Adam Gee
Title: Investigating Experiential Effects in Online Chess using a Hierarchical Bayesian Analysis
Date: July 16th, 2025
Time: 2:00pm
Location: LIB 2020 & Zoom
Supervised by: Owen Ward
Abstract:
The winner effect in animal behaviour is the increased probability that an animal wins their next encounter after having won previous encounters. Although widely documented in animals, it has long been debated whether these experiential effects exist in human compe- tition. Detecting these effects in most sports is challenging due to limited data availability and uncontrollable extraneous factors. The recent rise in online chess popularity provides a huge quantity of data in a controlled environment. Using a Bayesian hierarchical logistic regression model, we carefully investigate the presence of experiential effects in online chess. Our findings show no evidence for a consistent experiential effect across all players, but a small proportion of individual players show evidence of possessing the effect. Our results are consistent across amateur and professional players. We perform multiple model checking procedures to show that our model is suitable for these data.
Keywords: Chess, Experiential Effects, Hot-hand, Winner/Loser Effects, Hierarchical Bayesian Modelling, Sports Analytics