Is Everything You Know About Football Wrong? Why analytics won’t tell you everything

The news from the World Cup that among the victorious German sides arsenal was a sophisticated software analysis tool named Match Insights which is able to convert match data into simulations and graphs marks yet another triumph for the art of using in-depth statistical analysis to gain insights and improve performance.

Data analysis in football is of course nothing new. In 1968 Charles Reep co-authored a statistical paper Skill and Chance in Association Football in the journal of the Royal Statistical Society. The fruits of many years of having a pencil and notepad handy this was full of formulae seeking to unlock the secret of the most effective way of playing; Reep’s insights included the famous observation that most goals come from moves of three passes or less. Reep even achieved some influence – not least over Norwegian football – and is credited with popularising the long-ball game. Though that is perhaps also the reason for his downfall; long-ball being a style which wins few friends (though the writer Jonathan Wilson would later point out a crucial flaw in some of Reep’s analysis, namely that he had overestimated the actual efficiency of the three pass or less rule).

In any case Reep remained something of an oddity and the analytical approach firmly in the margins. That was until fairly recently. Footballs quantitative turn can be in part traced to events in another sport. It was the early noughties when the Oakland ‘A’s, an unfancied baseball team with a low budget exceeded all expectations. Their secret, captured in the book and film Moneyball: The Art of Winning an Unfair Game, was to use analytical techniques – often referred to as sabermetrics – to find players which had been hitherto undervalued by the market by finding new and better measures of ability.

It was nothing short of a revolution. Billy Beane the Oakland ‘A’s manager hailed the analytical approach as not just the key to improving performance, but also being able to deliver meritocracy and diversity whilst ending the insider/outsider divide in baseball. In football proponents of the data approach have also sought to challenge accepted wisdom; Chris Anderson and David Sally’s 2013 book, The Numbers Game being provocatively subtitled Why Everything You Know About Football is Wrong. Meanwhile clubs have clamoured aboard the analytic bandwagon; Last season it was reported that every single Premier League club was employing an analyst with the club who would go on to the championship, Manchester City, employing eleven.

That the German team used such analysis is therefore of little surprise. But there is a danger in over-emphasising the power of the technique. One issue is that football possesses a great deal of complexity, more so than a game such as baseball where the game is built around the key interaction between pitcher and batter. With 22 players on the field, all possessing free-will it can be hard to make accurate forecasts and simulations, not least when other factors come into play such as the effect of the crowd, psychology, the weather or even the performance of match officials. Furthermore in a game of football any one players individual performance data makes little sense outside of the context of their role in their team; for instance a player may be effective in one teams system, but then upon joining another team struggles to make any impact.

Although Beane, who points to the same issues in baseball, believes it is one which improvements in technology can, in time, overcome the issue remains that the problem with any analytical model is that it can only account for what is in the model. There are a multitude of events which will be outside the model such as the impact of a cold virus sweeping through the camp just before a game, or a key player becoming unsettled by a transfer bid. The success of a club is only partly achieved on the pitch, the rest – the financing, the corporate deals, the youth set up is also beyond the scope of match analytics. That is not to say that analytics – and God forbid heatmaps – are not worthwhile. They are. They are a valuable tool in reading and understanding the game. But that is just what they are – a tool. One of many. Alone they cannot hope capture football in all it’s glorious complexity.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s