of the Elo model, based on the results of the 44th Olympics in Chennai.