Abstract details
| Abstract-ID: | 1878 |
| Title of the paper: | Individual brilliance or team consistency? Evaluating player ratings as predictors of soccer success |
| Authors: | Ball, J., Varley, M.; Huynh, M. |
| Institution: | La Trobe University |
| Department: | Allied Health, Human Services and Sport |
| Country: | Australia |
| Abstract text | INTRODUCTION: Success in soccer is driven by individual performances and team cohesion (Filho et al., 2014). While player ratings provide an objective numerical measure of individual performance, their relationship with match outcomes remains underexplored. This study examines the extent to which individual performances, as reflected in player ratings, contribute to match outcome across elite-level soccer. A primary objective is to determine whether team-wide consistency (collective performance) or standout individual performances serve as the stronger predictor of match outcome. METHODS: WhoScored player ratings (a player performance score between 0-10) from three elite leagues (Premier League, La Liga, Bundesliga) between 2022-2024 were collected. Data was split into training and testing sets, and models were trained using linear mixed models. Predictions were made using the models on the testing dataset. The first model examined whether average team ratings predict match outcomes (win, draw, loss). A second model replaced the rating of standout players, determined as two standard deviations from the team average, with the team’s average rating. Models were compared to evaluate their impact on match outcome. RESULTS: Average player ratings for each team revealed that losing teams had an average rating of 6.38, drawing teams averaged 6.71, and winning teams recorded a mean of 7.1. Overall team ratings were a strong measure of performance, reinforcing the importance of collective performance. However, individual standout performances influenced specific games. Replacing a team’s highest-rated player with an average performer resulted in a slight decline in model performance when predicting match outcomes across the full dataset (92.2% vs. 91.1%). However, game-specific analyses revealed cases where key players had a decisive impact. For example, Nottingham Forest’s win probability against West Ham dropped from 96.4% to 43.2% when their goalkeeper, who saved a penalty and earned an 8.7 rating, was replaced with the average rating for that game. Similarly, Southampton’s win probability against Liverpool fell from 66.5% to 17.98% without their highest-rating player, who scored twice and received a 9.2 rating. CONCLUSION: These findings highlight the dual nature of success in soccer. While team performance is the strongest predictor of match results, exceptional individual performances can be decisive. A strong collective effort is necessary for consistent success, but teams with game-changing players may have an advantage in key moments. Identifying standout individual performances helps opponents strategize and neutralise threats. Future research should explore how playing styles and formations influence the balance between team cohesion and individual brilliance. Filho, E., Dobersek, U., Gershgoren, L., Becker, B., & Tenenbaum, G. (2014). The cohesion–performance relationship in sport: A 10-year retrospective meta-analysis. Sport Sciences for Health, 10, 165-177. |
| Topic: | Statistics and Analyses |
| Keyword I: | |
| Keyword II: | |
| Keyword III: | |