A Bayesian Mixture Model approach to expected possession values in rugby league.

Abstract

This study aimed to introduce a novel Bayesian Mixture Model approach to the development of an EPV model in rugby league, which could produce a smooth pitch surface and estimate individual possession outcome probabilities. 99,966 observations from the 2021 Super League season were used. A set of 33 centres (30 in the field of play, 3 in the opposition try area) were located across the pitch. Each centre held the probability of five possession outcomes occurring (converted/unconverted try, penalty, drop goal and no points). Probabilities at each centre were interpolated to all locations on the pitch and estimated using a Bayesian approach. An EPV measure was derived from the possession outcome probabilities and their points value. The model produced a smooth pitch surface, which was able to provide different possession outcome probabilities and EPVs for every location on the pitch. Differences between team attacking and defensive plots were visualised and an actual vs expected player rating system was developed. The model provides significantly more flexibility than previous zonal approaches, allowing much more insightful results to be obtained. It could easily be adapted to other sports with similar data structures.