Winning in Premier League 2: a statistical model of technical performance indicators

Abstract

Previous research has highlighted the importance of performance indicators in determining match outcomes in men’s senior professional football, yet limited evidence exists for elite youth competitions. This study examined the relationship between match outcome (win vs not win) and technical performance indicators in the Premier League 2 and developed a binary logistic regression model to predict match winning. The sample comprised 303 matches from 2022–2024 for model training and 64 matches from 2025 for testing. Mann-Whitney U analyses revealed winning teams executed more forward/backward passes, shots, shots on target, successful crosses, and corners, while facing fewer shots and shots on target (p < 0.05). Logistic regression identified increases in shots on target, backward passes, and interceptions increased the probability of winning, whereas shots on target faced, crosses, free kicks, and passes into the final third reduced it. Variables such as successful lateral passes and offensive duels faced were retained in the model to support interpretive context despite non-significance. The final model demonstrated good predictive performance (AUC = 0.771) and generalisability to unseen matches. Findings emphasise the value of creating high-quality scoring opportunities, structured vertical passing, controlled possession, and anticipatory defensive positioning for enhancing success in elite youth football.