This study developed a novel methodology to assess player skill in Australian Football by decomposing Champion Data Ranking Points (CDRP) into two components: involvement (number of on-ball actions per game) and impact (CDRP per involvement). Data was collected for players from the AFL and the three Tier 2 leagues (VFL, SANFL and WAFL) between 2015 and 2023. Separate Bayesian mixed models were used to estimate each component controlling for age, position and league. Predictions from these models were combined to reconstruct CDRP and were compared against a baseline model. Results indicated that players in Tier 2 leagues generally achieved higher involvement (IRR: VFL 1.43, WAFL 1.42, SANFL 1.39) and higher impact (Coefficients: VFL 0.17, WAFL 0.21, SANFL 0.26) compared to AFL players. Rucks had a higher involvement (IRR=1.01), but lower impact (Coefficient=-0.42) than other positions. The combined models demonstrated improved predictive accuracy (MAE=18.44, RMSE=23.32) over previous research and allowed random effects to be interpreted as player skill ratings. This methodology offers significant applications for professional sporting organisations, enabling robust cross-league comparisons, age-specific evaluations and longitudinal tracking of player skill. These insights can inform recruitment strategies, player contracting and development programmes.
© Copyright 2025 International Journal of Performance Analysis in Sport. Taylor & Francis. All rights reserved.