In the 2010 Champions League final, Bayern Munich appeared to decisively outplay Inter Milan. Bayern dominated possession of the ball, made more than double the passes Inter did, completing almost all of them, and delivered 23 threatening crosses in the offensive zone to set up potential goals. Inter made just three crosses the entire match.
Yet Inter won 2-0.
A deeper statistical analysis pulls away the facade of Bayern's seemingly superior game. Many of its passes were made in the middle third of the pitch and the team struggled to pierce Inter's wall of defence. When Bayern did squeeze through and sent in crosses from the wing, just four of 23 connected with a Bayern striker - real chances to score were actually rare.
Inter, meanwhile, made many fewer passes but was more aggressive. When its smothering defence stole the ball from Bayern, big counterattacks were launched by belting the ball up the field. On free kicks, Inter also fired the ball up the field. Using this strategy, Inter snatched its first European championship in 45 years.
The intricate patterns of a soccer match, for decades assessed by gut feel, are now laid starkly bare by software programs widely used by top-tier teams in Europe and, in the past couple years, by squads in North America's Major League Soccer. Toronto FC was among the first in the league to employ statistical analysis, to measure its own performance and failings, as well as opponents' strength and vulnerabilities. Vancouver Whitecaps FC, headed by former English premiership executive Paul Barber, signed on for its MLS expansion season.
"There's no hiding from a system like this," Barber said. "It's almost like an X-ray."
There are a small number of firms that produce such software. Toronto and Vancouver use programs from Prozone Sports Ltd., a 13-year-old company in Leeds, England, that popularized statistical analysis among Premier League teams. Soccer joins other sports such as baseball and basketball in the use of stats to divine subtle advantages and uncover problems.
Author Michael Lewis brought such work to wide attention in his 2003 book Moneyball: The Art of Winning an Unfair Game, which looked at the Oakland Athletics and general manager Billy Beane's data mining. The strategies have quickly spread, at least to an extent, to most pro sports, Lewis said in a 2009 story: "Each one now supports a subculture of smart people who view it not just as a game to be played but as a problem to be solved." And number-crunching has been elevated to the highest ranks of U.S. scholarship. MIT has staged an annual sports analytics conference for the past five years, with topics at this year's confab in March featuring research such as the basketball question: "How much trouble is early foul trouble?"
In soccer, the work begins high above the pitch. At Empire Field in east Vancouver, four small high-definition cameras are mounted under the awning above the press box. The one on the right peers leftwards, and the left camera records the right side of the field, with the middle two looking at the middle half. The four together produce a picture of the entire playing area at all times.
Soccer, especially to the eye of a spectator who doesn't know the game well, looks uneventful, particularly on the scoreboard, where 0-0 isn't a rare result, even in the unruly MLS. But the level of detail piles up. In the Bayern-Inter final, there were 2,842 "events" - the gamut of passes to saves, throw-ins to yellow cards - and, occasionally, goals.
After a match, a Whitecaps intern who monitored the quality of the video recordings compresses the files and uploads them to the Internet. Prozone takes information and codes the game and returns the package about 15 hours later. Everything can be parsed with exacting precision and all of it is directly attached to game video.
"It's really changed the landscape of European football," said Andrew Opatkiewicz, who runs Prozone's business in the Americas.
"Clubs like Chelsea will have a full team of six individuals who sit in a dark room in front of their individual computer screens, for eight to 10 hours a day, Monday to Friday, just crunching numbers, looking for whatever competitive advantage they can find."
In MLS, which is far below the Premiership in terms of skill and cash resources, using stats to score is just beginning. Only one club has an employee dedicated to the tools, the New York Red Bulls, who signed on a 24-year-old British analyst who had worked for a third-tier English side. MLS this week issued its first performance ranking, a cocktail blend of numbers compiled by London-based Opta that gauges all the players in the league on a scale of 10. Thierry Henry, the French star striker on the Red Bulls, ranked No. 1 for April.
"Within MLS, the use is still in its infancy," Opatkiewicz said.
For coaches, it quickly quantifies what they can see on the field. Teitur Thordarson, the Whitecaps coach, calls Prozone "a huge step forward." In one example, with his general pressing-offensive strategy that is underpinned by a demanding fitness regime, a Prozone feature Thordarson uses is the tally of a player's movements on the field. The results include how much ground players cover and their varying paces, sprints and strolls. The numbers illuminate the players who deliver and those who fade - and remedies can be made.
Coaches, executives and the developers of the software emphasize the key is analysis, not the numbers. Software and stats don't provide answers, they help show what questions to ask. "All of us can sit in front of a pile of data and not learn anything from it," said Barber, the Whitecaps chief executive officer.
And it doesn't always add up to wins. While Prozone users such as Real Salt Lake have had success on the field, others such as Toronto have not. Vancouver, in its debut season, has won just a single game, its first outing, and is 1-4-5. And the only team below Vancouver in the Western Conference standings is the San Jose Earthquakes, devotees of stats analysis because it shares the same ownership as the Oakland A's. The A's themselves, after a Moneyball-fuelled run of success in the early 2000s, haven't made the playoffs since 2006.
As analytics fixes itself in the mainstream of sport, skepticism remains strong.
"If you survey soccer coaches, you'll get the nod, 'Yes, we believe in analytics work, we believe in the study of the sport,'" said Mark Brunkhart, president of software maker Match Analysis in California. "Just because you have stats available doesn't mean anyone actually uses them to do anything."
The opposite problem also is prominent. With the prospect of unravelling soccer's mysteries, there's a lure that numbers will unveil all answers.
"Because of Moneyball, there's this desire where people want to solve soccer," Brunkhart said. "'Here, we're going to plug some numbers into an equation, it'll tell us what's wrong and we're going to fix things.' If one more person comes to me and says, 'We want to solve soccer, we're hiring an intern, can we have your data?' I'll cry. It's a very difficult thing to study."