Oxford's oldest student newspaper

Independent since 1920

Football’s statistics revolution?

If Manchester City’s Performance Analysis Department are right, then football’s big news in August had nothing to do with Robin Van Persie, will probably have no impact on the Premier League this season, and for once didn’t involve the exchange of millions of pounds.

Instead, it was the announcement on their own website of the launch of a new initiative, MCFC Analytics, and with it, the public release of all of last year’s Opta data for the entirety of the Premier League season. That means every touch of every player from every team in the course of the last season will be documented, categorised and put into one enormous spreadsheet that can be downloaded from the club’s website.

This wasn’t just a dream come true for the stats-minded football enthusiast: a rare chance to test that longstanding but somewhat controversial belief that Man United should have held on to Darron Gibson, or that Ramires is the best player in the league.

City are keen to instigate a lasting change in the way statistics are used in football, by making important data available to engage the ‘analytics community’ and emulate the success of amateur enthusiasts in the statistical revolution in American sports – in short to harness the ‘Moneyball’ effect.

Moneyball is the title of a book by Michael Lewis – turned into last year’s Brad-Pitt-starring film of the same name – documenting the success of the Oakland A’s statistically informed recruitment policy under general manager Billy Beane. This approach relied on the use of ‘sabermetrics’, a term coined by Bill James, the pioneer of statistical analysis in baseball, to designate ‘the search for objective knowledge about baseball’.

In James’ view, not only was such knowledge accessible, but it was also revolutionary, and would undermine traditional subjective measures of player and team value. And so it turned out. By making use of sabermetrics principles in his recruitment of players on a low budget, Beane (played by Brad Pitt in the film) oversaw a record-breaking run of 20 straight victories for the franchise.

By now the principles of ‘sabermetrics’ are well-established in the baseball world. As owner of the Boston Red Sox, current Liverpool owner John W. Henry was one of the first to embrace these principles at a big team, with remarkable success.

However, Henry’s attempts to apply Moneyball principles to football since FSG’s purchase of Liverpool have been almost entirely unsuccessful. Stewart Downing is the obvious example. When Liverpool signed him, Damien Comolli, then Liverpool’s Director of Football, a firm believer in statistical analysis, and, interestingly, an associate of Beane’s, said, ‘We look thoroughly into data before signing players, as well as statistics, and we really think we are getting a big, big asset throughout. Maybe his [Downing’s] talent has been undervalued in English football.’

Ironically his statistics in the Premier League last season are among the most commonly known (and gleefully recited) in the average football enthusiast’s armoury: 0 goals, 0 assists. No fan needs Opta’s full data set to tell them that.

According to the team behind MCFC Analytics, at least a part of the reason for the past failure of performance analysis in football has been the cost of procuring the necessary data: whereas baseball’s statistics revolution was based on the work of amateur hobbyists such as Bill James, and is still fuelled by a thriving community of bloggers and researchers making use of freely-available data; football statistics enthusiasts have had to make do with the paltry post-game statistics made publicly available. Thus it is with the hope of empowering a community of scientifically minded football fans that Manchester City have collaborated with Opta to reverse this situation.

Whether or not the launch of this initiative truly heralds a new era for the use of statistics remains to be seen, and no doubt those who will hasten to suggest that football’s relatively fluid nature makes it far less amenable to analysis than an inherently structured game like baseball have a point.

But that is not to say that numbers have no place at all in the beautiful game: perhaps Comolli and Liverpool were simply looking at the wrong stats, or the right stats in the wrong way; perhaps their analysis was just too unsophisticated. But the fact that such a high-profile misjudgement was possible suggests that there is in fact plenty more to come from the field of performance analysis. Indeed, in baseball, the greatest success of the ‘sabermetricians’ was to identify the ‘right’ stats, and it is something along these lines that the MCFC Analytics team hopes it may be able to achieve.

Check out our other content

Most Popular Articles