We analyse data from the Euroleague, between the 2000-1 season and the 2021-2022 season. Our aim is to identify the game-related statistics which discriminate between successful teams from unsuccessful ones, at a season-level, where success is identified as either (i) qualification to playoffs, or (ii) participation in the Final Four. We use both statistical and machine learning methods for the analysis. A logistic regression model identified, in decreasing order of significance, points per possession, defensive rebounds, possessions per game, fouls drawn, successful free throw percentage, offensive rebounds percentage and steals, as the most important factors regarding qualification to playoffs. For the Final Four, the key determinants were effective field goal percentage, total rebounds, pace, fouls drawn, turnovers, steals and blocks. In particular, we have found that the number of possessions per game and pace are negatively correlated with a team's season success. Among the machine learning methods, Support Vector Machines gave the best results for classifying the teams. We also discuss how the top performances of individual players affect season-long success of their teams. Some very high individual performances came from players whose team did not proceed further in the competition. For a team to qualify for the subsequent phases, it is not enough to have one player with a high scoring average or performance index rating.
© Copyright 2025 Journal of Sports Analytics. IOS Press. Alle Rechte vorbehalten.