Tomorrow's talent: it's all in the data
How AI and analytics is helping to create the skills index that will be used to determine future generations of football stars.
With the FIFA 2018 World Cup in full flow, fans around the world are marvelling at the skills demonstrated by the current crop of players. However, few have probably given thought to where the next generation of talented footballers will come from.
It is recognised that most successful sports teams already use advanced analytics to improve team performance, so it should be expected that this technology will also be used to ensure that those future players with the talent to grace the world stage do not remain undiscovered diamonds-in-the-rough.
However, there is more to it than simply discovering tomorrow's star players. It is also about understanding exactly how these players will combine with their teammates, since the best team is not only about individual skills but also about the sum of its parts, and how well players combine with one another. What we are seeing now with football, and many other sports, is that artificial intelligence is being brought to the pitch, says Aneshan Ramaloo, senior business solutions manager at SAS. He explains that SciSports, a Dutch company founded in 2012 by two self-proclaimed football addicts and data geeks, is one company that is innovating on the edge of what's possible.
"What this business does is it uses streaming data and applies machine learning, deep learning and artificial intelligence to this information, allowing it to capture and analyse it in such a way as to enable innovations in everything from player recruitment to virtual reality for fans," he suggests.
"The traditional method of finding new talent has always focused on scouts and coaches, who observe particular players, make use of rudimentary data and rely mostly on intuition to determine who the club should sign. Today, however, the more savvy clubs are using advanced analytics, rather than mere gut-feel, to identify rising stars and undervalued players."
Explaining how the SciSkill Index works, Ramaloo indicates that it evaluates every professional football player in the world using a single universal index. This is done by utilising machine learning algorithms to calculate the quality, talent and value of more than 200 000 players. This, in turn, helps clubs find new talent, search for players that fit a certain profile and even analyse their opponents. He adds that, every week, more than 1 500 matches across 210 global leagues provide the data that is put through the company's advanced analytics and machine learning solutions.
"What is being done here with analytics goes beyond simply studying the player that has the ball. Usually, even those forward-thinking clubs that seek more in-depth data on players tend to focus only on the individual who has the ball, leaving everything else undocumented. While better than sight and gut-feel, this nonetheless provides an incomplete picture of player quality."
"What SciSports is doing now, thanks to an advanced camera system, is capturing the immense amount of data happening away from the ball. The solution it uses is a real-time tracking technology that automatically generates 3D data from video. Fourteen cameras placed around the stadium record every movement on the field. The technology then generates data such as the precision, direction and speed of the passing, sprinting strength and jumping strength."
In other words, says Ramaloo, it allows the coach to form a complete picture of the game. Not only does it give insight into what players are doing off the ball, such as the angles being run by a striker, but it can help them determine if a player is getting tired and should be substituted. In addition, the data can also be used to enable fans to experience the game from any angle, using virtual reality, or to enliven sports betting and fantasy sports.
"The solution models on-field movements using machine learning algorithms, which by nature improve on performing a task as they gain more experience. So, for example, it automatically assigns a value to each action, such as a corner kick. Over time, these values change, based on their success rate. A goal will have a high value, but a contributing action, which may have previously had a low value, can become more valuable as the platform itself masters the game."
There is no doubt, he suggests, that artificial intelligence and machine learning are going to play a critical role in the future of football analytics.
"The current mathematical algorithms only model existing knowledge and insights in football. On the other hand, artificial intelligence and machine learning are now starting to make it possible to discover new connections, ones that people simply wouldn't make by themselves. In effect, although the saying goes that 'seeing is believing', which suggests that the eyes never lie, the truth of the matter is that, when it comes to identifying the talent of tomorrow, it's the data that never lies," concludes Ramaloo.