The sport which we all love watching on platforms like hesgoal is getting more and more complicated and in-depth behind the scenes. In the heart of Belgium, at one of the world’s oldest universities, a small corner is shaping modern football. The Sports Analytics Lab at the Catholic University of Leuven has been instrumental in understanding how the game functions today.
The lab is headed by Jesse Davis, a computer science professor from Wisconsin who discovered his passion for the beautiful game during the 2002 World Cup. Hired in Leuven in 2010 to research machine learning, data mining, and artificial intelligence (AI), Davis soon found himself supervising post-docs, PhDs, and master’s students working on football data.
The richness and complexity of this data lent itself well to AI studies. Their work significantly advanced the science behind sport through open-source analytics tools that changed some clubs’ approach to playing.
Overcoming Initial Challenges
In the early days of the sports analytics movement, it was believed that football didn’t lend itself well to advanced statistical analysis due to its fluid nature. Unlike baseball or basketball, where actions could be counted and assigned value easily, quantifying each player’s contributions in football was challenging due to fewer measurable actions like shots or goals.
However, as time passed by, most big clubs began to put together extensive data departments focusing on analytics, which has led to a surge in literature around this subject.
The Sports Analytics Lab published findings on optimal areas for taking long shots or whether it’s more efficient to boot the ball long than build out from the back. These insights were broken down into layperson’s terms through their blog posts, leading to collaborations with Red Bull Leipzig, Club Brugge, along with the German and United States federations, among other clubs and bodies.
The Value of Research
While there’s innovative work done within football clubs, it often remains unseen due to the competitive nature of professional sports. The scientific process requires taking risks and tinkering with experiment designs which may not yield immediate results or help a team win games in the short term.
Most basic football analytics, like shot value or types of passes producing danger, have already been explored. What remains are complex problems like tracking data and making sense of it.
For instance, while expected goal models quantify and tabulate chances a team created over a game course, they don’t work well in putting a number on a striker’s finishing ability due to biases in training data.
Moreover, when clubs find useful insights they need to implement these findings and track them over time which involves data engineering and software programming. This can take up most labour in analytics work but is crucial for long-term success.
The Sports Analytics Lab at Catholic University of Leuven exemplifies how university research can benefit the public and businesses.
Their pioneering work has not only advanced our understanding of football but also demonstrated the value academia brings even when under pressure from all sides.
