Winning gap

[MissTheApex]

Hi apex,
i just found this forum and must say i'm amazed by the statistics you can read here, very interesting!

I've gone through the statistics & analysis forum, but wasn't able to find anything about how the gaps between the leader and 2nd/3rd/4th... place have evolved through the history of F1.

Does anybody here in this forum have some numbers on this subject? For me, the average gap seems to be an indicator of how much better the winning car/driver combo is than the others and my impression is that the gap got smaller and smaller from 1950 to today.

Thanks for any help on this.

 

[teabagyokel]

Hi MissTheApex

This article is worth a read on the subject, from Galahad:

http://cliptheapex.com/threads/top-15-best-new-car-debuts-in-f1.2413/#post-41268

I would agree with your general trend, but I don't think we've got any numbers for it. Could be an interesting line of inquiry though.

 

[teabagyokel]

I've had a look and this is the historical distribution of races where a driver lapped the field:

... it oddly happened more in the 1960s than the 1950s. Its worth noting these gaps don't always suggest epic superiority - Patrese's win at Monaco in 1982 being the most obvious one, only Prost's disqualification at Imola in 1985 left de Angelis a lap clear and Damon Hill's 2 lap victory in Adelaide in 1995 was in the teeth of extreme unreliability from the other top runners.

 

[FB]

I'm not sure whether such figures would give any information as to how close, or otherwise, the racing was. I recall Senna used to slow down towards the end of a race to put less stress on the car (usually with Murray Walker squawking about how he must have a problem) and I'm pretty sure Vettel could have finished many races this year with a larger gap than he had but, again, he was "taking it easy". It would be a pretty massive task to take on if someone wanted to have a go at it and you assumption that it has got smaller is probably correct - then we would get into all sorts of debates about the general reliability of cars and number of retirements etc, etc.

Anyway, welcome MissTheApex , enjoy your stay here and keep firing those awkward questions .

 

[Brogan]

I have all the data in various spreadsheets and databases.
If/when I get the time I'll put something together

 

[MissTheApex]

1950        2nd                 3rd                 5th                 10th
Average     00:44,3    0,42%    03:01,0    1,82%    07:18,3    4,43%    13:52,0    9,35%
StdDev      00:44,8    0,40%    02:33,9    1,42%    04:11,1    2,20%    17:26,4    11,18%

2013
Average     00:11,5    0,20%    00:17,5    0,30%    00:36,8    0,63%    01:07,4    1,17%
StdDev      00:09,7    0,16%    00:11,9    0,20%    00:16,1    0,29%    00:22,9    0,42%

To start i compared the 1950 and 2013 season. This is of course only a very small part of all the F1 seasons, but should just illustrate what i want to analyse. The table shows the average difference and the standard deviation of the 2nd, 3rd, 5th and 10th place in relation to the winning car. The first number is the gap in mm:ss,0 and the second is the the gap divided by the overall race time. If a car was lapped, i calculated the gap to the leader as "total race time" divided by "laps behind".

This comparison shows a huge decrease in the gaps, as i expected. I will continue and i think i'll do the 1960,1970,1980,1990,2000 and 2010 season next. Let's see how the then more complete image does look like.

 

[MissTheApex]

Correction: The gap of a lapped car is "total race time" / "race laps" x "laps behind"

 

[Galahad]

I've done this analysis for qualifying times, not quite the same thing I know but hopefully of interest;

There has certainly been a substantial closing of the competition since the very early days of the world championship, culminating in the tightest spread for the top 10 positions in history in 1975. 1968 was one of the wettest seasons in history and that might explain that significant blip. From the late 1970s, the adoption of turbo engines caused the field to spread again into the 1980s. The dominant McLarens of 1988-89 and Williams' of 1992-93 produced the biggest field spreads of modern times.

Since the electronic "gizmos" were banned in 1994, the field has been close and has generally got even closer, with some blips from season-to-season. The closest 1st-2nd gap was in 2003, while the closest 1st-3rd and 1st-5th were in 2009.

There is some missing data for the top 10 spread in recent years; since the three-stage qualifying was introduced, there are a lot of cars qualifying 10th without setting a lap time, and even when they do record a lap, they're not always pushing 100% as we know. So maybe if the qualifying and tyre usage rules were different, the spreads of the mid-70s would be in range.

 

[Mephistopheles]

Which years had the one lap shoot out? To my mind there should have been a big spread then, as MS said the first cars out were track sweepers and you went out first if you won the previous race and DC was bloody useless at it...

 

[Galahad]

It was one-lap shootout from 2003-05.

Here's the same data taken from 1970, to show up more of the recent fluctuations.

 

[sushifiesta]

MissTheApex Welcome to the forum! There's a lot of great stats work on here and it's great to see new ideas coming in!

Personally I think looking at qualifying in the way Galahad has done is the way to go. The race data is much more sensitive to factors like safety cars coming in to play in recent times for example. Also, in the cases where cars have been lapped you immediately have an error of up to one lap in the gap you give (no way around that unfortunately) - in other words you don't know whether a car that finished 1 lap behind was only just lapped or whether he was about to be lapped for a second time. You might also want to look at median gaps rather than means to reduce the effect of any crazy races.

Now, to blow my own trumpet for a while I have a couple of articles where I've looked at the performance of the cars through the seasons from 2008 to 2013. Obviously you're really interested in a longer time scale but nevertheless you might find them interesting. The most recent article is here.

The two graphs that are most relevant to this are these two I think:

Both graphs are for teams rather than drivers, with data shown for the average of the top 3 teams, the 4th-6th placed teams, the 7th-9th placed teams and the 10th-12th placed teams. The top graph shows the gap to pole in qualifying and the bottom the points gained in the races (with all years adapted to the current 25-18 system).

A couple of features that stand out for me are -
1) The introduction of the new teams in 2010 is very apparent by the huge drop off in pace for the 10th-12th placed teams in qualifying.

2) The front runners were gaining a bigger and bigger advantage on the rest of the field up until 2012 when restrictions were placed on double diffusers and exhaust blown diffusers etc.

 

[jez101]

I'm afraid that I have to agree with my esteemed colleagues that have already pointed out that drivers are increasingly "managing the gaps" when leading so your are unlikely to get any meaningful data from what could be a ton of work.

I've always wanted to look instead at field spread after eg. 10 laps rather than at the end of the race. Unfortunately, that data is even harder to come by! Don't let me put you off though

 

[Fenderman]

It's an interesting exercise and I admire the mathematicians and stat's fiends who are taking on the task of analysing the data. There are a host of issues with regard to comparing across eras let alone the timescale that one needs to consider. First and foremost is one of my favourite old chestnuts from economics. That is the law of diminishing returns. The 2013 season marked the end of a period of unparallelled reliability in F1. That was mainly due to the freeze on engine development but the strictness of the technical rules, with regard to all manner of components and materials, resulted in much closer parity between teams. As alluded to above with regard to the introduction of new minnow teams, the only real reason a large disparity between teams appeared was because of finance. Had the new minnows the funds to be truly competitive any large gaps in recent years would not have existed.

We can speculate ad infinitum about whether or not the performance differentials between drivers has changed but in reality the closing of the gaps is a natural consequence of the diminishing opportunities for radical innovations. It is no coincidence that huge margins of performance have either followed or preceded changes to the sport's technical regulations. On the one hand new restrictions are imposed to stamp out innovations that are seen as conferring excessive or unfair advantage (Traction/Launch control, EBD's, F-Ducts, etc.) or seen as dangerous (sliding skirts, moveable aero', etc). Meanwhile changes in the reg's often hand an advantage to someone else who finds a wrinkle - usually Adrian Newey at least for a while until the rest catch up again!

So for me I don't expect to learn much that is new. Yes, it is an interesting exercise for those who want to get bogged down in the work and for us lazier types who will happily wait for the results of someone else's hard work so hats off to the chaps and chapesses for taking on the job!

Here though is what I think we might find out:

• The fortunes of the lower half of the grids is irrelevant. The disparity in terms of funding has never allowed them to meaningfully close up on the top four or five teams.
• As more commercialisation and "big money" has come into the sport the best funded teams have closed the gaps to each other in the development race as the they have been able to react more quickly to throw money at solving the technical challenges.
• The difference individual technical improvements bring has reduced over time for two reasons: 1) Technical restrictions impede radical developments from being introduced even when they are thought up and 2) As the engineering of a particular form (such as the V8) gets nearer to the ultimate iteration the expenditure outstrips the reward in terms of performance gains (i.e. the law of diminishing returns).
• The closer a formula gets to a spec' series the closer will be the gaps between finishers.
• Regulations that limit competitors cars outright performance (i.e. rev' caps, power output restrictions, etc.) and impose minimum reliability requirements (Kilometers per Litre, engine/power unit/gearbox lifespans, etc.) and budget caps close the performance advantages and gaps in the field.
• and last but not least Health & Safety regulations and the associated improvements (or not depending on how one feels about them) that reduce DNF's through accident and injury - meaning more finishers and the top blokes who don't crash finishing closer together.

So, whilst I don't expect the statistical analysis to significantly change my perception of what has happened over the decades i am interested to see if any startling revelations do emerge to surprise me. Happy New Year stat's fans!

 

[MissTheApex]

Thanks for all the input and thoughts on this topic. I agree that these statictics would be more useful, if safety cars, number of cars at the start and reliability would be considered, but that seems to be a huge task. So i'll just stick with the race results and don't go any further. If you would consider the reliability and the number of cars, i assume the 1950 and 1960 season would be closer to the newer seasons.

@ jez: http://en.mclarenf-1.com/ has what you demand, you can compared lap by lap with this site for the more recent races.

Here are my graphs for the 1950, 1960, 1970, 1980, 1990, 2000, 2010 and 2013 season. Only classified drivers were considered. I plotted the average and median gap to the leader relative to the total race time and you can clearly see that in the older seasons the these two numbers show bigger differences to each other. The overall trend is similar to the qualifying diagram that has already been posted.

What do you think will happen next season? New regulations reduce the amount of downforce, that should bring the field together IMO. The engines on the other side could cause the opposite, if the power differences are bigger with the newly introduced turbos.

 

[FB]

Now here's an interesting add on - does any one have the stats for cars starting and finishing races? I can imagine that that will have been a steady decline over the years, perhaps with a spike at the start of the 1st turbo era.

 

[Brogan]

I have all of that in the old database, including retirement reasons and classifications such as mechanical, fire, engine, etc.

 

[F1Yorkshire]

What's caused the spike in the averages in the 1960's?

 

[ZakspeedYakspeed]

What's caused the spike in the averages in the 1960's?

Jim Clark and Jackie Stewart ....

 

[MissTheApex]

My graphs only show the gaps for single seasons, so the spike is in the 1960 season, not in the average of the average gap of the sixties. This is one reason.

And then in the 1960 season in ten races only six times a driver was placed on the 10th place. The spike in the average gap is caused by Ian Burgess who placed 10th in France with 14 laps down and by Fred Gamble (what a name )in Italy who was 9 laps down. This is why the median shows huge differences compared to the average number.

Regarding reliability from 1992 onwards, there is an interesting article here: http://www.f1fanatic.co.uk/2013/12/30/2013-f1-stats-season-context-end-peak-reliability/

 

[MissTheApex]

I have all of that in the old database, including retirement reasons and classifications such as mechanical, fire, engine, etc.

Is your database programmed in a way that you easily can produce an graph for the winning gaps and for all the seasons from 1950 till 2013 or would this be to much work?
My "database" is just an OpenOffice/Excel sheet where i manually put in the data from F1.com