Two of the best all-time coaches in Dutch football, Louis van Gaal and Johan Cruijff argued about this for years already. How relevant are statistics in world football nowadays? Is the human eye still as important as it once was, when looking at the potential of a young talent? Cruijff argued that computers cannot see how a player controls and touches the ball, while Van Gaal is very much in favour of using modern techniques to see what players should improve on.
It is obvious that statistics are gaining more and more ground in modern football. Websites such as whoscored.com, OptaSports and Squakwa are getting more popular. Opponent set pieces such as corner kicks are heavily studied and analysed prior to matches and penalty shootouts are prepared before important cup matches as well. In the video below, you can see the latest final of the Johan Cruijff Cup, in which you can see that Feyenoord players continually tell goalkeeper Bijlow where the PSV player usually shoots their penalties. Bijlow guesses the right corner almost every time, even when the player still scores.
SciSports, a Dutch firm founded in Enschede, started a statistics firm with the intention to help clubs scout for new players that would not be possible if scouting was done the old-fashioned way: based on intuition. Their database features over 300.000 active football players, have deals with several Dutch clubs to do Data Intelligence Analysis (e.g. Heracles has 13 camera’s available for SciSports for 3D-analysis) and in 2016 they assisted with 13 transfers in the Eredivisie. A prime example of their scouting prowess is Wout Weghorst, who had to leave then Second Division side FC Emmen because he was deemed not good enough; but SciSports advised Eredivisie club AZ to sign him and he has now made his debut in the Dutch national team as well, playing three matches.
SciSports recently also acquired deals with clubs outside of the Netherlands, such as clubs in England and recently they signed a deal with French giant Olympique Lyonnais.
So in the end, who is right? Can computers scout players that humans would not be able to do? And vice versa? Would the 13-year old, growth-hormone deficient Lionel Messi have been scouted by SciSports’ technology? I have always been on Cruijff’s side of the debate, but technology is only getting better and better..
Sources:
https://www.ad.nl/nederlands-voetbal/nederlands-scisports-verandert-transferwereld~a989ef20/
http://www.scisports.com/
I think it is very interesting to think about how new technologies can or will influence the sports industry. I believe using algorithms for predicting opponents moves and for predicting talent are not far from being the determining winning factor across all sports. But that is not all. Already this year the goal-line technology and video assistant referees changed the dynamics of the soccer world club. Games had to be stopped several times so that the referee could check the cameras and conclude the right decisions. Of course, these technologies improve the fairness but do they not decrease the dynamics of a lot of games as well as the emotions which make soccer games so interesting? But since the sport industry is, like any other industry, determined by profits, I believe that we cannot stop the move towards being more dependent on new technologies.
Thank you for this interesting perspective on the application of sports analytics in Dutch football.
From my point of view, having more data-based insights is never wrong. They are free of emotions and can provide a good grasp of a player’s delivered performance. However, future predictions derived from statistical models should always be taken with a grain of salt, since what they show are merely possible tendencies that might occur with a higher likelihood. While it is true that analytics methods are getting better and better and data can nowadays be gathered at any desired level of granularity, the biggest caveat of models is still that they can never capture the whole picture. At the end of the day, players are not machines and performance can change dramatically even within a single season. The rise of players like Weghorst and Messi provides the best examples, but it can also easily go the other way if we just look at how many athletic careers have already been ruined by personal tragedies or simply media pressure. Especially in sports like football, a player’s performance also heavily depends on his attitude towards the game, his capability to integrate himself into the team, and how well he gets along with the coach. Factors like these are game-changing as they can determine whether the player can actually unfold his potentials on the field. So far, it is impossible to capture these sides of a player with data. And it probably also doesn’t make sense to do so – oftentimes, a quick glance at the player already tells the experienced coach enough to know whether the candiate is blessing or curse for the team. But since player transfers in major football leagues are expensive decisions, it is still advisable to look at cold-hearted rational data to not get fooled by media hypes and first impressions.
To sum it up, purely relying on either analytics or intuition is both wrong. It is rather a combination of both that can provide the best outcomes.
I think that the use of statistics can help the scouting improve, but only for a bit. It has to become a standard tool next to other methods a scouting department uses. As Pieter Zwart (a journalist highly involved in football statistics) argues: good use of statistics can make it possible for Excelsior to become 14th instead of 15th, but won’t get them in the upper half of the table. But the switch from 15th to 14th might actually make the investment worth it. The problem in the football world is that it’s conservative. This is also a problem that SciSports (which is a great example in this case) suffers. To make progress with data in the Dutch league, the clubs have to make a deal and follow the example of Heracles: putting those cameras everywhere. The NBA is a great example of where this succeeded.
To conclude, I don’t think statistics can take over the whole scouting department (and find the new Messi). For that to happen, the game is too difficult, because every second multiple unique situations occur. But the role of statistics will increase significally in the future; if the clubs are open and ambitious enough to invest in it.
This is a very interesting article! If I had to make a prediction about the impact of statistics in the football world, then I would say that data based approaches to recruiting footballers will lead to a leveling of the playing field. Why? Well because a club like Excelsior probably does not have the budget to send a dedicated scout to Peru to watch the Peruvian Segunda División. However, with the usage of the services of SciSports or Prozone, even Excelsior could analyze the strengths and weaknesses of Jair Ayrton Cordova Carpio. Thus, these statistics and data sets provide even a smaller club with a potential access to players from all around the world. What will not change, however, is that in the end, it will be the prestige and the finances of the clubs that will lead to the final decision where a particular player will go. One interesting anecdote in relation to the title: “Will the next Messi be found through the use of statistics?” is an Guardian article from 2014 entitled “Why clubs are using Football Manager as a real-life scouting tool”, available under this link https://www.theguardian.com/technology/2014/aug/12/why-clubs-football-manager-scouting-tool. The article tells the story of how the son of a former Rangers manager told his father to sign a particular youngster of the Barcelona B team based on his performance in a video game he was playing. The name of that football player: Lionel Messi. So it is not that the “next” Messi will be discovered using statistics, the “old” one already was!
Nice article to read Lars! Have you ever heard of the club Midtjylland from Denmark? They do their whole scouting process through statistics on football players. Their philosophy is, maneuvering yourself into a scoring opportunity is skill, while actually scoring the goal is also part luck. Based on this statistic, a lot of different decisions are to be made in the selection of football players on a younger age. This would mean that at the end of the day the whole scouting process is going to change.
Regarding your question, will the next Lionel Messi be discovered through the use of statistics, my answer would be no. Lionel Messi, has some points in has game that certainly make him stand out in a game, but he also has some very unfavorable characteristics in terms of being a football player. For example: he is not tall, not strong, does not have a magnificent right foot, isn’t the greatest header of all time and seriously lacks defensive abilities. There are a lot of players that do have these attributes nowadays, so Messi would be an outlier in this system based on statistical characteristics. Also, based on the fact that ‘talent’ would be very difficult to measure. In that sense it would be easier I believe to discover the new Cristiano Ronaldo, since he has some of the characteristics I mentioned before. I would like to hear your thoughts on the matter as well!
This use of statistics is looking very promising indeed! It looks like until now it’s mainly used for analysis of players’ external behavior. However, in sports the psychology plays a very large role as well. A 13-year old football player can have the right body skills but his mindset is also extremely important to take into account. Perhaps this will be the next step.
The thing that is putting a break on this kind of developments is the romantic ideas of talent and intuition. People have to start getting ready for the world of sports to change drastically.