Lies, Damn Lies and Football Statistics

I have a secret passion. A love that dare not speak its name. A lust so dark, so devious, so downright sickening that I have never til now made it public. After a long day at the office, I run home, up the stairs, close my curtains and double lock my door. My eyes dart around the room as the website loads, my face blushes, the photos of my friends and family on my wall stare out at me with dead-eyed condemnation. Finally, the release comes - I can gorge myself on football statistics until morning comes.


There's a stigma to football statistics; that they don't show real ability, that they blur the waters of the game and that they stand sharply against everything about the fluidity and beauty of football. Of course, statistics have their limitations, and it is these I intend to point out, but it is important to note they're of growing import in the game. Quite apart from the sheer joy of noting the pass completion rate of Angel Rangel or swooning over the goals-to-shots ratio of Papiss Cisse, they reveal a stark truth behind misconceptions. To the untrained eye Adel Taarabt may look like the beating heart of QPR, but when we can see that he has scored 1 goal in 75 attempts he looks more burden than brilliance. Likewise, a casual glance at Stewart Downing might lead one to conclude that he is nothing but a mediocre Championship player with one foot, no footballing brain, lacklustre pace and no apparent quality in any part of his game. In actuality, a glance at his stats shows us quite the...alright, Stewart Downing is probably not the best example.

I mention Downing not because I enjoy kicking a man when he's down, or rather not merely because of that, but because Damien Comolli's sacking today casts a sharp spotlight on the use of statistics, or more particularly Sabermetrics, at Liverpool. For those unfamiliar with the concept, the theory was developed to judge players not by traditional scouting techniques, but rather by heavily detailed statistical analysis. The idea is to judge both real merit, untainted by the misleadingly flashy play, and the ability of the player to contribute to a wider, team effort. Thus the quiet centre midfielder with high pass completion and interception rates can be deemed more valuable that the glamorous colleague with the penchant for Hollywood passes.

Formulated in Baseball, as the excellent Brad Pitt film 'Moneyball' recounts, the system has now been transferred to football and, if rumours are to be believed, drove the signing spree Liverpool has embarked on over the last year. In this context the purchase of the reliable winger, big man centre forward and ball-holding midfielders seems understandable as part of a wider team-building effort; eschewing big-name signings for those who would slowly increase the quality of team play according to the hard facts. It seems all very worthy, and indeed expectations for Liverpool's new swathe of players was high at the beginning of the season, with Phil McNulty at the BBC far from alone in tipping them for a return to the top 4.

So what has gone wrong? How have statistics, so beloved by the footballing scholar and so effective in Baseball, so radically failed this season for Liverpool? The answer is not a simple one. Of course, we must consider the possibility that the system has not yet failed; one of my Liverpool-supporting colleagues told me he fully expected Comolli's techniques to eventually bear fruit. This may well prove true, but it does little to stem the discontent of fans and neutrals alike at some truly terrible performances in the present. Instead, I'd tentatively point the finger at two flaws in the statistical ethos applied at the club.

The first seems to be a misunderstanding of the Sabermetrics system, or at least a misapplication. The point is to sign players not only underrated, but also undervalued. Liverpool have contrived to sign players who may well be impressive statistically, but whose value is at best inflated, and at worst absurd. Take poor Andy Carroll; he is young, he offers options going forward and will undoubtedly find some goals. He wouldn't be seen as a terrible mistake at £10m or less, but at more than 3 times that he smacks not of value but of extravagance bordering on insanity, especially when Cisse was available for a quarter of his price and does a similar, but far more effective job on Tyneside. The same is true of Downing, and may well prove true of Jordan Henderson; they could turn out to be fine players in a broader Liverpool playing revolution, but they do not embody the emphasis on value which underpins Sabermetrics.

Perhaps more fundamentally though, Sabermetrics seems an odd system to apply to the game we love. Baseball, like Cricket, is made up of a series of set-pieces. Actions are repeated constantly, success measured easily with little room for variation. Football, by contrast, is a sprawling, fluid game in which almost anything can happen. The potential for mistakes soars, as does the capacity, and indeed need, for players to react on the spur of the moment, as opposed to with well-drilled discipline. In short, basing an entire philosophy on statistics seems fatally flawed when the game is so unpredictable, and when certain facets of ability (drawing a marker, unsettling a defender, inspiring team-mates) are simply impossible to measure.

It is entirely possible that Comolli will be recognised as a forward thinking genius in years to come, especially if the players he splurged money on come good and the team gels to its statistical potential. What seems more likely, however, is that the signings of the club were based on a system ill-suited to football, and even then with its principles poorly applied. The Reds will bounce back, but it seems to do so they need take a leaf out of Newcastle's book; statistics are all well and good, but hard-working scouts can find gems even the best mathematicians miss.