While baseball has the most storied quantitative tradition, plenty of other sports now use sophisticated statistical analysis in the pursuit of success. It’s the latest thing in hockey, and some particularly die-hard fans of professional cycling dug into the data in 2010 to see if changes in Union Cycliste International’s doping rules had an impact on that year’s Tour de France.
Soccer might be next if Laszlo Gyarmati, Haewoon Kwak, and Pablo Rodriguez get their way, though there are technical challenges. Discovering what leads to success in soccer is tricky business, they argued in research presented last month in a workshop at the Association for Computing Machinery’s Knowledge and Data Mining Discovery conference in New York. As plenty of Americans learned for the first time this past summer, soccer is not a high-scoring game, so much so that matches are frequently scoreless after 90 minutes of play. And although one person gets official credit for a goal, it’s usually a sequence of well-chosen, well-timed passes rather than a single player’s heroic charge that wins games.
While there were different styles—the researchers identified three other more or less distinct styles, including one characterized by increased use of the ABCA motif as implemented by Atletico Madrid and others—they were generally matters of degree.
That latter observation inspired Gyarmati and his colleagues’ methodological approach. Drawing on detailed, publicly available data, the trio looked for “flow motifs,” patterns of three passes between players that could be used to identify a team’s footballing style. The ABAB motif, for example, indicates that two players passed the ball back and forth a couple of times. ABCA, on the other hand, implies that one player passed to a second, who passed to a third, who finally passed back to the first.
In an age where every team can watch tape of every other teams’ games, the researchers expected that teams would adopt similar strategies and similar styles of play, and to some extent that’s true. Across the Spanish premier league, known as La Liga, nearly all teams used the various motifs in roughly, though not perfectly equal, proportion. While there were different styles—the researchers identified three other more or less distinct styles, including one characterized by increased use of the ABCA motif as implemented by Atletico Madrid and others—they were generally matters of degree.
The exception, they found, was the perennially successful FC Barcelona. Despite being known for its fast-paced “tiki taka” style based on groups of three players passing to each other, it was actually back-and-forth passing—ABAB, ABAC, and so on—that characterizes Barca’s play.
Though they’ve yet to analyze whether one style of play wins more matches or analyze how individual players contribute to a team’s style, the researchers write that their analysis could help football clubs select players that better fit within their larger line-up. Top players, they note, don’t always work out, and statistics that help identify whether an individual can work well as part of a particular team could make a crucial difference.
