World Cup Teams Are in a Race for AI Dominance

The sheer scale of data being recorded at this summer’s World Cup is unprecedented. FIFA, the tournament organiser, will track around 150 million data points per match. Inside the ball alone, sensors monitoring IMUs (Inertial Measurement Units) will log 500 movements per second to trace the ball’s motion.

If that sounds excessive, Patrick Lucey can go further. “The thing with soccer is that there are more permutations (in a game) than there are atoms in the universe,” he says.

Lucey is chief scientist at Stats Perform, the data and AI company whose work underpins almost the entire global soccer ecosystem. Their statistics are used across every aspect of the modern game. It powers player scouting and multimillion-dollar fees for player transfers, helps coaching staff choose tactics and lineups, and devises corner and free kick routines. Players use it to negotiate contracts, broadcasters to entertain.

AI now enables data to be collected across matches around the globe like never before, and staff inside teams are pushing boundaries to crunch that data at unprecedented speed. At the World Cup, swathes of information will be manipulated and analysed, by humans and AI, to find a cutting edge.

Teams at this year’s Cup will also have access to a bespoke AI agent powered by Lenovo. It’s FIFA’s attempt to level the playing field. Whether or not it will be enough to do so is another matter.

“The data’s fine-grain, multi-agent, adversarial. What we do in sport is most similar to autonomous vehicles—you’re looking at trajectories,” says Lucey. “If you think of one team, there are 10 factorial permutations, just in terms of ordering players. If you include the opposition, it just explodes.”

Even smaller nations have found innovative ways to leverage technology. Curaçao, a Dutch Caribbean island with a population of roughly 159,000, became the smallest nation ever to qualify for a World Cup at this tournament after they used their own data and technology for “diaspora tracking”: mapping parentage, identifying eligible players, and using geospatial data to plan scouting trips and organise trials.

“Only one player of the Curaçao 26 was actually born on the island of Curaçao,” says Alex Stewart, chief executive of data-driven sports consultancy Analytics FC. “The rest of them were born in the Netherlands.”

Another growing use of data and AI in national federations is manager selection. Tools can analyse the pool of realistic squad options and identify managers whose tactical strengths best suit them. Teams can further use AI to help shape squad composition ahead of a tournament, based on group-stage opponents.

England are using AI for penalty analysis, knowing a penalty shoot-out can knock them out. What once took five days—analysing every penalty taker for an opponent—can now likely be done in five hours, the Football Association’s head of performance insights and analysis told the BBC.

Marcelo Bielsa, the Uruguay manager, once said when he was in charge at Premier League side Leeds United that his staff spent around 300 hours analysing an upcoming team. “We can do this automatically,” Lucey says. He shows a video of red and blue dots moving around a pitch chasing a yellow ball. Analysts can ask questions—how often a move has led to shots or goals, all the other times it occurred—each one revealing a fresh layer of information.

“You can compare this situation today with access to the web,” says Jan Wendt, cofounder and CEO of PLAIER, an AI platform working with clubs and national teams. Both British Airways and Amazon built websites in the early days of the internet. One became an information and airline ticketing platform, the other changed commerce globally, Wendt says. AI has a similar spread, changing both routine tasks and whole industries. Or, in the case of soccer, sports franchises.