Doing The Maths

This week, I'm going to play true to my reputation as a bit of a statistics anorak and do a bit of simple analysis to work out a couple of things about the pattern of this year's Formula One championship. Things which might not necessarily be entirely clear from the simple headline figures of who has scored the most points.

What I have done is calculated the extent to which, over the season, each driver has been faster or slower than the average fastest lap for the series. There are a number of limitations to this as an analytical tool. For one, it tells you only how teams and drivers are going when their cars are at their fastest. If they have a particular problem on worn tyres, or with heavy fuel, then these figures will not show it. Secondly, it is systematically biased in favour of those drivers who have tended to run shorter fuel stints, as their best laps will tend to be quicker, all else being equal. Thirdly, I have had to remove from consideration times set at races where, due to early retirement, the driver never really got in a representative laptime. There is a subjective element to this. Few would disagree with removing Klien's times from Australia (where his best lap was ten seconds or more off the pace) or Albers' time from Bahrain (he went just two laps before retiring), but what about Webber at Malaysia? Well, I took him out - he never got as far as the first fuel stops. And what about Fisichella at San Marino - did he ever get a clear lap in? Well, I kept him in - removing drivers' times on the grounds they spent most of the race being baulked by slower cars takes subjectivity to whole new levels - especially given that I didn't have the chance to see the whole race from his cockpit.

So here, without further ado - are the results:

1. Fernando Alonso +1.61s
2. Michael Schumacher+ 1.34s
3. Kimi Raikkonen + 1.27s
4. Giancarlo Fisichella + 1.00s
5. Felipe Massa + 0.96s
6. Juan Montoya + 0.91s
7. Jenson Button + 0.70s
8. Mark Webber + 0.54s
9. Nico Rosberg + 0.38s
10. Nick Heidfeld + 0.32s
11. Ralf Schumacher + 0.32s
12. Rubens Barrichello + 0.21s
13. Jacques Villeneuve + 0.25s
14. Vitantonio Liuzzi - 0.01s
15. David Coulthard - 0.08s
16. Scott Speed - 0.10s
17. Jarno Trulli -0.20s
18. Christian Klien - 0.23s
19. Christijan Albers -1.26s
20. Tiago Monteiro - 1.59s
21. Takuma Sato -2.84s
22. Franck Montagny -3.13s
23. Yuji Ide +5.28s

One of the most striking things is how the top three exactly reflects the top three in the current drivers championship. Further down though, there are odd surprises. The Williams turn out to be a good deal quicker than their qualifying performances or their points tally suggests, with Webber rated 8th overall, ahead of Barrichello and both Toyotas and Sauber BMWs. The Toro Rossos are in fact marginally quicker in race trim than the Red Bulls, and within a few tenths of the BMWs and the faster Toyota, suggesting perhaps that the Cosworth V10 remains an unfair advantage.

The comparisons between team mates are instructive too. The biggest differential was of course between Sato and Ide, but the 0.6s difference between Alonso and Fisichella would make grim reading for the Italian, and rather serves to underline the feeling that he has underperformed in the Renault this year. Ralf Schumacher appears to be doing well in the Toyota too, in an unobtrusive kind of way, comfortably faster than team mate Trulli. By contrast. there is almost nothing to separate the BMW drivers (0.07s between them) or the Toro Rosso newcomers, who are separated by just 0.1s, despite Liuzzi having appeared to have the edge over Speed for much of the season so far.

Some of the backmarker teams might find it useful for working out just how much more pace they've got to find. For instance, Super Aguri have got about 4.5s to find (less than I might have thought, although tyre degradation etc is doubtless worse on the old Arrows-hack) while Midland need about 3s, and Williams would be right in the hunt if they could find about 1s.

So, the definitive answer to who is quick in 2006? Nah, you need to read the season as a whole for that, but an interesting way of looking at it? Well, I hope so, my eyes have gone all screwy from mucking around with the Excel spreadsheet this evening, and I hope it wasn't entirely a wasted effort.

[NB - Anyone who wants a copy of the raw data from which this has been calculated (you can error-check it if you like) simply leave a comment below with your email address, and I'll send it on to you]