The fastest driver: a systematic approach
Posted 14 October 2000 - 00:21
It has ocurred to me tht a similar technique could be used tocompare the performance of drivers. Suppose, for example that drivers A and B are team mates. If we can accept that they had equal equipment (not always the case) we can perform a direct comparison of their respective performance. Suppose now that driver B changes team and that his new team mate is C. We can perform a similar comparison of B and C; this will enable us to compare a with C even if they never drove in the same team.
An example might illustrate this better. We will use Fangio as a baseline and arbitrarily give him an index value of 100. In 1950 and 51 Fangio and Farina both drove for Alfa Romeo. THe best measure of a driver's speed are practice times. We can compare Fangio's and Farina's practice times in the races they drove togeether. We find that on average, Farina was 0.94% slower than Fangio. this would give him an index value of 99.06.
In 1953 and 1953 Farina drove for Ferrari. We can perform a similar comparison of Farina's performance in those years with that of Ascari. In this case we find that Ascari was 0.91% faster than Farina. This will give Ascari an index value of 99.95.
We now have a comparison between Fangio and Ascari, namely that the Argentinian was 0.05% faster. THe comparison is independent of the performance of the cars they drove.
We can use the same technique, and cmparison with Fangio to give figures for Moss (1955), Collins (1956) and Behra (1957). Moss will give us a figure for Brooks (1957/58) and Brabham (1959, if we can assume that the Walker Cooper and the works car were equal in performance). We could even venture tentatively back in time using our index value for Farina to give figures for Varzi and Wimille. In fact,it is only the absence of reliable data which would prevent us ultimately being able to answer the question of whether Szisz was a faster driver than Schumacher
The only calculation I have done so far is the one in the example. I would like to know what other people think of the technique before I venture further.
Posted 14 October 2000 - 03:36
In the days of skating on rubber, Denis Jenkinson thought that Ascari had a finer touch than Fangio when 4-wheel drifting. Which opinion presumably gave him many opportunities to share a beer or two. Maybe that's why we like such discussions.
Posted 14 October 2000 - 04:03
As soon as I saw the heading - and knowing of your interest in Pomeroy's index - I knew what this was going to be about.
I think it is a very interesting exercise. Of course none of these things based on statistics is perfect and ultimately there is no definitive answer, but it does put it into some perspective.
Far better than various people's opinions based on litle other than passion.
But you're looking here, I think, at a driver's sheer pace.
I always have believed that a grand prix driver's job is to win races - although since 1950 the focus has shifted more and more towards winning championships rather than races.
Grand Prix races are won on a mixture of pace and intelligence. How do you quantify this? Probably best by average points accumulated. But this obviously favours those who were "in the right place at the right time" regarding being in the best car - or very near to it.
The Pierluigi Martinis of the world would come up with a ranking far below their true potential.
There also are a lot of drivers who have sheer pace, not much in the way of results.
Personally, while I always have found the likes of Gilles Villeneuve and Ronnie Peterson good entertainment value and can admire their sheer car control, when it comes to doing the job they fall far short. You don't win a lot of races by shredding the tyres and bouncing off the scenery.
I have never made the time to analyse it, but I always have had the suspicion that Tazio Nuvolari was more in the Villeneuve (Snr)/Peterson mould rather than the Fangio/Ascari/Prost category.
That is, some absolutely brilliant wins, but a lot of broken cars and lost opportunities.
Before I raise anyone's ire, let me again state that this is only an underlying feeling I always have had; I have never made the time to try to prove or disprove it.
But think of Achille Varzi, Nuvolari's main rival. He rarely (never?) crashed a car apart from that final, fatal time (and then after he had racked his body and brain by using drugs) and he won a hell of a lot of races.
It is interesting that it is the spectacular drivers who seem to accumulate hordes of worshippers; the true artists are so good they sometimes can be boring (Prost, for example). They make it look all too easy.
Fangio was spectacularly efficient - does that tell us something about the cars and the circuits? Modern tracks with no bumps, cars with high aero downforce, and modern race tyres have killed much of the spectacle.
I wonder how spectacular Prost might have looked in a 250F Maserati or an Alfa Romeo Tipo B?
Since Hans has studied the pre-1950 era so thoroughly, I would like to hear his opinion on this.
Posted 14 October 2000 - 05:54
I think Roger has here an excellent idea and I really would like to see the numbers. For starters, it would be nice to compare the drivers of the Fifties and Forties. Once we can look at the result, the discussion can begin. Probably something good will come out of this.
Roger, I think you can only go back to 1933 when, for the first time, reliable practice times were available at the Monaco GP grid, assembled according to best lap times.
Posted 14 October 2000 - 06:56
But in several cases you'll end up with the following: Driver A was 10 points better than driver B, who was 10 points better than driver C , who was 10 points better than driver A and you stand there as clever (or dumb) as before.
By the way, I did not use one particular driver as baseline but worked out the mean value of the practice times of most drivers for each race, leaving out only the top 4(?) and bottom 4(?) results.
Posted 14 October 2000 - 11:16
First, one can obtain result of comparison between two drivers in many ways, each of them giving, if not significantly, different result. Perchance it could be accounted for (giving mean value of all possible results), but we should consider the time also. If a certain driver was off-form in particular season, his performance should reflect on results of his teammate, and subsequently of all comparisons derived from his performance (and perhaps overall career performance mean value could sort that out).
And the problems of drivers getting older (if not more prudent in the process). For example G.Hill, at the beggining of his career, when compared to '50 drivers, must (in my mind, at least) performed better (because of his ,relatively, young age and eagernes for recognition) tha in his latter days (when compared to the drivers of '70)...
Many a issue can be raised at the method (as well to any other method), but I belive it could provide interesting data. I think it quite a good idea.
Posted 15 October 2000 - 10:58
Posted 15 October 2000 - 11:34
Second, it is systematic, in that it follows a system, but it is emphatically not subjective. you have to make judgements as to whether a driver ws under-performing because he was at the start or the end of his career and whether he had a bad day due to mechanical or personal problems. Somebody once said that tere is no fact concerning the history of Motor Racing that cannot be proven to be manifestly untrue. Anything that I put forward in this analysis is not a fact, even in that definition of the term.
However, if you believe that Ascari was far beyond Fangio, it would be useful to know why. Very few of us can have seen Ascari race and it is far too easy to form an opinion based on legend or on the opinion of professional writers.
I have tried to apply the method to the drivers of the forties, but failed due to lack of data. I have the Sheldon book, but there's just not enough races to perform a valid analysis.
Posted 15 October 2000 - 12:00
It might not necessarily prove anything, but it might disprove a legend or two and it is all grist for the mill.
And, while "the fastest" and "the best" are two very different things, it is interesting to know both.
Posted 15 October 2000 - 22:39
I must say after seeing Jimmy in the Tasman series in my youth and being stunned at the sheer effortless poetry of his driving I have no problem with this result.
Also it mut be remembered that at Monza in 1967 Fangio himself proclaimed Jimmy as the best racing driver in the world.
Posted 15 October 2000 - 23:15
Posted 15 October 2000 - 23:21
There was one writer who said that 'races surrendered to Jim Clark'... no more apt description of how he often won, don't you think?
Posted 16 October 2000 - 00:10
As for Jimmy words cannot express the depth of his talent, I doubt we shall ever see his like again.
Posted 16 October 2000 - 00:19
Posted 16 October 2000 - 04:28
Did I ever mention that I didn't really like Clark?
Posted 16 October 2000 - 04:53
Posted 16 October 2000 - 05:15
Posted 16 October 2000 - 09:04
Jim Clark - 372.6
Juan Fangio - 369.3
Jackie Stewart - 362.7
Michael Schumacher - 348.7
Ayrton Senna - 347.6
Alain Prost - 342.6
Alberto Ascari - 340.8
Stirling Moss - 339.6
John Surtees - 332.0
Jack Brabham - 330.3
Posted 16 October 2000 - 12:20
I've tried on many occasions to do something similar, but you always run into boundaries. By comparing teammates you rule out the major differences in cars (not the fact that driver A gets new parts sooner than driver B, but that's something we cannot rule out because those details aren't very public). However, there's also the experience these drivers have in different years. When Prost & Lauda were at McLaren, Lauda had 15 years of experience and Prost 5. When Prost later teamed up with Senna, he had 7 years experience and Senna 4. There's a vast difference between Prost anno Lauda and Prost anno Senna, so in fact we're comparing A to B and A+ to C. This could possibly be overcome by making an analysis of the average careerpatterns in all of F1 and setting that off against the careerpattern of the drivers you're comparing.
But even that's not the end of it. Take 1992 for instance. When you look at the ice cold result sheets, Mika Hakkinen didn't cause anything more than a few ripples in the F1 pond. However, when you consider the competition of a) the overwhelmingly strong Williams, b) McLaren with the experienced duo of Senna & Berger and c) the fast rising Benetton star Schumacher, Hakkinen took more out of the dying Lotus than anyone could think was in it. So there's also the level of competition a driver battles against in each of his seasons.
All in all, to achieve an honest comparison between drivers of all times, one would need an incredibly detailed and complex model to evaluate all differences in circumstances. For all we know the greenhouse effect might influence F1 (more wet races = better results for Schumacher, unless he gets a dog of a car like Alesi)
Maybe this is exactly what F1 Racing have done, I don't have the issue on hand at the moment. It'll be hugely interesting, but I fear nearly impossible.
Posted 16 October 2000 - 13:14
There also is the way the chemistry works within a team. Sometimes a driver and team "click" and the driver's very best efforts emerge - way beyond anything he had done or could do elsewhere.
Jim Clark at Lotus is an example and I always have believed that the difficulty of judging Clark's true ability as a GP driver is hampered by not having seen him in any other team.
And there are examples of drivers wilting completely when teamed with the likes of Senna or Schumacher - largely because the team has swung completely behind one driver to the exclusion of the other.
In Australia we have a classic recent example in the dominant V8 Supercar outfit, Holden Racing Team. When Mark Skaife moved there from Gibson Racing, HRT was 100 per cent behind Craig Lowndes and many of the personnel had previously moved to HRT from Gibson because they didn't like Skaife.
Lowndes won everything and Skaife really struggled for a couple of years to the extent that some thought he had seen his best years.
Exactly what changed, I have yet to discover, but this year Skaife and the team gelled and Skaife has had a great year, demonstrating his former peak form, while Lowndes has been struggling a lot (but not all) of the time.
On which year's results would an analyst rate these two drivers against one another?
However, I still think this is an interesting exercise of Roger's. He strikes me as a level-headed person who knows how to treat and to analyse statistics. He's not trying to give us the definitive answer, just some more clues.
I, for one, am interested in the outcome - as I am in the list Bernd has shown us. I don't necessarily agree, but I find it interesting.
Further difficulties in this sort of analysis include things like Alain Prost's relative lack of interest in gaining pole positions, preferring to concentrate on setting up his car for the race, compared to Ayrton Senna's near obsession with proving he was the fastest by taking pole after pole. Check the fastest race lap statistics and it gives a very different picture.
Another problem is in trying to analyse a driver like Jack Brabham, who never went any faster than necessary to win a race but who, like many other top racers, was capable of blinding speed when the need arose.
Posted 17 October 2000 - 00:07
And if they've now changed their approach, is it possible Skaife has been undergoing a personality change?
There was some room for change, I guess...
Posted 17 October 2000 - 08:21
Posted 17 October 2000 - 10:26
Posted 17 October 2000 - 10:28