Affiliation:
1. UFMG Soccer Science Center, Sports Department, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil
2. Chair of Performance Analysis and Sports Informatics, Technical University of Munich, Munich, Germany
Abstract
This study identified which key performance indicators that predict the probability of winning are similar and specific for the top CONMEBOL and UEFA leagues. The analysis covered 12,289 games from the 2017/2018 to 2021/2022 seasons, excluding draws to focus on decisive results. Performance metrics were extracted using the InStat Scout® web platform, evaluating 55 offensive and defensive performance variables from the CONMEBOL and UEFA leagues. For data analysis, binary logistic regression was performed with the Forward Stepwise method. 14 similar predictive variables were identified between the CONMEBOL and UEFA leagues. We also identified 6 win-predictor variables specific to CONMEBOL, 2 win-predictor variables specific to UEFA, and 1 contextual variable that was predictive of victory for the CONMEBOL and UEFA leagues. The KPIs that are associated with an increased likelihood of winning games were identified. These include offensive actions, such as creating scoring chances, entering the final third of the opponent's field, shooting on goal, and being effective in counterattacks, converting penalties, and direct free kicks and making important passes and passes. For defensive actions, it was found that fouls, recoveries, possession interceptions, and low block defensive pressures are also associated with an increased likelihood of winning in the five main CONMEBOL and UEFA leagues. Particularly, for CONMEBOL leagues, counter-attacking actions, ball possession (%), free-kicks and efficiency for high pressing can also be associated with increased probability of winning. Specific to UEFA leagues, advancing actions down the left flanks and efficiency for team pressing are associated with increased probability of winning.