Phrases like “striking the post” and “direct free kick outside the 18” may seem foreign if you’re not a fan of football (for Americans, see: soccer). But for a football scout, it’s the daily lexicon of the job, representing crucial language that helps assess a player’s value to a team. And now, it’s also the language spoken and understood by Scout Advisor—an innovative tool using natural language processing (NLP) and built on the IBM® watsonx™ platform especially for Spain’s Sevilla Fútbol Club.
On any given day, a scout has several responsibilities: observing practices, talking to families of young players, taking notes on games and recording lots of follow-up paperwork. In fact, paperwork is a much more significant part of the job than one might imagine.
As Victor Orta, Sevilla FC Sporting Director, explained at his conference during the World Football Summit in 2023: “We are never going to sign a player with data alone, but we will never do it without resorting to data either. In the end, the good player will always have good data, but then there is always the human eye, which is the one that must evaluate everything and decide.”
Read on to learn more about IBM and Sevilla FC’s high-scoring partnership.
Benched by paperwork
Back in 2021, an avalanche of paperwork plagued Sevilla FC, a top-flight team based in Andalusia, Spain. With an elite scouting team featuring 20-to-25 scouts, a single player can accumulate up to 40 scout reports, requiring 200-to-300 hours of review. Overall, Sevilla FC was tasked with organizing more than 200,000 total reports on potential players—an immensely time-consuming job.
Combining expert observation alongside the value of data remained key for the club. Scout reports look at the quantitative data of game-time minutiae, like scoring attempts, accurate pass percentages, assists, as well as qualitative data like a player’s attitude and alignment with team philosophy. At the time, Sevilla FC could efficiently access and use quantitative player data in a matter of seconds, but the process of extracting qualitative information from the database was much slower in comparison.
In the case of Sevilla FC, using big data to recruit players had the potential to change the core business. Instead of scouts choosing players based on intuition and bias alone, they could also use statistics, and confidently make better business decisions on multi-million-dollar investments (that is, players). Not to mention, when, where and how to use said players. But harnessing that data was no easy task.
Getting the IBM assist
Sevilla FC takes data almost as seriously as scoring goals. In 2021, the club created a dedicated data department specifically to help management make better business decisions. It has now grown to be the largest data department in European football, developing its own AI tool to help track player movements through news coverage, as well as internal ticketing solutions.
But when it came to the massive amount of data collected by scouters, the department knew it had a challenge that would take a reliable partner. Initially, the department consulted with data scientists at the University of Sevilla to develop models to organize all their data. But soon, the club realized it would need more advanced technology. A cold call from an IBM representative was fortuitous.
“I was contacted by [IBM Client Engineering Manager] Arturo Guerrero to know more about us and our data projects,” says Elias Zamora, Sevilla FC chief data officer. “We quickly understood there were ways to cooperate. Sevilla FC has one of the biggest scouting databases in the professional football, ready to be used in the framework of generative AI technologies. IBM had just released watsonx, its commercial generative AI and scientific data platform based on cloud. Therefore, a partnership to extract the most value from our scouting reports using AI was the right initiative.”
Coordinating the play
Sevilla FC connected with the IBM Client Engineering team to talk through its challenges and a plan was devised.
Because Sevilla FC was able to clearly explain its challenges and goals—and IBM asked the right questions—the technology soon followed. The partnership determined that IBM watsonx.ai™ would be the best solution to quickly and easily sift through a massive player database using foundation models and generative AI to process prompts in natural language. Using semantic language for search provided richer results: for instance, a search for “talented winger” translated to “a talented winger is capable of taking on defenders with dribbling to create space and penetrate the opposition’s defense.”
The solution—titled Scout Advisor—presents a curated list of players matching search criteria in a well-designed, user-friendly interface. Its technology helps unlock the entire potential of the Sevilla FC’s database, from the intangible impressions of a scout to specific data assets.
Sevilla FC Scout Advisor UI
Scoring the goal
Scout Advisor’s pilot program went into production in January 2024, and is currently training with 200,000 existing reports. The club’s plan is to use the tool during the summer 2024 recruiting season and see results in September. So far, the reviews have been positive.
“Scout Advisor has the capability to revolutionize the way we approach player recruitment,” Zamora says. “It permits the identification of players based on the opinion of football experts embedded in the scouting reports and expressed in natural language. That is, we use the technology to fully extract the value and knowledge of our scouting department.”
And with the time saved, scouts can now concentrate on human tasks: connecting with recruits, watching games and making decisions backed by data.
When considering the high functionality of Scout Advisor’s NLP technology, it’s natural to think about how the same technology can be applied to other sports recruiting and other functions. But one thing is certain: making better decisions about who, when and why to play a footballer has transformed the way Sevilla FC recruits.
Says Zamora: “This is the most revolutionary technology I have seen in football.”
Source: ibm.com
Posts
0 comments:
Post a Comment