Soon over, the “glorious uncertainty of sport” ? We are not there yet, but it is clear that over the last 20 years, the statistical study has changed the approach to sport, and football is obviously no exception. Cameras, sensors, GPS… The tools capable of dissecting performance, collective as well as individual, have multiplied. And whether they are actors or observers, journalists or analysts, all (or almost) now have recourse to it. With different aspirations but a similar objective: to try to rationalize the most irrational of sports.
What is data?
The concept of data may still seem hazy to most of us, so let’s define it first. Florent Toniutti, now editorial manager at Coparena and video analyst, created the first site dedicated to tactics in France. According to him, the data is similar to “all the elements that can be gathered around performance. It can be match data or physical data used to evaluate a player, his form on and off the field”. This is just as much raw data (distances covered, number of high intensity sprints, action zones, shot conversion rate, etc.) as it is graphic elements and more and more videos.
So what does data actually bring to those who use it? “This will serve to reduce the part of the hazard”, estimates Philippe Doucet, journalist at Canal +, and author of The palette told by its inventor *. “For the coach who wants to control everything, down to the hyper detail on which the match will probably not be played, it is a way to optimize his own team”.
Where does the data come from?
It is difficult to estimate precisely when the beginning of the generalized use of data dates back, especially since it differs according to sports, cultures or even geographical areas. Billy Beane, general manager of the Oakland Athletics baseball franchise – who inspired the movie The strategist, released in 2011 – was one of the first to rely on sabermetry, which consists in basing its recruitment on a purely statistical approach, in order to optimize the ratio between the cost of the players and the results obtained. In cycling, the hegemony of Team Sky during the 2010s finds part of its origin in the crossing of data allowing to optimize both the race and its preparation (developed power, pedaling cadence, heart rate), with in particular the use of sensors.
And in football then? “When I released the first data, the first pallets, very few clubs were using it”, remembers Philippe Doucet, who launched the famous tool 20 years ago on Canal +. “It was minimal, both in France and abroad. Gérard Houllier was one of the first to want to develop around data. Elie Baup, in Bordeaux, was working a little on it but light years away from what ‘we could offer, with VHS, rude editing … Canal had technical means that the clubs did not yet have “.
An impact on the club model
Now, the technical staff of professional clubs all have a unit (more or less extensive) dedicated to analysis. “Before, a staff, it was a trainer and an assistant who took charge of the analysis but also the physical preparation”, continues Philippe Doucet. “Today, a staff is at least 8 people, including the video analyst. There are even clubs where the analyst is on the bench, and will be able to act live during the match, intervene at the part-time in collaboration with the trainer. There are new jobs that have appeared, with coaches who are increasingly specialized in a very specific performance data “.
Optimizing your performance, analyzing that of your opponent, that’s when it comes to the field. But now, and more and more, data is also inviting itself into the offices of clubs and more particularly of recruitment units. With an approach that “varies from club to club”, explains Enzo Djebali, deputy head of the Stade de Reims recruitment unit. “For some, data provides information that will then be confirmed by the field. Personally, I believe that data comes later, that it is intended to confirm or deny what your eye sees. I first ask questions. of football with figures. But in all cases, there is the need to cross-reference the data available, whether it is between them, to give them a real value, or with this human gaze “.
Far from opposing ideologies, between an approach which would be exclusively human and the other purely scientific, the elite clubs adopt an increasingly diversified approach. Liverpool, for example, has been relying for a few years now on a full-fledged research hub with a dedicated director, Ian Graham, analysts and even more surprisingly, an astrophysicist, with the aim of establishing potential targets as well as possible. “We can put the return to the foreground of Liverpool in perspective with their pole of analysts, says Florent Toniutti. The biggest clubs develop their own models in-house, with algorithms that match their way of playing and their project. Because when you’re Klopp’s Liverpool, you don’t necessarily need the same players as Guardiola’s City “.
Other clubs go even further by putting this data-recruitment at the heart of their sports policy. This is how Brentford, a small club in London until then relatively anonymous, was able to make two consecutive climbs until reaching the Premier League this season, while selling in the process for 100 million pounds of players. His secret? A former club president who bet on almost everything on this very “scientific” approach to recruiting, like other clubs such as the Danes of Midtjylland or the Germans of Hoffenheim.
And now ?
By helping to find these famous “small details that make the big differences”, it is obvious that data has contributed to the professionalization of sport as a whole, and made certain practices evolve. But to what extent will data continue to shape the football of tomorrow? “We are still at the embryonic stage in its large-scale use, underlines Enzo Djebali. We have not yet explored all the football questions that the data could answer. US sports such as basketball or baseball have, for example, tools that allow us to assess the potential for integration of a player into a style of play, via a score. Football is only at the beginning of advanced statistics, with a level of data use that would be similar to what the NBA experienced in the mid-2000s “.
A “delay” which can be explained by technological reasons, by a different cultural approach, but also the essence of the football discipline, a “fluid” sport unlike US sports, much more sequenced and offering phases of play much less diverse, and therefore easier to study. A specific feature that should prevent football from switching to a standardized form that some fear seeing appear with the proliferation of data. Because as Florent Toniutti reminds us, “We will not be able to standardize the game, because it is the technical quality and the nature of the players that prevails over the models. We can prepare our players very well for all situations, but once on the pitch, it is them and their success which makes the match “.
* The palette told by its inventor, 366 pages, Editions Solar
