In 2003, the sports director of the Seville F.C.., Ramon Rodriguez, Monchi, He received a call from his leading scout in Latin America to sign Dani Alves, an 18-year-old footballer who played for a team in the Brazilian second division. There were no more references than the personal bet of that employee, Antonio Fernández, who came to put his position at the disposal of the president of the club in the event that the footballer turned out to be a fiasco. Sevilla signed him, after a season on loan, for 800,000 euros and four years later sold him to FC Barcelona for 36 million euros. If this operation had been carried out today, Fernández would probably have breathed easier because his intuition would have been supported by the avalanche of empirical data developed and processed by an application that the club has developed with its own software to fine-tune decision-making and that forms part of a transversal project that seeks to apply the algorithms of machine learning or machine learning to the rest of the club to gain sporting and financial efficiency. “It is the necessary step to update ourselves and adapt to the new times, because they are here,” says Monchi.
In a social context where there are few things left that cannot be measured, translating intuition into numbers seems like a matter of time. The world of sports is no stranger to this quest to transform subjective perception into numbers and in this process of dataficationSevilla has found “the square of the circle”, in the words of its sports director. The Nervión team has created an application based on artificial intelligence, AIFootball, which draws on the subjective reports of the scouts and the objective analysis of the data. “This allows us to shorten the time in the search for footballers and the margin of error,” says Monchi.
With him at the head of the sports management, Sevilla has perfected in the last 20 years a successful and profitable business model based on the generation of capital gains from the sale of players, a system in which the design and management of databases had already become a key element. With the departure of the team manager in 2017, both he and the club began parallel paths in deepening the adaptation of algorithms to technical management. Monchi learned it first-hand in his new destination, Rome, where his then owner, the North American businessman Jim Pallota, had incorporated systems of machine learning that were used by the NBA. Sevilla, faced with sporting uncertainty, wanted to implement a system that, supported by data, could guarantee reliability when making transfers, regardless of the coach or sports director who was in charge, an effort that began to lead in the Andalusian capital its general director, José María Cruz de Andrés and for whom they contacted Elías Zamora in Germany, one of those responsible for GetCapital AGan industrial asset management company that develops artificial intelligence applications for investments.
By the time Monchi returned to Sevilla, in March 2019, the collaboration between José María Cruz Gallardo, head of the club’s data section, and Zamora had germinated into a project that the Sevilla manager finished defining, providing it with content that has allowed to develop the application with the introduction of methodological improvements provided by its team of scouting. “We introduced what they called game principles: not just how many passes a player makes, but what it means to pass well; what does it mean to enable…”, explains Zamora. “They are subjective elements that we transform into a mathematical formula,” he points out.
With the arrival of Monchi, the R&D department commanded by Cruz Gallardo and directly linked to the sports technical management was launched, and in 2021, as the complexity of the AIFootball application was defined based on the demands of the sports director of Seville, the Data department was created, directed by Zamora, and which functions as a transversal section that supports the rest of the entity. “What makes us different is that all the development and software capacity is generated in Seville, we don’t have to outsource it, and with Andalusian professionals,” Zamora points out. Between both departments, 22 highly qualified professionals are employed.
The AIFootball application not only includes objective and measurable parameters —physical, contextual, cinematic or technical data—, nor the methodological and tactical concerns of the technical team —such as defining sequences such as shots on goal after passes from outside the area, or what is a game elaborated or direct game, where, as Zamora explains, not only the specific player has to do but the collective, also aspires to look for players who are not only the best in each position at the data level, “but from the point of view of the opinion of our technicians”, Zamora abounds, who gives examples of soccer players who have triumphed in Seville and who, a priorithey would not have had the endorsement of the measurable requirements.
“Charisma cannot be measured, but it is very important, hierarchy cannot be measured, but it is essential,” says Zamora. That subjective perception is provided by Monchi’s team of scouts through their own assessments, which are also included in the shaker of machine learning of the club application.
That’s why when Monchi affirms that a player “has passed the filters that the club has to decide us”, It means that it is backed by the exhaustive analysis provided by the Sevilla software that allows understanding and evaluating the footballer’s performance and its potential adaptation to the squad and the coach’s requirements, and by the intuition and experience of the technical team. “The application allows us to play with a wide variety of parameters,” says the club’s technical director.
In the times of modern football, it is not enough to find the right signing, it is vital to get ahead of the rest and make the most of a sports project. And there, Zamora’s experience in the field of finance is also an asset in the datafication 360º to which the club has been entrusted. In addition to the AIFootball application, Sevilla has been running AIRadar since February 2022, which allows young promises under 23 years of age to be detected based on how their performance is being in their respective leagues through the monitoring of 32 competitions throughout the world.
Another application that has been in operation since December 2021 is AITracking, which tracks all the players who have passed through the Sevilla youth academy to detect the transfer operations from which the club can obtain benefits for their training.
Sevilla is also developing another application to optimize ticket sales based on subscriber data, their influx to the stadium… “This allows us to retain those who usually come or put more seats on sale through predictions” , indicates Zamora. His team is also working on the design of other applications to detect possible sponsors. “The advantage is that it is our own software and it can be constantly evolving and updating according to our interests, without having to outsource it or depend on third parties, and we can also market it,” explains Cruz Gallardo, in another example of the versatility of the company’s business model. that Sevilla boasts. There are teams and federations interested in the software designed at Sevilla FC that the club is interested in marketing.
Zamora defines the use of machine learning of these applications as a “refinery” that distills the data it processes from other information bases. But the use of data does not end in the design of tools to optimize the operation and income of the entity. The R&D department that Cruz directs seeks to incorporate specialists from outside the world of football, linked to science, aeronautics or engineering, into this data management and give them an outlet through training, with the aim of improving the training systems of the quarry and detect the talent that can reach the first team or be sold to other clubs.
In a world where nothing escapes the algorithms, Sevilla not only does not want to be an exception, but has positioned itself as a reference club in the generation of advanced and specialized software with a pioneering system. But in which, as Monchi warns and until the mathematical formula that can define it is found, common sense and experience also count.
Apply the data to the quarry
The R&D department directed by José María Cruz Gordillo is supported by a Sports Medicine and Science team, which addresses issues related to training and nutrition; in the group of former managers of the sports technical management who work side by side with the Zamora department in the development of applications; a team of analysts who record all the training sessions of the lower categories to extract relevant information about players and coaches, which allows them to determine which ones have the conditions to go to the first team, and another team made up of an aeronautical engineer and a computer engineer. “We integrate all areas, which allows us to incorporate technology in a coherent and reasonable way to all work processes,” explains Cruz Gallardo.