Parameter's for judging car advantage.

Post here all non technical related topics about Formula One. This includes race results, discussions, testing analysis etc. TV coverage and other personal questions should be in Off topic chat.
User avatar
raymondu999
54
Joined: 04 Feb 2010, 07:31

Re: Parameter's for judging car advantage.

Post

beelsebob wrote:
raymondu999 wrote:IMO reliability should not be a "multiplier" in this thing. When your car gives you a DNF - whatever the cause may be... faulty clutch... blown engine etc... then you have absolutely no car advantage. It doesn't matter if you completed 1 lap out of a possible 66, or 65 laps. Fact of the matter is, even if you had a 2s/lap advantage on the field, on that day your car precisely has as much points-scoring capability as a 2012 HRT in 2013.
I don't understand how your evidence backs up your argument. Quite the opposite in fact. If your car is unreliable, it doesn't matter how fast it is, it's still ---. That means that in a fast, unreliable car, you have not got a car advantage, but a car disadvantage. So yes, surely reliability should be taken into account.
You misunderstood my argument. I'm not saying that reliability should be left out. Rather, I'm saying reliability should not be a "multiplier" - but rather a "subtractor." A DNF immediately "zeroes" any other car stat, as opposed to reducing it by a certain factor. If we had a scale of 0-1, and a dominant, all-conquering reliable car was 1, to an HRT's 0, let's say. Now we have a car that is a 1 for 65/66 laps. But it blows its engine on the last lap. That car, in my eyes, is not a 0.99, or a 0.9 or whatever number. In my book that's a 0 for that car.
失败者找理由,成功者找方法

beelsebob
85
Joined: 23 Mar 2011, 15:49
Location: Cupertino, California

Re: Parameter's for judging car advantage.

Post

raymondu999 wrote:
beelsebob wrote:
raymondu999 wrote:IMO reliability should not be a "multiplier" in this thing. When your car gives you a DNF - whatever the cause may be... faulty clutch... blown engine etc... then you have absolutely no car advantage. It doesn't matter if you completed 1 lap out of a possible 66, or 65 laps. Fact of the matter is, even if you had a 2s/lap advantage on the field, on that day your car precisely has as much points-scoring capability as a 2012 HRT in 2013.
I don't understand how your evidence backs up your argument. Quite the opposite in fact. If your car is unreliable, it doesn't matter how fast it is, it's still ---. That means that in a fast, unreliable car, you have not got a car advantage, but a car disadvantage. So yes, surely reliability should be taken into account.
You misunderstood my argument. I'm not saying that reliability should be left out. Rather, I'm saying reliability should not be a "multiplier" - but rather a "subtractor." A DNF immediately "zeroes" any other car stat, as opposed to reducing it by a certain factor. If we had a scale of 0-1, and a dominant, all-conquering reliable car was 1, to an HRT's 0, let's say. Now we have a car that is a 1 for 65/66 laps. But it blows its engine on the last lap. That car, in my eyes, is not a 0.99, or a 0.9 or whatever number. In my book that's a 0 for that car.
Right, but now lets say that we have a dominant car that has a 5% chance of blowing up over a race distance – that means that (on average) you'll win 95% of races, so a 0.95x multiplier makes absolute sense. Similarly, your incredibly unreliable car that always blows up would gain a 0.0x multiplier because of it's reliability track record.

User avatar
FoxHound
55
Joined: 23 Aug 2012, 16:50

Re: Parameter's for judging car advantage.

Post

Generally speaking, a car that does more laps over a season can be considered more reliable than one that does less.
I do get that a car that fails on lap one compared to that that fails on lap 65 of 66 will distort the figures.
But, bear in mind also this will not really be uniform.

Hence why Raymond, if you have a formula to circumvent the multiplier I would be all ears.
Put something forward. 8)
JET set

User avatar
raymondu999
54
Joined: 04 Feb 2010, 07:31

Re: Parameter's for judging car advantage.

Post

Are we doing this in a "prospective" fashion with probabilities into the future, or are we doing it "retrospective" and looking at past statistics?
失败者找理由,成功者找方法

tim|away
15
Joined: 03 Jul 2013, 17:46

Re: Parameter's for judging car advantage.

Post

beelsebob wrote:
raymondu999 wrote: You misunderstood my argument. I'm not saying that reliability should be left out. Rather, I'm saying reliability should not be a "multiplier" - but rather a "subtractor." A DNF immediately "zeroes" any other car stat, as opposed to reducing it by a certain factor. If we had a scale of 0-1, and a dominant, all-conquering reliable car was 1, to an HRT's 0, let's say. Now we have a car that is a 1 for 65/66 laps. But it blows its engine on the last lap. That car, in my eyes, is not a 0.99, or a 0.9 or whatever number. In my book that's a 0 for that car.
Right, but now lets say that we have a dominant car that has a 5% chance of blowing up over a race distance – that means that (on average) you'll win 95% of races, so a 0.95x multiplier makes absolute sense. Similarly, your incredibly unreliable car that always blows up would gain a 0.0x multiplier because of it's reliability track record.
I think raymondu999 suggests that reliability based on completed laps or km can be rather inaccurate and paint the wrong picture. Let's consider the following theoretical example in a season with 20 races where each race has 60 laps:

Car A:
- breaks down in 5 out of 20 races with only 3 laps to go each time.
- laps completed 1185 of 1200 -> 98.75%

Car B:
- breaks down once during the warm-up lap of one race and finishes the remaining 19 races.
- laps completed 1140 of 1200 -> 95%

Having said that, basing it on mechanical failures causing a DNF might be more advisable:

- Car A has 5 mechanical failures and finishes 15 out of 20 races -> 75%
- Car B has 1 mechanical failure and finishes 19 out of 20 races -> 95%

The fun part comes if one were to suggest that Car A has 5 failures per season and Car B has 1 failure per season. Surely Car A is 5 times as likely as Car B to have a mechanical DNF, right? WRONG!

If we take Car B as a baseline, Car A would have 4 more failures over 19 races - in other words Car A would be 21% more likely to experience a failure each race. If Car A manages to score 22%+ more points on average per race than Car B, it can statistically be considered the better car, because the gain outweighs the loss. It's absolutely fine treating reliability as a multiplier here by the way.

User avatar
WhiteBlue
92
Joined: 14 Apr 2008, 20:58
Location: WhiteBlue Country
Contact:

Re: Parameter's for judging car advantage.

Post

If the car advantage is viewed by statistical parameters it is not very interesting for the engineers who participate in this forum. It is just an exercise of book keeping to figure it all out. Even fans without any technical understanding can do that. I would find it much more interesting to identify technical parameters such as superior power, superior aerodynamic efficiency better use of the tyres and superior reliability.

I admit those things are more difficult to isolate but it would also be more rewarding to understand how a team has managed to get in front. It also has a big impact on the politicking which helps you to understand what is going on.

To give an example Ferrari used their Bridgestone connection in the early noughties to gain superiority by doing a huge amount of testing. Tyres were a big part of the Schumacher/Ferrari dominance between 2000 and 2004. But the pattern was picked up by Renault who got behind Michelin to create them a bespoke tyre that worked well on their cars and they did beat the Ferrari connection twice in 2005 and 2006. It pissed off Ferrari big time and they decided to lobby for a single tyre supplier. Because the other teams loved the idea of eliminating the cost of tyre tests the idea immediately worked and Michelin went away. Renault lost their competitive advantage and never recovered from it.

The next scheme for Ferrari was the new engine formula which they thought would shift the things a bit away from the aerodynamic contest that is largely at work since 2007. It took very long to introduce the new formula and by the time we will see it the aero masters Red Bull have lobbied so many restrictions that the 2014-2019 formula will probably not be dominated by engine power but by aerodynamic efficiency again. If it were different Red Bull would have bought an engine company some time ago and joined the race for the most powerful engine. They found a way to avoid that.

These two examples show how analysing the technical source of competitive advantages helps you understand the politics in F1.
Formula One's fundamental ethos is about success coming to those with the most ingenious engineering and best .............................. organization, not to those with the biggest budget. (Dave Richards)

User avatar
FoxHound
55
Joined: 23 Aug 2012, 16:50

Re: Parameter's for judging car advantage.

Post

WhiteBlue wrote:If the car advantage is viewed by statistical parameters it is not very interesting for the engineers who participate in this forum. It is just an exercise of book keeping to figure it all out.
I'm not sure you understand the premise here WB, not to mention that I actually disagree with your assessment.
Your post also seems to assume the position that you speak on behalf of "engineers who participate in this forum".

Further, "technical source" of advantage or not, there is a thing called car advantage. A car is the sum of all its part's, something you have glaringly missed.
Now if I may, the thread is Parameter's for judging car performance. This thread is here to get a guide on how to best judge a car's advantage over a season. It can be applied retrospectively or prospectively.
In my mind, this is a useful tool for any engineer and can also settle tedious circular arguments that blight alot of threads.
JET set

Blanchimont
214
Joined: 09 Nov 2012, 23:47

Re: Parameter's for judging car advantage.

Post

WhiteBlue wrote:I would find it much more interesting to identify technical parameters such as superior power, superior aerodynamic efficiency better use of the tyres and superior reliability.
Indirectly this can be done by analysing the speed of the car round the tracks throughout the whole season. A chart of maximum corner acceleration for different corner radii and straightline acceleration should reveal pretty much of what you're asking for. Tyre degradation can be defined when you analyse how much the lap times increase/decrease for a given stint.

BTW: Is it really necessary to downvote WhiteBlue for his post above? Come on guys
Dear FIA, if you read this, please pm me for a redesign of the Technical Regulations to avoid finger nose shapes for 2016! :-)