[Driver] Team mates comparison 2018

[Schummy]

After years of indolence(?) I am going again to post the (ex)customary Table of Team Mates Comparison (name no copyrighted).

I have chosen these teams in particular because I think they are the most closely fought or at least the most interesting to follow. I think Mercedes and Ferrari pairs are not very difficult to guess (and the poll in this forum shows it very clearly, although it also shows how few Finnish people we have here ;-) ).

         AU BA CH AZ SP MO CA FR AU GB GE HU   score    conf need grid race   
RIC-VER  01 10 01 10 00 11 00 00 00 01 00 01 =  8-16 VER 85%  2x0  3-9  5-7 
HUL-SAI  11 11 11 10 00 01 11 00 00 11 11 00 = 14-10 HUL 46%  5x1  7-5  7-5
GRO-MAG  01 00 10 00 00 10 01 10 11 00 01 10 =  9-15 MAG 69%  3x0  5-7  4-8
PER-OCO  11 00 10 01 01 00 00 01 00 00 11 00 =  8-16 OCO 85%  2x0  3-9  5-7

In each GP, first number means "grid" (actually qualifying time) and second is race classification. A 1 means the first driver was better and 0, obviously, means the second driver was better. For example, 01 for RIC-VET in AUS GP means Verstappen was better than Ricardo in qualifying but Ricardo had a better classification in race.

score = total account of 0s and 1s for both teammates (Ricardo was 8 times better and Verstappen was 16 times better).
conf = level of "confidence" in that the better performing driver is really better and not just by random events. When it is more than 95% it could begin to be considered "proven".
need = An "obscure" stat :-) . it means the amount of races the better performing driver has to be better to "prove" he is really better.
grid = the scoring just counting qualifying
race = idem just counting races

So, none of the drivers has proven to be better than his team mate beyond a reasonable amount of randomness. In other words, those results could be made just by random events, similarly to how one can get more heads or tails by chance in a fair coin.

VER and OCO, however, are with a clear advantage and nearing a strong enough performance to prove their superiority. They would need two consecutive dominant races to enter in the "proven" zone.

MAG is outperforming GRO but a bit less convincingly, so "probably" he is actually better than the French but is not yet a proof.

HUL's loss in Hungary has put his advantage in a weak position. He has still a long road ahead to show he is really better than SAI beyond just luck.


Just to put a bit of perspective, a score 18-6 would mean a level of confidence greater than 95% and a score 19-5 implies a confidence greater than 99%. Those kind of scores could be considered as proofs of real dominance beyond random (bad) luck.

[crmalcolm]

Hey - good to see something that resembles a scientific approach to determine who is better. Appreciate the work that you have put into this - and look forward to seeing if 'proof' develops over the next few races.

[Akrapovic]

I know a lot of people hate these comparisons, but I like them. I find them interesting. I don't necessarily put much faith in them, but I do find them interesting to read, so thanks for this,

[dsg]

Originally Posted by Akrapovic

I know a lot of people hate these comparisons, but I like them. I find them interesting. I don't necessarily put much faith in them, but I do find them interesting to read, so thanks for this,

I like them as well however the binary choice for each event is why I don't have a great deal of faith in any conclusions based solely on this type of analysis. They can certainly be used to support an argument however the devil is in the details and any number of factors can skew the statistical information.

Thanks Schummy, they are interesting to look at.