At ETH Zurich, Ulrik Brandes is teaching the Swiss national team to see football the way artificial intelligence does—not as formations drawn on a whiteboard, but as constantly shifting networks of movement and intention. As the 2026 World Cup approaches, the professor of social networks has transformed tracking data from international tournaments into a tactical weapon that reveals what television formations never will.
Most modern football stadiums are filmed by multiple cameras capturing the position of every player and the ball at every second. For years, Brandes has mined this data to understand how teams actually behave during matches, going far beyond the static 4-3-3 or 3-5-2 diagrams that dominate halftime analysis. "Football often appears to have an unpredictability and spontaneity about it," Brandes explains. "But if you look at enough data, stable patterns start to emerge."
The difference between formations and reality matters enormously in modern football. Players constantly switch roles, swap sides, and adjust their positioning based on whether their team has possession. A center forward might drop deep into midfield. A left back could roam so far forward he functions almost as a winger. A defensive midfielder becomes the playmaker and therefore a vulnerability to aggressive pressing. These dynamic patterns are invisible to traditional analysis.
When Switzerland's coaching staff prepare opponents for the 2026 World Cup—particularly under the time pressure of group phases where traditional video study becomes impractical—they now rely on Brandes' rapid visualizations of tactical patterns. His data might reveal that an opponent's left back is far more attacking than the right back, or which midfielder operates as the team's creative hub and where they can be pressured. The Swiss federation uses these insights to identify weaknesses that exist only in the actual flow of play, not on paper. Brandes is careful to note he only prepares and presents the data; the coaching staff alone decides what conclusions to draw from it.
Beyond formations, Brandes has also pursued a fascinating scientific puzzle: what makes penalty shootouts genuinely different from penalties during normal play? After analyzing nearly 2,000 penalty shootouts, he and his co-authors reached a surprising conclusion. The widely repeated claim that teams shooting first gain an advantage proved false—"The statistics contradict the claim that teams that go first are more likely to win," Brandes says.
But they found something clearer: conversion rates collapse under the added pressure of shootouts. During normal play, roughly four out of five penalties go in. In shootouts, only three out of four succeed. The culprit is not improved goalkeeping or superior technique by defenders. Instead, Brandes identified the core difference: more misses by the players themselves. "Even players who are otherwise very reliable score less often in penalty shootouts," he observes. The psychological weight of sudden-death moments reshapes even the most accomplished athletes' accuracy.
For the 2026 World Cup in Canada, Mexico, and the United States, Brandes will be at work both before and during the tournament, distilling patterns from raw tracking data into tactical intelligence. In a sport that prizes intuition and experience, data has become another form of seeing—one that reveals the hidden geometry of play beneath what the naked eye observes.