Goooal! Scientists Score in Ranking Soccer Stars

If you're new to World Cup soccer madness and wondering which team is tops, science has an answer. Researchers have developed software that can measure soccer success, creating a more sophisticated ranking system for the sport and a tool for analyzing performance in other team activities.

Unlike baseball and American football, the world's most popular sport doesn't produce many stats. Only a handful of soccer ranking systems exist, most of which rely on limited information: the number of goals scored in a match, the number of goals assisted, and some indices of a match's difficulty and importance. Although such systems are meant to assess which teams and players are the best, their focus on scores alone limits their accuracy, says Luís Amaral, a complex-systems engineer at Northwestern University in Evanston, Illinois. "You end up with no information on anyone but the player who made the goal and the player who assisted."

An avid soccer fan, Amaral wanted to measure team and player performance in a way that takes into account the complex interactions within the team and each player's contribution. So he turned to an unlikely source: social networks. Applying the kinds of mathematical techniques used to map Facebook friends and other networks, Amaral and colleagues created software that can trace the ball's flow from player to player. As the program follows the ball, it assigns points for precise passing and for passes that ultimately lead to a shot at the goal. Whether the shot succeeds doesn't matter. "There's lots of luck involved in actually getting it in," Amaral explains. Only the ball's flow toward the goal and each player's role in getting it there factors into the program's point system, which then calculates a skill index for each team and player.

When the researchers used the program to analyze data from the 2008 UEFA European Football Championship, the indices closely matched the tournament's outcome and the overall consensus of sports reporters, coaches, and other experts who weighed in on the performances. The results appear online today in PLoS ONE.

"One of the issues with any kind of teamwork is assigning the right credit," says Amaral. "The wild, loud people get more credit, but with this analysis you can get a picture of how much an individual really contributes to an outcome."

"This is a nice paper, and it's timely—the world is holding its breath for the [World] Cup," says Alessandro Vespignani, an equally soccer-smitten computational scientist at Indiana University in Bloomington. "But it is much deeper than that. It is a powerful way to analyze any team performance: scientific teams, companies, creative groups."

Amaral plans to apply the program to other sports and fields and is already supplying it with 2010 World Cup data—with one ulterior motive. "I am from Portugal, and I was disappointed in how we played Tuesday," he says. (His team tied 0-0 with Ivory Coast.) "I'm very curious to see what score the program gives us—maybe we played better than I thought."