Bootstrapping F1 (2010)

Schummy · 20 May 2010, 03:49 (Ref:2694473)

Like last season, I have applied a "bootstrapping" method to F1 results (so far). The graph shows probability of win the WDC title after each GP run until now.

The calculations are done supposing in future races cars/drivers will perform equally well, i.e. without the current advantage for Red Bull and so.

If we consider current differences in cars/drivers and carry it for future races (i.e. Red Bull will go on enjoying its advantage) probabilities become more extreme:

WEB 35%
VET 32%
ALO 21%
BUT 10%

The remainder drivers would get 1% or less.

**Born Racer** · 20 May 2010, 17:15 (Ref:2694880)

Hi Schummy. In the calculations based on Red Bull having their advantage reduced, how much have you reduced it by and how have you decided where the teams' relative performance level will sit?

Schummy · 21 May 2010, 01:26 (Ref:2695153)

Quote:

Originally Posted by Born Racer

Hi Schummy. In the calculations based on Red Bull having their advantage reduced, how much have you reduced it by and how have you decided where the teams' relative performance level will sit?

It has been supposed every team has the same level of performance in the future, so RB's advantage has been entirely removed as if in the next race Ferrari, Mac, etc, will perform in equal terms.

In fact that "future configuration" doesn't correspond with any past GP, because in each one of them there was differences in teams level. It is a idealized framework reflecting we cannot know who is going to end the season on top form. Moreover, it avoids to introduce personal opinions (about future teams development) in the calculation.

phoenix · 21 May 2010, 09:31 (Ref:2695270)

Hi Schummy

How does Button's DNF in Monaco affect his future predicted fortunes? I ask because it appears from your graph to have had a dramatic and irreversible effect, though I see that Vettel's unfortunate start to his season has not had the same effect...

Schummy · 21 May 2010, 12:57 (Ref:2695382)

Quote:

Originally Posted by phoenix

Hi Schummy

How does Button's DNF in Monaco affect his future predicted fortunes? I ask because it appears from your graph to have had a dramatic and irreversible effect, though I see that Vettel's unfortunate start to his season has not had the same effect...

Vettel's event was less severe because there were more races to amend it. As GPs pass by, errors will be more relevant. But in Button's particular case, the main problem is he is now below three competitors, and he will have to "overtake" all of them to win the title; it has a dramatic effect in his PoT (Probability of Title).

Anyway, he is now at 13%, it is not so bad taking in consideration the level of competition this season.

Marbot · 21 May 2010, 18:19 (Ref:2695602)

Quote:

Originally Posted by phoenix

How does Button's DNF in Monaco affect his future predicted fortunes? I ask because it appears from your graph to have had a dramatic and irreversible effect, though I see that Vettel's unfortunate start to his season has not had the same effect...

I'm predicting that the wavy line will stop going into freefall for the next race....hopefully.

Schummy · 21 May 2010, 23:39 (Ref:2695733)

(... wait a moment, I'm redoing the 2007 graph... )

Schummy · 22 May 2010, 01:46 (Ref:2695769)

The celebre 2007 season.

For most of the season, McLaren had a collective 90% of PoT (Prob of Title), and ultimately that pesky Kimi won

. With barely a 7%-8% in the last race, Raikkonen raised one of the most surprising titles.

Hammy had 99% two races to go and yet... the rest is history.

Schummy · 17 Jun 2010, 15:48 (Ref:2713778)

Probability of Title after Canada:

Current percentages are:

HAM 28%
BUT 23%
WEB 20%
ALO 11%
VET 9%
ROS 3%
KUB 3%
MAS 2%
SCH <1%

Team wise (title for a driver in the team, not team title):

MAC 51%
RBR 29%
FER 13%
MER 3%
REN 3%
AnOther <1%

The maximum PoT reached by each driver along the season. It is a sort of hierarchy in merits for the season so far:

HAM 28%
WEB 25%
BUT 23%
ALO, VET 20%
ROS, MAS 13%
--------------
KUB 7%
SCH 5%

I have separated them in two groups because usually reaching a 10% of PoT in some moment of the season is approx the mark for being a relevant driver in that season (in terms of driver-car pair, not purely in driver skill). So, Kubica and Schumacher have not reached yet the level of "season relevance" required. In fact Michael is moving away of it.

As season pass, less GPs remind and it becomes harder and harder to overcome a points deficit. That's the reason PoT will tend to concentrate in less drivers (the true title contenders) and the others will lose the thread eventually. We are beginning to see as Rosberg, Massa and Kubica seem to slowly go down in their numerical expectations.

Alonso and Vettel naturally still have a good shot at the title, but these calculations show that it would be dangerous to have another zero GP in the few next races.

dsg · 18 Jun 2010, 01:57 (Ref:2714091)

Love your work Schummy.

Schummy · 19 Jun 2010, 12:38 (Ref:2714692)

Thanks!

In one moment or two, I'll post the evo of some past season as 2009 or 2008. So one can have a perspective about how PoT moves along a season. Moreover it can be fun to look on.

JeremySmith · 19 Jun 2010, 14:54 (Ref:2714727)

Great work Schummy as always!

pannenmann · 20 Jun 2010, 22:43 (Ref:2715372)

Thanks, all that statistical analysis is really interesting. Might be something official F1 should incorporate. Does your job involve such analysis or do you only do this in your spare time?

Schummy · 21 Jun 2010, 01:36 (Ref:2715439)

Quote:

Originally Posted by pannenmann

Thanks, all that statistical analysis is really interesting. Might be something official F1 should incorporate. Does your job involve such analysis or do you only do this in your spare time?

I'm silly enough to do it in my spare time

. But my job is related to math & stat applications, so I think I'm probably sick

Schummy · 12 Jul 2010, 14:27 (Ref:2725513)

Evolution of Probability of Title after British GP.

Current PoTs:

Code:

Drivers:
HAM  46%
BUT  23%
WEB  17%
VET  11%
ALO   2%
ROS   1%

Teams (WDC, not WTC):
MCL  69%
RBR  28%
FER   2%
MER   1%

Farewell Alonso. His title hopes are virtually gone. It's not just the point gap but the fact he has to overcome four drivers.

Title chase is reduced to Mac and Merc drivers, with advantage for the Dennisians. Hamilton looks good for WDC, but it's still too soon to jump to too much conclusions. He lost the 2007 title while having a much bigger PoT.

shiny side up! · 12 Jul 2010, 23:44 (Ref:2725817)

Love the analysis, thanks so much Schummy!!!!

Teretonga · 13 Jul 2010, 00:01 (Ref:2725824)

According to Autosport Alonso has been adamant he is going to win the title....According to this he has a 2% chance.....

Does he know something none of us do?

(Is Mr Todt going to declare RBR and Macca's DQed from the championship for......?)

Or is he playing the game for the Italian press....

Schummy · 13 Jul 2010, 01:03 (Ref:2725841)

Quote:

Originally Posted by shiny side up!

Love the analysis, thanks so much Schummy!!!!

Again thanks, you are really kind!

Schummy · 13 Jul 2010, 01:18 (Ref:2725847)

Quote:

Originally Posted by Teretonga

According to Autosport Alonso has been adamant he is going to win the title....According to this he has a 2% chance.....

Does he know something none of us do?

(Is Mr Todt going to declare RBR and Macca's DQed from the championship for......?)

Or is he playing the game for the Italian press....

I have heard Alonso speaking about the point deficit. I think he is not totally aware that confronting a given gap with 3 drivers in between is a lot harder than the same gap against only one driver. When you are 5th you depend on the other 4 drivers faulting here and there. Hard.

The easiest/only way for Alonso would be Ferrari becoming a very dominant car. So, the gauntlet is for Ferrari's tech department in a greater degree than for Alonso.

**Born Racer** · 13 Jul 2010, 10:43 (Ref:2726008)

Quote:

Originally Posted by Schummy

Evolution of Probability of Title after British GP.
Hamilton looks good for WDC, but it's still too soon to jump to too much conclusions. He lost the 2007 title while having a much bigger PoT.

True, but can your (enthralling and fascinating, so thanks for them) statistics take account of a driver's 'potential', i.e. that in 2010, Hamilton is a more complete driver and would be less likely to make mistakes than in China 2007?

m1fcf · 13 Jul 2010, 10:50 (Ref:2726013)

Quote:

Originally Posted by Schummy

Title chase is reduced to Mac and Merc drivers, with advantage for the Dennisians.

I think you meant Mac and RedBull drivers, but the same still applies. As far as WDC, and maybe WCC, it is down to a two-horse race.

Thanks for the info.

Schummy · 13 Jul 2010, 15:35 (Ref:2726154)

Quote:

Originally Posted by Born Racer

True, but can your (enthralling and fascinating, so thanks for them) statistics take account of a driver's 'potential', i.e. that in 2010, Hamilton is a more complete driver and would be less likely to make mistakes than in China 2007?

No, it doesn't consider those factors. One of the difficulties to try it is to avoid overdoing it. For example, Alonso is now more experienced and he is making more "rookie" errors.

One way to, indirectly, doing it in an objective way is to "bootstrap" taking into account current performances (whatever they are), i.e. forecasting future races using past scores (in this season). I have try it and globally speaking I'm not very convinced yet.

It is an interesting point

Schummy · 13 Jul 2010, 15:36 (Ref:2726155)

Quote:

Originally Posted by m1fcf

I think you meant Mac and RedBull drivers, but the same still applies. As far as WDC, and maybe WCC, it is down to a two-horse race.

Thanks for the info.

Lol, yes. I meant RB. Probably a freudian lapsus

**Born Racer** · 2 Aug 2010, 11:42 (Ref:2738115)

Schummy, what are Alonso's percentage chances of winning the world championship now? To ask someone such a question feels a bit like asking an octopus to predict a football match result, although this is grounded more in mathematical reasoning and you are more intelligent than an octopus.

It seems like Alonso now has a major chance, correct?

How can this bootstrapping allow for changes in relative car performances, such as we are seeing (e.g. the McLaren struggling a bit to a lot for the last 3 Grands Prix)? Can it allow for it at all?

Schummy · 2 Aug 2010, 15:43 (Ref:2738258)

Quote:

Originally Posted by Born Racer

Schummy, what are Alonso's percentage chances of winning the world championship now? To ask someone such a question feels a bit like asking an octopus to predict a football match result, although this is grounded more in mathematical reasoning and you are more intelligent than an octopus.

Definitively I'm a fan of Paul Octopus! I don't think I can equal him but I'll try to do the best a human can do. He is psychic and I only use maths, so I'm in a severe disadvantage, however I'll put the bootstrapping analysis soon, as you request!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
F1 2010 (Codemasters)	ECW Dan Selby	Virtual Racers	630	24 Jun 2011 10:38
[Official] 2010 F1 Predictions and Fantasy F1 Competitions	Ralf's Girl	Formula One	4	8 Mar 2010 17:10
Bootstrapping F1 (stats)	Schummy	Formula One	39	2 Jun 2007 13:49