Zen and the Art of Soccer Moneyball Maintenance

♠ Posted by Emmanuel in at 6/19/2011 02:57:00 PM
What constitutes knowledge? Questions of epistemology underpin many of the debates in the social sciences--especially the supposed divide between the "objective" and the "subjective": Should we focus on the generalities of several cases or the particularities of each? Are there universal laws governing behaviour or is it relativistic? These are hard questions to answer that surround any number of human endeavours. The global financial crisis is certainly one, when many attributed the desire to blandly generalize Anglo-Saxon assumptions as leading many into temptation. Hence Greenspan's (semi-)confession that the self-regulating market may not always correct itself. Today, though, consider a slightly sportier field encountering a similar debate--soccer. Money changes everything, including time-tested managerial practices, it seems, judging by a number of things.

Many, many years ago, Robert M. Pirsig had a surprise bestseller with the creatively titled Zen and the Art of Motorcycle Maintenance that considered the objective/subjective divide through the lens of keeping his bike in good condition throughout his travels. While quite an interesting, introspective read, let me confess that I didn't know what he was writing about most of the time. (At a lecture I attended, famed sociologist Anthony Giddens joked he'd been around when even he had trouble understanding his earlier writings, and I suspect Pirsig may encounter the same problem re-reading this book.) At any rate, I also read the follow-up book Lila which pursues similar themes. My takeaway--don't ask me to explain--is that Pirsig believes the objective-subjective distinction is an artificial divide.

This medium-winded intro bring me to an interesting article in FT Weekend by sportswriter Simon Kuper on the growing statisticization of soccer (football). Although baseball has long been the domain of number crunchers due to its stop-start nature which is said to make it amenable to statistical analysis, the international big bucks in soccer have made it a natural target. Moreover, with another fluid game having been "colonized" already in basketball, it was only a matter of time before the quant jocks invaded soccer's locker room.

Talking about motorcycle maintenance, we turn to the matter of football team maintenance. It certainly looks as if the quant jocks will be given more input in managerial decisions, especially squad selection. It is here where supposedly "objective" number-crunching will begin bumping up against "subjective" gut feel and intuition. At any rate, the questions are similar for soccer as they are in other activities where the objective/subjective distinction manifests itself:
  • Is soccer an activity in which statistical analysis plays a role?
  • If so, what statistics are meaningful bases for player comparisons?
  • Which team statistics correlate with win/loss records?
  • Do stats provide a superior basis for making decisions than managerial experience?
There are dozens and dozens of these sorts of questions you can come up with. Kuper provides many examples of the quants and traditionalists brushing up each other, often unhappily. What follows though is a mini-history of the use of data collection in the Premier League by two top managers, Sir Alex Ferguson of Manchester United and Arsene Wenger of Arsenal that paint a decidedly mixed picture. The former has developed a more cautious view of the technology after mistakenly letting go of a player based on "Type II" error and remains a seat-of-the-pants manager who's won many more trophies since. The latter is still a numbers junkie who famously hasn't won a major trophy in five years:
But the broader breakthrough came in 1996, after the Opta Index company began collecting “match data” from the English Premier League, explains the German author Christoph Biermann in Die Fussball-Matrix, the pioneering book on football and data. For the first time, clubs knew how many kilometres each player ran per match, and how many tackles and passes he made. Other data companies entered the market. Some football managers began to look at the stats. In August 2001 Manchester United’s manager Alex Ferguson suddenly sold his defender Jaap Stam to Lazio Roma. The move surprised everyone. Some thought Ferguson was punishing the Dutchman for a silly autobiography he had just published. In truth, although Ferguson didn’t say this publicly, the sale was prompted partly by match data. Studying the numbers, Ferguson had spotted that Stam was tackling less often than before. He presumed the defender, then 29, was declining. So he sold him.

As Ferguson later admitted, this was a mistake. Like many football men in the early days of match data, the manager had studied the wrong numbers. Stam wasn’t in decline at all: he would go on to have several excellent years in Italy. Still, the sale was a milestone in football history: a transfer driven largely by stats.

At Arsenal, Wenger embraced the new match data. He has said that the morning after a game he’s like a junkie who needs his fix: he reaches for the spreadsheets. In about 2002 he began substituting his forward Dennis Bergkamp late in matches. Bergkamp would go to Wenger to complain. “Then he’d produce the stats,” Bergkamp later recalled. “‘Look Dennis, after 70 minutes you began running less. And your speed declined.’ Wenger is a football professor.”
Go figure. I'll make that time-tested cop out and say that managerial judgement is important for complementing technology in determining what statistics matter and their proper interpretation. At any rate, I certainly would welcome judicious use of number crunching to avoid tremendous wastes of money a la Abramovich.

It gives a whole new meaning to booting up, doesn't it?