Wednesday, November 07, 2007

Why Statisticians Get It Wrong

It may surprise you to know that I am not a big fan of statistics. If you put a big table of facts and figures in front of me, my head starts to spin and a small bead of foam is likely to form in the corner of my mouth. I am greatly pleased that, in spite of the best efforts of the likes of shadowy-named companies like "Pro-Zone" and "Opta", football is still a game that can be enjoyed without having to have a reference book by one's side (for the purposes of elucidation, I should point at that, for writing this blog, I keep a copy of the complete Football League tables and Simon Inglis' still magnificent "The Football Grounds Of Great Britain" by my side - anything else is, frankly, superfluous) or an encyclopaedic amount of knowledge. I can only imagine how, say, an American getting into baseball for the first time copes with the telephone directory sized books which accompany that particular sport.

The work of statisticians is valuable for the purposes of archiving, but occasionally they wander into the realms of opinion, and suddenly all hell breaks loose. This is because the usual work of the statistician is concerned with absolutes - there is no question, for example, that Liverpool are the winners of the most League championships, for example, because, well, they have won it more times than anyone else. There is no doubt that England's record goalscorer is Bobby Charlton, because he scored 49 goals for them, and that is more than anyone else has ever managed. However, when they expand their remit beyond merely recording and start venturing into the altogether murkier world of what is the "best" and the "worst", though, the results make for strangely entertaining reading. Consider, if you will, this list, compiled by The Association Of Football Statisticians, which claims to give a definitive list of the 100 greatest ever footballers. Pele, of course, is number one, but you don't have to go too much further down it before things start getting very strange indeed. The Daily Telegraph, who reported it, were most amused by the presence of Gary Neville in it at number 86, but there else in it that makes one wonder whether the AFS spend so much time analysing their data that they don't actually bother watching the games themselves at all.

The key problem with the list is the very way in which it has been compiled. The list is, as all bad lists of this nature are, hopelessly skewed in favour of the modern era. I can only assume that the inclusion of a large number of these players has come about because of their clubs' perpetual involvement in the Champions League (it would certainly account for the likes of Neville and Jon Dahl Tomasson). No goalkeepers make the top thirty-five, Ronaldo and Romario are "better" than Diego Maradona and Franz Beckenbauer. Ali Daei, the Iranian striker who scored 109 goals for his country, is ranked higher than Thierry Henry or Dennis Bergkamp. Bobby Moore doesn't even make the list at all. I could go on. The biggest single omission on the list is that there are no players at all from before 1945. It's almost certain that statistics from the pre-war era are considerably harder to track down than more recent statistics, so why, if you're going to carry out this sort of exercise, try to claim that this is some sort of definitive list. If your list can't be definitive, then why bother at all?

Ultimately, the Association of Football Statisticians made this list public to publicise their organisation, and it has become commonplace in recent years for these lists to be deliberately controversial in order to garner more column inches in the press (and I do appreciate the irony of my saying that). However, for an organisation which exists, presumably, to chronicle the game, it strikes me as strange that they should make public a list which demonstrates their shortcomings. Football is the world's most popular game in part because it is, by its very nature, transcendental. We can't prove it with facts and figures, but we all know that George Best and Eric Cantona were greater players than the likes of Gary Neville and Kanu. It says something about the very nature of statisticians that they can go public with this list when their methodology is so ridden with holes as to be practically worthless.

Ballhype: hype it up!


Design by Dzelque Blogger Templates 2007-2008