F1 Stats: Reliability and some probabilities for this season

Schummy · 26 Mar 2002, 04:50 (Ref:244225)

Last season total race reliability (this is, if car finish the race ot not, whatever the reason) had an average of 61% (minimum 10 finishers, maximum 17 finishers). It implies that approximately one driver has a prob of 61% to finish.

This year we have:
AUS: 8/21 (38%)
MAL: 13/22 (59%)
-----------------
Total: 21/43 (49%)

We know in AUS we had a 10 car pile-up in first lap, and it made a very low race reliability (last year the minimum number of finishers was 10). Anyway, the group of 11 survivors had a reliability of 8/11 (73%) that is not low at all.

Being a bit playful with the numbers, we can deduce some curious calculations. Reliability less than perfect is what makes possible for lesser teams to get some points, and to even get podium positions. If we supose that overtaking is minimal and reliability was about the same for every team, we can derive probs for those events. I put them here just for entertainment and not claiming any particularly good forecasting accuracy:

Prob of podium with Ferrari-Williams-McLaren: 83%
Prob of podium position with another team in it: 17%
Prob of points position for Sauber: 63%
Prob of podium for Sauber: 10%
Prob of points position for Renault: 64%
Prob of podium for Renault: 6%
Prob of points position for Jordan: 44%
Prob of podium for Sauber: 3%
Prob of points in any of the next 15 GPs for Minardi: 2%

Thanks for the patience

DavyboyLT1 · 26 Mar 2002, 06:50 (Ref:244253)

Based on statistics I've compiled from the first 2 races this year, there is a 100% probability of a Williams and a Ferrari having an "incident" of some sort at the first corner on lap 1 in Brazil.

DNQ · 26 Mar 2002, 07:25 (Ref:244260)

Excellent, well done.

Glen · 26 Mar 2002, 09:46 (Ref:244290)

Excellent stuff - thanks.

One problem - why assume that overtaking is minimal? The likelihood of a poor qualifyier finishing much higher than his start position is minimal, but overtaking is not neary as limited as it first seems. TV coverage has been rubbish, and we have missed alot of moves, but overtaking is there.

Renault have got to be the hot tip for Brazil, in my opinion. Going well, Michelin tyres, good drivers...

Champ69 · 26 Mar 2002, 11:44 (Ref:244319)

Good work. I have just completed a Statistics exam today and I looks as if I may get you to sit the next one for me.

Schummy · 26 Mar 2002, 12:12 (Ref:244326)

Champ69 LOL.

Glen, I have assumed minimal overtaking just because it is easier to calculate

and it's not totally unaccurate. I agree with your point of view about Renault, I hope it isn't wishful thinking of us.

Thanks.

PS: My guess about the first corner customary W/F incident is MS with Ralf or Rubens with JPM (those are the remaining combinations)

Mr V · 26 Mar 2002, 20:04 (Ref:244639)

Quote:

Originally posted by DavyboyLT1
Based on statistics there is a 100% probability of a Williams and a Ferrari having an "incident" of some sort at the first corner on lap 1 in Brazil.

has anyone worked out the proberbility of jpm being blamed and getting a penalty?

AndyF · 26 Mar 2002, 20:07 (Ref:244641)

It's lies, damned lies and statistics - Minardi will get more points this season!!

Mark F1 · 27 Mar 2002, 00:39 (Ref:244842)

Nice stuff Schummy

What I'd like to know is the probability that one of the Arrows remains stranded on the grid at the start of a race

dman007 · 27 Mar 2002, 01:01 (Ref:244848)

What's that saying, figures lie, and liars figure?

Interesting stuff Schummy. Are you going to give us up to date stats following each race? Undoubtedly, you will be able to come up with some new categories as each race unfolds.

Schummy · 27 Mar 2002, 02:35 (Ref:244873)

Sure! I have a lot of statistical sillynesses yet in my pocket

If next races are boring I'll have to make calculations to camouflage it

I hope when season advances Minardi will get more (statistical) chances to take points. I always thought Giancarlo Minardi was a nice person, and I find Stoddart as a worthy follower (c:

Bibendum · 28 Mar 2002, 16:04 (Ref:245962)

Playful with numbers, yes. Forgive the pedantry, but this ain't statistics -- it's reading chicken entrails.

Or, to put it more technically, the extrapolations are based on an entirely inadequate data set!

Schummy · 28 Mar 2002, 17:12 (Ref:245988)

Yes, I know

This is not Levy-Pareto stable distributions or General Autorregresive Conditional Heterocedasticity time series model

But, what the hell, we are having a bit of fun here

**Adam43** · 28 Mar 2002, 17:15 (Ref:245991)

nice!

Should get a bit better with more races.

Bibendum · 28 Mar 2002, 19:10 (Ref:246074)

Out-syllabled, by god! I'm humbled; gag me with a two-tailed, bi-modal distribution! I limp from the statistical field trailing a few irrational numbers behind me.

(And I cannot resist the observation that, based on a statistical analysis of my single prior contribution to this thread, I'm a humourless dolt

.So I'll try to resist giving further data points to support that inference...)

Schummy · 28 Mar 2002, 19:53 (Ref:246115)

So funny, Bibendum! And that use of the "smiley-in-the-bag" is a classic!

Now, let's see what the weekend brings to us... maybe a lot more (admittedly silly) numbers. The statician in me would hate it but I'd like to see a Minardi in the podium =:-o Oh, well! It would be as rare as Ferrari and Williams going thru first corner untangled! (zero, null, vanishing probability?)

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Reliability after Canada	Schummy	Formula One	5	19 Jun 2004 02:28
Reliability after Imola	Schummy	Formula One	10	30 Apr 2004 15:37
F1 Reliability	Wrex	Formula One	23	11 Aug 2003 07:46
Another interesting season stats!	Gt_R	Formula One	11	28 Nov 2000 12:58