Professor Jesse Davis and his team at KU Leuven in Belgium have transformed soccer analytics using machine learning models that reveal subtle tactical advantages, including seemingly counterintuitive moves like kicking the ball out of bounds on the opponent’s side to create scoring opportunities.
- 1.4 million passes and 60,000 throw-ins analyzed to identify new tactical patterns
- Kicking the ball out near opponents' goal can position teams within striking distance
- Open-source tools and standardized data are expanding soccer analytics adoption
What happened
Jesse Davis, a computer science professor at KU Leuven in Belgium, leads the university’s Sports Analytics Lab, which has become a leading force in applying AI and machine learning to soccer. By assembling an extensive dataset that includes over 1.4 million passes and 60,000 throw-ins from major competitions such as the 2022 World Cup, Davis and his team have uncovered tactical insights that change how teams view seemingly non-intuitive plays.
One such finding revealed that deliberately kicking the ball out of bounds on the opponent’s side when the ball is in the middle third can strategically position a team to recover possession closer to goal within about ten subsequent actions. This challenges conventional soccer strategies and offers pro clubs new ways to gain competitive advantages by leveraging data-driven decision-making.
Why it matters
Soccer traditionally relies on experience, intuition, and conventional tactical wisdom, but the KU Leuven lab’s work introduces a new dimension by quantifying the effects of specific in-game actions. With matches featuring thousands of actions yet relatively low scoring, even small tactical advantages can significantly impact outcomes in high-stakes competitions such as the World Cup or elite European leagues.
Moreover, Davis’s commitment to open-source tools and data standardization initiatives addresses a key barrier in sports analytics: disparate data formats and limited accessibility. By providing freely available resources, KU Leuven facilitates broader adoption of advanced analytics, allowing clubs of varying sizes to incorporate sophisticated AI models into scouting, strategy development, and player performance evaluation.
What to watch next
KU Leuven’s continued development of standardized in-game data protocols is expected to streamline analysis across leagues and improve the reliability and comparability of soccer analytics worldwide. This development could catalyze widespread use of AI-driven insights during live matches and in training environments.
Additionally, the impact of these data-driven tactics on coaching styles, player development, and even fan engagement will be closely monitored. As more clubs integrate machine learning insights into their decision-making processes, the dynamic between traditional soccer intuition and analytical strategy will evolve, potentially redefining the sport’s future competitive landscape.