Any player from a soccer school is capable of emulating how Haaland, Mbappe, Benzema or Messi celebrates goals, even reconstructing their avatar in a video game. It is the power of the image to which the new generations spend thousands of hours exposed. Could that same strength of the image allow us to emulate them? It is not a conditional, it is a reality, and each time extended. Artificial intelligence already opens the door to complement and improve the natural talent that accumulates in the quarries. As? By dint of knowing the behavior on the field of the best players, assimilating it and learning to react in a similar way to the same situations.
You can 'manufacture' a player who learns to make the movements behind the defensive line that City's goalscorer makes or how he occupies space in the area to have an advantage in shooting and entrust the rest to instinct. Even testing, by comparison, how the Norwegian behaved on the field of play at the age of 17, what were the movements that predicted a bright future for him, and how the most promising forward of a quarry behaves at that same age, whose most promising parameters conventional ones make you think that they have similarities.
In this task, the automated tactical analysis of the game made by the startup Kognia appears as a weapon, and which only some youth academies are beginning to explore, but which already adds value to coaches such as Xavi Hernández at Barça, Unai Emery at Aston Vila or Paulo Sosa in the Salernitana. Its technology is based on software that, through algorithms and advanced computer vision techniques, detects more than 300 tactical situations of a player, a line or the entire team in the video of the game and collects them in video clips with the graphics (telestation ) of the tactical concept.
"Nobody in the industry generated that automatically," explains the CEO of Kognia, Maurici A. López-Felip, to EL MUNDO, who began to see the potential of video analysis in soccer in 2008, when he was studying Activity Sciences. Physics and Sport in Barcelona and collaborated with Joan Vilà, former coach of La Masía, considered the discoverer of Xavi Hernández and head of the Methodology area at the Barça club for years.
López-Felip edited videos manually to highlight tactical concepts that Vilà wanted to transfer to players to improve their game. However, the unique tool that has become Kognia is underutilized in the quarries, although it is beginning to make its way. There are teams like Villarreal that monitored all their 11 football teams, so they have been generating concepts of each of their players. «A great opportunity opens up for them to be able to work in a very professional way with the youth academy with all the information that we generate through telestation. Some clubs in Spain, England or Germany have been recording the teams from their academies for years, even if it is with portable cameras and lower quality, and now they are beginning to have the will to, with all that data generated, confront players and Find out how to make them improve. We have helped some in that comparison and perhaps the next step is to offer it in the tool, “says its creator.
This comparison can be with players in direct competition, by age or category in which they play, with great historical references that have emerged from the club's own factory or even with stars from the past. "If we had enough videos of Andrés Iniesta or Fernando Torres when they were youth players, we could analyze them and extract all the data from their game, and that would allow us to make a comparison of what their movements were like to be able to measure them with those who occupy their positions in the categories today. inferiors of Barça or Atlético. But it is that we could analyze any player, even Maradona if we had the videos of his matches and they were of quality, something that is not possible, "he adds.
«Until now, the data analytics industry linked to sports, which is barely 20 years old, compares one player with another based on superficial data such as the number of shots, trips, recoveries, fouls or the distance covered in a match, but those are not representative. We need data on tactical concepts, on how he participates in the game and, for example, creates chances for himself or for his teammates or how he moves in the area to take advantage of a cross. This is impossible to do manually, and we have provided the automatic filter to be able to respond to the needs of sports management and technical bodies, "insists Maurici A. López-Felip.
Where the idea of monitoring all grassroots football to control every detail takes shape is in the United States, their sport is very familiar with data analysis. Major League Soccer (MLS) created its Next category in 2000 in which teams from under-13 to under-17 compete and has already been interested in how to collect the data generated in the almost 10,000 games played each season.
“I had not found a way to do it at scale because the costs in personnel and infrastructure were unfeasible. We can do it because we only need the video to generate all the information they need, facilitate their analysis and provide them with underlying data that their own data scientist teams can process, as we do with the Bundesliga, for example”, recalls the CEO of Kognia, who finds an explanation for the fact that its technology can be implemented earlier in the great American quarry than in the European ones. “It is a league that thinks more about the future and is very concerned about the structure. Perhaps because of its franchise economic model where there is no uncertainty of relegation and a sports director can work on a long-term project. In Europe everything is more geared towards immediate performance”.
According to the criteria of The Trust Project