Abstract
The analysis of box-score performance indicators has traditionally been used to classify player roles in women's basketball based on the five conventional positions: point guard, shooting guard, small forward, power forward, and center. However, this framework may not reflect the current tactical and functional demands of the game. The aim of this study was to identify and redefine functional player roles in professional women's basketball using performance data derived from actual competition. A total of 36,204 individual player records from 3,894 games in the Spanish Liga Femenina Endesa (2012-2022) were analyzed. Game-related statistics were normalized by effective playing time and scaled to a 40-minute format. One-way ANOVA revealed significant differences across traditional positions, but also indicated considerable functional overlap. Unsupervised learning techniques, including k-means and Gaussian mixture models, were applied to identify underlying performance-based player profiles. The analysis yielded nine stable and interpretable functional roles, offering a more nuanced classification than the traditional five-position model. These roles capture offensive, defensive, and hybrid specializations, providing coaches and analysts with a practical framework for tactical planning, scouting, and individualized player development. The findings support a shift toward data-driven classification systems that better reflect the functional realities of modern elite women's basketball.