US-style data analysis scores in baseball, but isn’t winning big in soccer
Inquiring minds want to know: Who, if anyone, is using sabermetrics — “Moneyball’’-style statistical analysis — in the World Cup? And is it doing them any good?
“Moneyball’’ was Michael Lewis’s mammoth 2003 bestseller that explained how Billy Beane of the Oakland Athletics crunched data to discover hidden value in unheralded baseball players, and to dictate strategy in certain key situations, such as never steal, never bunt. Bill James, who works with the Red Sox, is the father of so-called sabermetric analysis, which has spread to other professional sports, most notably pro basketball.
So what about soccer? The United States invented sabermetrics — but not for soccer. “If you and I were discussing the relative merits of two baseball players, obviously we’d talk about batting averages, or ERAs,’’ says Mark Brunkhart, president of Emeryville, Calif.-based Match Analysis, which sells “soccermetric’’ analysis to some professional teams. “But most soccer people are fervently opposed to stats, to the idea that you can take this beautiful thing and measure it.’’
Match has one client in the World Cup, which it won’t identify (I’m pretty sure it’s Mexico), and its London-based rival, Prozone, has two — England and the United States. “They are both heavy users,’’ says Prozone’s business development manager Simon Edgar. “We’ve worked for three years with [US coach] Bob Bradley and his team. They are one of the forerunners on the international scene.’’
In Britain’s famous Premier League, “everyone uses statistical analysis,’’ according to Edgar. But of the 32 teams in the World Cup, he estimates that no more than five countries have embraced soccermetrics. “In the American context, diving into stats has a certain machismo,’’ MA’s Brunkhart notes. “In Europe, it’s the opposite. Coaches try to convey that success is based on their magical understanding of the game, and if they do use statistical analysis, they try to keep it as much of a secret as possible.’’
What are some examples of soccermetrics? Touches, meaning when feet or heads strike the ball; passes; and contested balls won are a few. (Interestingly, soccer has never kept data on assists, which are crucial to basketball and hockey.) Match Analysis found an “insanely high’’ correlation between number of touches and international soccer rankings in the first 16 World Cup games, according to Brunkhart. On the other hand, I couldn’t help noticing that their touches and passes data for the close-fought Uruguay-Mexico game favored Mexico, which lost. “It’s by no means a definitive science,’’ says Edgar. “It’s 100 percent objective, but it’s not subjective at all. You can never account for luck, refereeing, or team strategy. England could have a superior pass rate, but if they have decided to play a defensive game, that statistic is irrelevant.’’
Steve Gans, a former professional player who works with the Premier League team Fulham on business matters, is somewhat skeptical about reliance on soccermetrics. “The analogy from baseball to soccer is not a neat one,’’ he says. “Soccer is not a natural statistical sport. You can come up with these categories of measurement, but can you be certain of the relevance?’’
One stat that can be telling, Gans says, is completed passes. “On the face of it, that number says how good you are under pressure. If it’s low, it suggests your team isn’t performing well under pressure. On the other hand, you can take a team that plays a probing or defensive style, and they will hit a lot of consecutive passes, but not necessarily under pressure or to the offensive good.’’
Brunkhart agreed to send me his soccermetric analysis of Wednesday’s dramatic US-Algeria game just a few hours after Landon Donovan’s goal found the back of the Desert Foxes’ net. Here it is, in part:
“In the second half, Algeria’s ability to retain the ball player-by-player dropped from 71.4% to 58.5% (US’s increased from 67.2% to 69.7%). Algeria’s touch count in the second half also dropped 24%. That seems to indicate a weary and manhandled Algerian side.
“Donovan is known for being able to perform in the big matches. In the US-Slovenia match, Donovan had only 48 touches and drifted to the left side of the field for over half of them, lost the ball over half the time he touched it, and made no passes leading to shots. His goal was critical but his performance was not as strong as it was against Algeria.
“In the Algeria match, he stayed wide right with 91% of his touches right of center, saw the ball 67 times, played 4 passes leading to shots, and retained possession 69% of the time he touched it. Much better game for Donovan against Algeria.’’
Too geeky for the “beautiful game’’? To quote the mathematician Paul Erdös: “If numbers aren’t beautiful, nothing is.’’
Alex Beam is a Globe columnist. His e-dress is firstname.lastname@example.org.