How much do Stats tell us about driver performance?

[MOWOG]

When I read the first half of your response, SF, I though "Uh oh. I'm about to get blown out of the water!" But then I read the rest of it and said "Phew. Someone who knows statistics actually agrees with me."

In my personal experience, based upon more than 6 decades of travelling the road of life, I find that "Figures lie and liars figure" is a pretty good aphorism, whether talking about sports, politics or economics. More times than not, good old fashioned common sense seems to be as effective as reams of data.

[sennafan24]

When I read the first half of your response, SF, I though "Uh oh. I'm about to get blown out of the water!" But then I read the rest of it and said "Phew. Someone who knows statistics actually agrees with me."

I would not say I am a expert, like I said I am "studying" but I know enough to not take stats as gospel, as you said above critical thinking is a good thing. I can kinda agree with the "its the results that matter" and maybe I was a bit too simplistic in saying results/stats are one in the same. For example, in a darts match a winning player can average a score of 90 according to the stats, and the losing player can average 100 plus. That is how stats can be misleading, and not represent the results as it were.

To bring it back on topic, I believe it can happen with F1 for the reasons of "context" like the poster G-Berger said above. Without mentioning drivers names, I think some closely fought title wins have been more impressive than some years where the driver has won the WDC by a mile, due to the context of how the WDC's have been achieved. Some years where drivers have produced the best stats have done little for me, due to the context in which they achieved these stats in.

[Stradivarius]

To answer the question posted initially: The stats tell us everything that we can possibly know about driver performance. I would like to ask a question in return: Is there any way to evaluate drivers, other than stats? I can't think of any. To me, stats is information and the information about driver performance is reall nothing else than stats. To use the same example as in the opening post, Prost won the title in 1989, but some people feel that Senna deserved the title more, because he beat Prost 9 - 1 in those races that both of them finished. This isn't a matter of either putting weight on stats or on other things, it is about weighing different stats against each other. If you exclude all the races when they ran into trouble, Senna seemed better. But that is simply because Prost's strength was his ability to stay out of trouble, while this was Senna's most obvious weakness. What these particular stats show is that Senne was quicker, but F1 isn't only about being quick, and it never has been.

There is always different opinions about what qualities are most important. A good example is speed and consistency and the example of Senna and Prost illustrates this well. Senna was better at one thing, while Prost was better at another thing and then people let their feelings decide what should be given most weight, i.e. the balance between speed and consistency. But the way I see it, the rules clearly define this balance. In 1988 and 1989, consistency wasn't as important as today, since the 6 worst race results were excluded from the championship. This meant that Senna won the championship in 1988. But in 1989 Prost won the championship and the simplest interpretation of this is that Senna simply wasn't good enough that year. He failed too many times to score high point's positions. You may blame it on reliability or other things and say that he was quicker, but that doesn't change the fact that reliability issues always have been part of the sport and a driver who is quick, but inconsistent will suffer more form poor reliability than a driver who is more consistent, but not as quick. In 1988, Prost could have afforded 4 additional reliability issues and he would still have only 1st and 2nd places counting in the championship. But Senna was on the limit, so if McLaren had been less reliable, Prost would probably have won it. But under the actual circumstances, Sennas qualities were worth more points and he won the championship, just as deserving as Prost did the year after.

One could of course also say that with today's points systemt, Senna wouldn't stand a chance against Prost, and use that as an argument that Prost was better. But it doesn't make sense to judge the drivers on other terms than those they were competing under. The old point's system encouraged drivers to take risks and be quick rather than cruising home in 2nd place. And that is exactly what Senna did and that is why he won in 1988. So in my opinion, we all have to accept that the rules decide what is most worth and then it's hard to argue with the stats. For the record, stats may very well contain information about points lost due to reliability issues, team orders or other things that is considered outside the control of the driver. Stats may also contain information about how many points the team mate was able to score with the same car, or other things that can give indications on how strong the car is compared to other cars. It is all stats.

What I think is an interesting question is which stats should be considered important when evaluating a driver. Is fastest lap important, for example? Is qualifying important? What about podiums? Points? Race wins? Titles?

[sennafan24]

But that is simply because Prost's strength was his ability to stay out of trouble, while this was Senna's most obvious weakness.

What I think is an interesting question is which stats should be considered important when evaluating a driver. Is fastest lap important, for example? Is qualifying important? What about podiums? Points? Race wins? Titles?

I enjoyed your post, but I disagree with the top part. Senna did tend to stay out of trouble in 1989, just his engines blew up which was out of his hands. 1988 Senna did get himself in some trouble if I am fair, like Monza for example.

I agree with your general point of playing towards the rules, in 1988 I believe Senna raced in a different way than in 1991 slightly to accommodate the rule change. All 13 times he finished in 1991 he was in the points, and all but one or two of those were podiums, even with the faster but less reliable Williams to challenge him. He showed he could be as consistent as the best of them that year.

The second part is one I am also interested in. I have always the valued the "finished first when both finished" stat more than most do, as you could say it takes out the variable of reliability. However, the reasons for not finishing are not always reliability based due to crashes and collisions.

Like I said I do not put much weight into titles, Stirling Moss's legacy is testament that amount of titles do not tell the full story of a drivers career. I do not really value fastest race lap, as in F1 you get no points for it, like Vettel said for him its just for "satisfaction"

[dren]

What I think is an interesting question is which stats should be considered important when evaluating a driver. Is fastest lap important, for example? Is qualifying important? What about podiums? Points? Race wins? Titles?

Yes, just what I was going to post. Stats most certainly can and do tell us about driver performance.

Order of importance:

1. Finishing position (direct relationship to points)
2. Qualifying position
3. Driver related DNFs

These would have to be looked at as against the field and also against their teammate. You'd have to try and level out the field performance wise, so you'd have the top team and each below it would receive a multiplier. This is hard to do, but it would be fun to try. A team with two bad drivers would give a car a higher multiplier, so it could be skewed a bit. If one of the two bad drivers significantly outperformed the other, the high car multiplier could make him look much better than he is. Only the teams will know what the car is capable of in simulations and how the drivers perform vs that. That would be good data to have. Although I think it's safe to assume the top drivers are getting close to the max out of the car, so we can use them to figure what the car's potential is vs the others.

You could dive into all sorts of statistics on just overtaking. That would be interesting.

All

[Tim.Wright]

Stats won't tell you anything about a driver because they are a result of the driver/car combination. To separate out just the driver component you need to make some corrections for the car performance.

At this point it all gets extremely hand wavy and IMO a waste of time. Like all driver discussions.

[sennafan24]

Stats won't tell you anything about a driver because they are a result of the driver/car combination. To separate out just the driver component you need to make some corrections for the car performance.

This gentleman here - http://grandprixratings.blogspot.co.uk/ has tried to form analysis based on car/driver combination, basically rating the drivers performance based on what car they had to play with..

The guy who does it is a smart and nice guy, I used to post with him on another F1 forum. His methods are not perfect as he will admit, but in my opinion they are quite reflective.

[SectorOne]

Stats are good for facts. Things you can quantify. Like, who´s the most succesful. easy. Schumacher is by far the most succesful.
Or who´s fastest of the two teammates, who´s done the best job this season etc.

But the best? There´s no numbers in the world you can really legitimately lean on for that. It´s really about knowledge and visuals at the end of the day.
You need to know the back story and you need to see the Quali/Race yourself to get a nice big picture and then give your opinion on that.

[sennafan24]

But the best? There´s no numbers in the world you can really legitimately lean on for that. It´s really about knowledge and visuals at the end of the day.
You need to know the back story and you need to see the Quali/Race yourself to get a nice big picture and then give your opinion on that.

=D>

Well put

[Tim.Wright]

Stats are good for facts. Things you can quantify. Like, who´s the most succesful. easy. Schumacher is by far the most succesful.
Or who´s fastest of the two teammates, who´s done the best job this season etc.

I'd strongly disagree there. Stats are not good for demonstrating anything. You can "prove" almost anything you want by choosing the right stats. Its been shown numerous times on these boards even.

For motorsport its even worse. Like I said before, you have the car factor but also strategy, weather, rule changes, collisions, important changes of personel etc.

But the best? There´s no numbers in the world you can really legitimately lean on for that. It´s really about knowledge and visuals at the end of the day.
You need to know the back story and you need to see the Quali/Race yourself to get a nice big picture and then give your opinion on that.

The best? Every discussion I've ever seen on this has never even reached a consensus on what "the best" actually means.

[sennafan24]

I'd strongly disagree there. Stats are not good for demonstrating anything. You can "prove" almost anything you want by choosing the right stats. Its been shown numerous times on these boards even.

I agree.

But the stat Sector one stated about Schumi being the statistical best is a fact , due to Schumi having the highest amount of WDC's, Poles and Race Wins.

In terms of quantity stats, Schumi is the king!

[Juzh]

Order of importance:

1. Finishing position (direct relationship to points)
2. Qualifying position
3. Driver related DNFs

All three points are held by Vettel on the current f1 grid.
1. Not entirely sure what you meant by this one, but vettel had collected the most amount of points in the last 3 seasons, hence making him WDC.
2. Vettel has the most pole positions.
3. Vettel holds the record for longest run of non-driver induced dnf in history. 61 races and counting.

So by that stats, Vettel is clearly miles better than anyone else on the 2013 grid

[sennafan24]

So by that stats, Vettel is clearly miles better than anyone else on the 2013 grid

http://www.youtube.com/watch?v=xF1nze1Oxxc&t=8m13s

[GitanesBlondes]

You may blame it on reliability or other things and say that he was quicker, but that doesn't change the fact that reliability issues always have been part of the sport and a driver who is quick, but inconsistent will suffer more form poor reliability than a driver who is more consistent, but not as quick. In 1988, Prost could have afforded 4 additional reliability issues and he would still have only 1st and 2nd places counting in the championship. But Senna was on the limit, so if McLaren had been less reliable, Prost would probably have won it. But under the actual circumstances, Sennas qualities were worth more points and he won the championship, just as deserving as Prost did the year after.

The thing I would say is that reliability isn't always down to driver input. It depends on what the circumstances of the failure were, and whether that can actually be attributed to a driver error or not. Some cars are just poorly designed that they can't handle a great deal of anything - BAR-Supertec 1999. An engine can be poorly designed, or something could have changed with the engine that no one was aware of - Andrea de Cesaris' Jordan 191 Cosworth engine running out of oil because the mechanics weren't told more oil needed to be added by the Cosworth guys. Most of the guys driving in F1 are consistent. They may not be consistently fast, but they're generally consistent with the speed that they drive at.

Senna suffered plenty of misfortune in 1989 that I don't think can be proven was due to being on the limit. His differential failed when the race was restarted. Can that be proven that it had anything to do with being on the limit? Senna also did not drive on the limit for the duration of any race. It's such a misnomer...he usually drove extremely hard at the beginning on cold tires just to open up a gap that he could then control for the rest of the race. He drove much smarter than anyone seems to realize anymore. Not quite as smart as Prost perhaps, but his racing smarts were far better than seems to be common knowledge now. Had things gone differently, would he have won the championship? Who knows. Prost won, and there's no way to know otherwise.

Some guys have bad luck some years that has little to do with anything they've done to the car. Niki Lauda had a ton of reliability problems in 1985. The same guy who won the title in 1984, and was known for his consistency managed to have the worst luck of any full season in his career. Did he suddenly forget how to drive from 1984 to 1985? Doubtful. This is the guy who won at Monaco in 1976 and said the only difficulty he had in winning the race was to not fall asleep from sheer boredom.

[Stradivarius]

Stats are good for facts. Things you can quantify. Like, who´s the most succesful. easy. Schumacher is by far the most succesful.
Or who´s fastest of the two teammates, who´s done the best job this season etc.

I'd strongly disagree there. Stats are not good for demonstrating anything. You can "prove" almost anything you want by choosing the right stats. Its been shown numerous times on these boards even.

Incomplete stats can be used to fool some people into believing almost anything. Complete stats, however, cannot be used that way. If you compare two drivers by looking at their points, there is only one possible outcome of that comparison. The driver with more points will always come out on top. What happens a lot is that people search for arguments to exclude certain results and then they compare the points. Then it is possible to change the order, but only because the stats are incomplete. Another thing that occurs is that people choose not to look at the points, but other things like fastest lap, qualifying, number of wins and so on. If your favourite isn't in front on one, you simply emphasize something else. But it is possible to be objective and acknowledge the fact that fastest lap and qualifying doesn't really mean anything unless it's reflected in the points. The competitors are trying to do as well as possible in the competitions, not in the forum polls. Then it's most natural to me that their performance is also evaluated based on the actual criteria of the competition. Points matter. Qualifying doesn't matter, except that it influences the points somewhat.

For motorsport its even worse. Like I said before, you have the car factor but also strategy, weather, rule changes, collisions, important changes of personel etc.

So if you try to account for all these factors, aren't you still looking at stats? I think everyone agrees that Alonso is most probably a better driver than Massa. But how can we say that if we don't rely on the stats? In the end of the day, I really don't see that you would have any information at all about driver performance if you remove all the stats.

But the best? There´s no numbers in the world you can really legitimately lean on for that. It´s really about knowledge and visuals at the end of the day.
You need to know the back story and you need to see the Quali/Race yourself to get a nice big picture and then give your opinion on that.

The best? Every discussion I've ever seen on this has never even reached a consensus on what "the best" actually means.

That is very true. But I would like to ask SectorOne, who claims that it's all about knowledge and visuals: What type of knowledge are you talking about? How can you know that one driver is better than another if this isn't reflected in the stats?

The way I see it, it's not about knowledge at all. It's all about feelings and nothing more. If your feelings tell you that one driver is better than another, you will automatically, probably both consciously and unconsciously, try to justify your feeling and look for information that support it, while you will be less likely to accept information that points in the oposite direction.