"Average" positions (4 GPs)

[Schummy]

---(Somewhat boring details section)-------
For some reasons I think average final positions (in GPs) are a good indication (not the only or necessarily the best) about the relative form of driver/car; just an item more to get information.

But simple average is no convenient because "anomalous" data can destroy the intended effect of "averaging" as a "predictor" for (imaginary) future races.

To overcome this difficulty, sometimes it's used some "modified average". I will use here an average dropping "anomalous" classifications. With only 4 races, this average is similar to the well known "median" (the "middle" number in the data).
---(Ok, now the real data)------

   | Positions |   Stats
    B  M  A  S  Me   MAD  Av  Rank
FA  1  2  1  2  1.5  0.5  1.5  1
MS  2  6  -  1  4    2.5  4    2/3
KR  3  -  2  5  4    1.5  4    3/3
GF  -  1  5  8  6.5  3.5  6.5  6
JM  5  4  -  3  4.5  1    4.5  4
JB  4  3 10  7  5.5  2    6    5
FM  9  5  -  4  7    2.5  7    7
JV  -  7  6 12  9.5  3    9.5  8
RS 14  8  3  9  8.5  3    8.5  9

The first four columns are the positions acheived in the GPs so far.
Me=Median
MAD=A measure of the irregularity of results.
Av=The corrected average
Rank=Ranking by corrected averages

Average positions expressed in a semigraphic way

1 2 3 4 5 6 7 8 9
<------------------
 FA   MS  JB   RS
      KR   GF    JV
       JM   FM
<------------------

Fisichella looks worse than in FIA classification, and he is the most irregular of that set of drivers. Appart from his (deserved) win in MAL he has been disappointing.

JPM is doing a regular, consistent work, its only fault was that semiaccident in the infamous last corner in Melbourne. Anyway he still is devoid of a top race/result.

[Dani Filth]

i love rainy days

[Marbot]

Originally Posted by Schummy

MAD=A measure of the irregularity of results.

Not an irregularity of the driver then.

[marzF1rocks]

seems to pretty all over the place..

[Inigo Montoya]

Nice stuff again Schummy ! As I always say, keep it coming!

Just out of curiosity, how is this central tendency measured - i.e. how do you decide what is "anamolous"?

[Schummy]

Inigo:

Thanks

"Anomalous" is, in this context, defined in terms of median and MAD. It is any data that deviates from median more than four times MAD

x < Me - 4*MAD, or x > Me + 4*MAD

Usually they are retirements (or accidents).

BTW, MAD is calculated as follows. You have to calculate absolute values from each datum respect median, and then you calculate the median of those absolute values.

Median and MAD have the advantage (over average and standard deviation) of being almost not disturbed by "anomalous" (or even infinite or undetermined) values. So they are a good base to help to identify what are "anomalous" values.

When you suspect they are not "anomalous" data, it is better to use the usual averaging procedures.

I hope it helps

[Bellof]

and the conclusion :
Me - MAD 2...

[Dan Fielden]

It all sounds very complicated. Cant you just copy and paste the championship table of f1.com or something

[VilleneuveTracy]

Look at JV ranked 8th despite having medicore machinery.

[Knowlesy]

Yes, that BMW Sauber is a real dog isn't it.

[VilleneuveTracy]

Originally Posted by knowlesy

Yes, that BMW Sauber is a real dog isn't it.

Well, "Quick Nick" struggled to even beat a Toro Rosso driver en route to qualifying 15th in the second session.

That said, I see your valid point. However, BMW is not in the same class as the top four teams at the moment. BMW is, at best, on par with Williams and Toyota, at worst slightly behind these two teams. JV is doing better than the car warrants. This was especially the case today. When Heidfeld, who won the pole at this track last year, could only muster 15th in qualifying (beating only the Aguris, Toro Rossos, Midlands, no big deal. The only decent car he beat was one of the Red Bulls), that suggests that BMW is significantly behind the top teams. I still believe NH is a very good driver but the 1997 WDC is flattering the BMW, in my (unbiased ) view.