What drives success in FIBA U18 women’s EuroBasket? A machine learning analysis of key performance indicators

This study aimed to investigate the game-related statistics influencing winning probabilities in the Under-18 Women’s EuroBasket Championship. The official FIBA game-related statistics of 348 games played in Division A (n = 168) and Division B (n = 180) Under18 Women’s EuroBasket Championships in 2022 (n = 117), 2023 (n = 116) and 2024 (n = 115) were analysed. Games were classified on the basis of playing level (Division A vs. Division B), round (final stage vs. group phase) and, using a two-step cluster analysis, as close games (n = 172, mean ± SD score difference = 6.8 ± 3.5 points), balanced games (n = 115, mean ± SD score difference = 20.1 ± 4.7 points), and unbalanced games (n = 61, mean ± SD score difference = 41.9 ± 10.6 points). A decision tree using the Classification and Regression Tree (CART) approach was adopted, with the outcome as the dependent variable, and game-related statistics, division, round and cluster as the independent variables. No effect for division, round or cluster was found, whereas the field goal percentage (FG%) emerged as the primary game-related statistic representing the most decisive variable at the first node of the decision tree. Additionally, a combination of steals (STL) and defensive rebounds (DREB) was identified as a further relevant indicator differentiating winning from losing teams in the analysed competition. Overall, the combination of FG% and STL was associated with the highest probability of winning (92%). These findings suggest that Under18 Women’s EuroBasket coaches should emphasize offensive drills to improve FG% and defensive drills to increase STL and DREB, thereby enhancing winning chances.