Comparing player rating systems as a metric for assessing individual performance in soccer

Abstract

This study investigates associations between soccer player ratings and performance metrics while comparing three major rating systems: WhoScored, FotMob, and Sofascore. Using data from over 2,100 players across two seasons in top European leagues, generalised linear mixed models assessed the influence of 73 performance metrics on ratings. Offensive metrics, such as shots on target (β = 0.21 WhoScored, β = 0.16 Sofascore, β = 0.17 FotMob, all p < 0.001), key passes, and successful take-ons had the highest impact, while defensive metrics like clearances (β = 0.12 WhoScored, β = 0.06 Sofascore, β = 0.12 FotMob, all p < 0.001) played a smaller but consistent role. Comparisons across rating systems revealed systematic differences, with WhoScored ratings significantly lower overall (β = -0.20, p < 0.001). These findings underscore that rating systems capture different aspects of player performance and are not directly interchangeable. The study highlights their potential for more systematic player evaluations in scouting and development while raising questions about transparency and biases. Future research should investigate their utility in recruitment and talent identification. This research advances understanding of single-value ratings, emphasising their complexities and limitations in assessing individual performance.