Coaches and sports scientists are looking for a way to predict performance in complex team sports such as basketball. However, concerning knowing what type of player`s profile is needed to win the competition, there is not too much information in the literature. Hence, our study had two aims: (i) to identify how the individual game-related statistics discriminate between winning and losing among different player positions through a cluster analysis; (ii) to elaborate predictive models that explain better performance through a decision tree analysis. 335 matches of the men`s Spanish League 2018/2019 were analysed, with a total of 7,345 individual statistics performances. The cluster analysis identified 3 performance groups formed by foreigners with both low (FLC; 23.8% shooting-guards) and high contributions (FHC; 32.1% centres) and Spanish with low contribution (SLC; 32.9% shooting-forwards). The decision tree analysis revealed that having players of SLC and FHC profiles predicts better results in the competition. Coaches can apply these profiles to build team composition.
© Copyright 2023 International Journal of Performance Analysis in Sport. Taylor & Francis. All rights reserved.