TONY BLAIR is right. The choice now really is “Europe — in or out?” And so it should be. Away with all that round-robin league rubbish. The early-season skirmishing has finished and this week the Champions League turns into the European Cup. About time, too. So now the question is — who is going to win?
Working out the probable winner in a clash between, say, Chelsea and Barcelona, poses all sorts of problems that simply aren’t there when assessing domestic games. The volume of recent encounters between comparable sides that you need in order to work mathematical wonders isn’t available. Nevertheless, the Fink Tank has been on the case.
Each week in The Times, the Fink Tank Predictor gives probabilities for all Barclays Premiership or FA Cup games. Yet those who go to the website can find much more. Here, among many other things, are to be found probabilities for all English domestic games and data on all European clubs. It is this wealth of background data, updated after each round of domestic matches all over the Continent, that allows us to provide probabilities for European Cup games.
Using data from European games to allow us to calibrate different leagues, the Fink Tank is able to produce a European ranking. Unlike similar projects of this kind, this ranking is entirely objective. All rankings involve not only data but weighting — how much value to set by goals scored, for instance, or domestic competitions won. Most use subjective judgments to produce weighting. The Fink Tank ranking does not.
The value attached to different achievements is entirely dependent on its contribution to the ability of the model to predict the outcome of games. The predictions are tested against games that have already happened to check their accuracy.
Advertisement
Fink Tank’s top-ranked European team is Chelsea. This means that, on a neutral ground, Chelsea would be favourites to beat any side in Europe. The side ranked second, Manchester United, are only 78 per cent as good as Chelsea. Barcelona, who are third, and Villarreal, in fourth, are 71 per cent and 63 per cent as good respectively. Critics of Peter Kenyon, the Chelsea chief executive, may note that the European ranking last year had Valencia in first place. Claudio Ranieri, the Chelsea refugee, has guided the Spanish club to twelfth spot.
Using the rankings allows predictions for different games. But beyond this, of course, the tournament draw is unpredictable. So Dr Henry Stott and Dr Alex Morton have simulated the final stages of the European Cup 100,000 times to provide the probabilities of progressing and winning the tournament.
Despite having a tight tie against Barcelona, Chelsea are clear favourites to lift the European Cup. There is an eyebrow-raising 50 per cent probability of an English side winning overall. The data also allows a judgment to be made on which teams benefit from moving to the knockout stage from the league competition. Are there any knockout specialists? And is it true, as has been suggested, that English teams prefer the league rounds?
The answers are no and no. There are no teams whose improvement or deterioration in the knockout stage is large enough to be statistically significant and, naturally, that includes the English sides.
One odd feature, however, is that, when taken as a group, performance in the league stage is a poor predictor of performance in the knockout stage. In other words, a team doing better in the league stage than domestic results should not be expected to repeat this in the knockout stage.
Advertisement
In fact, oddly enough, overperformance in the league stage is negatively correlated with performance in the knockout stage. Perhaps, although this is speculative, sides are overconfident after good early performances. They should get to the Fink Tank website. Then they wouldn’t make that mistake.
finktank@thetimes.co.uk