Understanding the statistical properties of performance indicators in Australian Football

This study evaluates the statistical properties of commonly used performance metrics in Australian Football (AF) using the meta-metric framework of discrimination and stability. Drawing on data from 1,383 players across 1,823 matches and 9 seasons (2015-2023), 33 player and team-level metrics were analysed to assess their statistical properties and utility for performance evaluation. Results indicate that count-based metrics such as kicks, handballs, disposals, and goals exhibit high discrimination and stability, making them robust indicators of player performance. In contrast, percentage-based metrics (e.g. effective clearance %, effective kick % and effective disposal %) demonstrated lower discrimination and stability, particularly at the team level, due to contextual variability. The advanced performance metrics Ranking Points, and Rating Points also had high discrimination and stability for all positional groups, suggesting their value in capturing overall player performance. Positional analysis revealed that discrimination and stability vary across position, underscoring the importance of contextual interpretation. This study advances the understanding of performance indicators in AF and establishes reference values to inform metric selection for decision-making and performance evaluation. These findings support the use of stable and discriminative metrics for decision-making in professional sport, while highlighting the need for further research integrating contextual data.