Quantification is taking over the world of sport. Ruben Saavedra, a Catalan neuroscientist who runs Metrica Sports, is part of the data revolution in football. He explains why and how data matters in football to Pinaki Chakraborty
No. I always make the comparison with chess. Football is similar to chess in a lot of ways. If you move a piece it has an impact on other pieces. You can analyse all of that. Not everything can be done with data, when they scout, they also scout their personal lives. Not everything can be measured. The more information you have the better decision you can make.
How has game evolved with data sciences?
We saw while watching games what media were doing with heat maps and we thought it was very basic. A player ran 5.2 km, but then when did he run, at what speed, with ball, without ball? I am talking about 2012-13, there were no clubs in the world that had proper data science departments, none. Some were starting in England, mainly Liverpool, and in the US because there is a culture in sports of data analysis. It comes from baseball, basketball etc.
Didn’t big clubs have data analytic teams then?
Barcelona was starting, they had only one person. All top clubs nowadays have data sciences department. But if you go below those, they don’t have it yet. Clubs like Athletic Bilbao and Real Sociedad are now building data science teams. We think we are in the middle of a revolution.
How is the data used at different levels?
Top clubs have a data science department and they use the data for whatever else they want along with football. The data can be used for scouting. You can get data from any video. If they upload a video from Brazil, then they can see his (player’s) workload, physical speed and high intensity runs. They also use the data for opposition analysis: where they shift the ball, the transitions.
That’s pattern recognition.
Yes, but it is also for analysing your own game. Clubs have their own philosophy and they have their own way of going through the game and they take their time to analyse the game. These are very fine details, for example, in some situation the players should not be further away than 5 metres. In the lower leagues you don’t have data scientists, it is all based on videos. Now these clubs can add data to their video analysis.
How does data impact football philosophies of clubs?
Data can be used to see if the philosophy is actually happening. Clubs, reportedly like Barcelona and Ajax, have a fixed philosophy. They can build algorithms to detect those very specific movements: if we lose the ball here, we press for three seconds and step back, a certain player will instantly move to a position. Data will instantly tell you when this happens and when this does not. You can specify algorithms for positioning. For right backs and wingers there is a set of patterns. This enables them to compare a player who is playing u-17, u-19 and u-21 to those in the main team. They did it in the past by watching players, now they do it with data and scouts.
Can all clubs afford this?
For sure, this is not a cheap department. It also needs a change in culture. That is why the top clubs are leading the revolution. We believe that in a few years every professional club will make decisions based on data. When we started, the data trackers were bought by broadcasters and then in some clubs and leagues the broadcasters would try and sell this data. Now, it is the clubs that are buying the data directly.
How do they use the data for scouting?
Twenty years back all teams had a crew. They had 20-50 people travelling around the world watching games. They would send a report and then if the player was interesting, they would send another analyst. Now, data is used to support the decision. A club once told me that they were scouting a player in Brazil and all videos that they had was that he was superfast. Until we got the data. He was fast compared to Brazilian average, was low compared to our average. Now, it’s come a lot further. Data can tell you whether the player will fit into your team. So you buy a defensive midfielder and look at the data and find how many tackles he makes, how many recoveries and where he sprints. Then you put that data in your model and take out your defensive midfielder and put this one in. The more information you have, the better decision you can make.
Tell me what does your firm do?
We provide video and data analysis to football clubs. We provide it at all levels. The idea is to provide video and data analysis to football coaches, players, analysts, fans at all levels. We have an AI based tracking technology. We can track any football video whether it is training, from broadcast, from a hand held camera, anything. The data that is generated is the difference. We provide clubs like PSG or other clubs with highly sophisticated data, event data i.e. all ball events like passes, shots, tackles, all events during a game.
Is it because the so-called not well off clubs cannot afford expensive softwares like yours?
For sure, this is not a cheap department. It is not just the people, but the data is very expensive. It also needs a change in culture and like any innovation there are early movers. Thos early movers are those who have enough money to “waste” or give . That is why the top clubs are leading the revolution. It is a matter of time when the low clubs in Spain, England and Germany, Holland now have the budgets to have such a department. When we started the data trackers were bought by broadcasters and then in some clubs and leagues the broadcasters will try and sell this data.
You spoke about democratising football. What is the Tesla model that you follow?
What I meant was Tesla launched a very expensive product in the beginning, which only some privileged customers could pay and with that they financed some of the lower versions of the product. This is our model. When we founded the company in 2014, we only had the elite plan. We only worked with big clubs like the Reals, the Valencias, the Spanish national football team and covered the MLS, the American soccer league. That was our company back then. We wanted to democratise access to these kind of solutions for everyone now.
Views expressed above are the author’s own.
END OF ARTICLE