Football and Philosophy of Statistics with Joost van der Leij

Joost van der Leij is a university philosophy lecturer specializing in philosophy of mind and action and philosophy of statistics. He fell into the world of football (what us Americans call soccer) when in 2016 he was asked to help improve the players of Hannover 96, who had just been relegated, using his methodology for mental coaching. He created a model that could track whether players actually improved and tested it during the Euro 2016 tournament. The model correctly predicated the winner of the tournament: Portugal. In the following years, he fine-tuned the model and started his company Football Behavior Management. He also runs the website, which showcases the philosophical side of football. Together with Patrick Busby, the current technical director of FC Volendam, Joost wrote a book on football statistics from the Bayesian point of view. He also teaches the Football Behavior Management course at the VU-university in Amsterdam every year. In this interview, we discuss football and the philosophy of statistics he applies to his analyses.

Greg: What led you to create an entire website about how to apply philosophical lessons to football?

Joost: Since 2016, I have been working in football using Bayesian network models (more on those models here) to predict three things:

  1. which players would do well at a specific football club in a specific league
  2. which starting eleven has the most chance to win
  3. what is the most likely league standing in three months

As I use subjective Bayesian statistics based on the work of professor Bruno de Finetti, often this needs a bit of explaining and the explanation is quite philosophical. Which is understandable, given that I am a philosopher with a background in the philosophy of statistics. Philosophy is not a very popular subject, let alone the philosophy of statistics. Outside of the philosophy department, much to my surprise, I discovered that there is genuine philosophical interest in the world of football. Working with clubs, I noticed that quite a few of the decision makers really want to understand how things hang together in order to learn how to maximize results.

Joost van der Leij, an expert in football and philosophy of statistics
Joost van der Leij

Greg: What is the overarching mission of your website

Joost: The overall mission of the website is to give me a platform where I can organize my thoughts. It is set up as an encyclopedia so that many different topics can be discussed. Most of the topics center around my broader philosophy outside of football, which is statistics, cybernetics and behaviorism.

Greg: I’m not as familiar with football as I am with other sports, but it seems like almost every sport has been impacted by revolutions in statistical analysis in the twenty-first century. What can philosophy contribute to football analysis that math and statistics cannot?

Joost: Math and statistics are also part of philosophy, so I don’t see a divide between philosophy on the one hand and math and statistics on the other hand. In fact, due to a lack of philosophical fundamentals, many people misuse math and statistics in sports. The main mistake people make, in my opinion, is their belief in truth and objectivity. While there are many respectable philosophical schools that share these beliefs, in my view, they are wrong. That is not to say that relativism is right, because relativism shares too many of the same concepts as do theories about truth and objectivity. The third way is the Bayesian pragmatism of Wittgenstein, Quine, Davidson and De Finetti. Personally, I think that the subjective approach to math and statistics is the best way for sport, especially for scouting. Of course, this is a very deep topic, and it takes ten days of explaining as I do in the Football Behavior Management course at the VU-university here in Amsterdam.

Greg: Has analyzing football from a philosophical perspective changed the way you watch matches?

Joost: Analyzing football from a philosophical perspective has only changed the way I watch matches indirectly. In order to test the Bayesian network model that I built for players, teams, matches and leagues, I have analyzed two to three matches per day for almost two years. That has had a big impact on how I watch football. For instance, I can’t stand highlights anymore. One of the biggest mistakes that scouts make is to look for the best a player has ever played, rather than look at what the player most likely will do at a different club in a different league.

Greg: What discovery has most surprised or inspired you since you started this site?

Joost: It’s hard to say what discovery surprised or inspired me the most, but highlights are:

  • the fact that people think that describing what happened during a match is the same as analyzing the match
  • the fact that people think that formations are real rather than a mental construct that is useful
  • the fact that people think that it is good idea to talk about or think in terms of the potential of a player
  • the fact that decision makers in clubs are clueless about the problem of ruin
  • the fact that expected goals (xG, more detail here) is such a useless stat that can easily replaced with Shots on Target

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.