Yet Another Combined Championship (American Sportscar)

[Danske]

YACC: Points scaled by number of entries in the class; last place is worth one more place than not participating (which garners zero points); theoretical maximum for a win is 100 points, but would require 200 entries to be rounded up to that; average point value awarded is 50; a combined ALMS P* class also included, although it makes little difference in the rankings; all non-excluded finishes score points regardless of distance completed; excluded entries do not count toward the number of entries in the class; totals averaged by number of races each series has run; top five cars in each of the seven classes included in the table:

Pts #       Cls
94  01  Ganassi     DP
88  07  Banner      GT
86  7   Penske      P2
82  91  BS/RM       DP
82  7   Penske      P*
80  71  Tafel       GT2
78  2   Audi        P*
75  9   Highcroft   P2
74  2   Audi        P1
73  69  SpeedSource GT
73  67  TRG         GT
73  9   Highcroft   P*
73  60  MSR         DP
71  46  FLM         GT2
70  45  FLM         GT2
70  57  Stevenson   GT
69  72  Autohaus    GT
68  99  BS/RM       DP
67  61  Risi        GT2
64  1   Audi        P1
64  3   Corvette    GT1
62  1   Audi        P*
61  44  FLM         GT2
58  76  Krohn       DP
55  16  Dyson       P*
53  4   Corvette    GT1
50  16  Dyson       P2
49  6   Penske      P2
46  26  AGR         P2
32  37  Intersport  P1
19  12  Autocon     P1

By design, high finishes in classes with more competitors award more points than high finishes in classes with few competitors. Conversely, finishing last in a class with more entries awards fewer points than last in a class with fewer entries. Fifty points is the average number of points awarded in each class, and also the number of points awarded for a median finish. A win in a class with N cars is still worth more than a 2nd in a class with N+1 cars.

Anyway, with different classes varying so much in number of entries I thought it would be interesting to see how a combined championship would work-out correcting for the number of entries. One other system under consideration was awarding points proportional to the ratio of each finisher's average speed to that of the class-winner's, but for the moment that seems like too much of a pain to calculate and the teams are actually racing for position, not a higher average speed if no positions will be gained by that.

[Dani Filth]

you're actually trying find a system for GA teams winning

[Speed-King]

Originally Posted by Dani Filth

you're actually trying find a system for GA teams winning

Hey, he could've included Koni Challenge!!

[AstonGeoff]

Well if your formula is based purely on finishing position you could add just about any championship into that table! From Formula Ford to F1, from Caterhams to ALMS.

Its really rather pointless isn't it???

For it to be taken seriously you'd need to base it on pseudo-race results, i.e. use the fastest times from tracks where both series run to work out what the results might have been if they had all been on the track at the same time. I don't really follow GA, but I presume the ALMS cars are faster so they would win!

Is this just a case of "Lies, Damm Lies and Statistics".....

[dxk1]

Originally Posted by AstonGeoff

Is this just a case of "Lies, Damm Lies and Statistics".....

Yes!

[Danske]

Originally Posted by Dani Filth

you're actually trying find a system for GA teams winning

The difference between Ganassi winning the Rolex 24 and Penske winning the Sebring 12 Hours in the combined P* class is only three points out of one hundred: 96 vs. 93. The number of entries for each was 25 and 13. Second place in the Rolex 24 was 92 points under the system. Conversely, last in the Rolex DP class was worth only 4 points, while last at Sebring in the combined P* was worth 11. The i-Moto Acura would get 98 points for winning the 54-entry Koni opener. Yes, points per position are double in a class with twice as many entries, but 1st in a 10-entry class is still worth slightly more than 2nd in a 20-entry class.

Guess I should have struggled with a race-by-race table, but it was annoying enough getting that one presentable.

And yes, it is an adjusted percentile result. Finish mid-pack regardless of the number of entries and the result is 50 points. The motivation was to rank teams relative to their competition, regardless of the number of entries. For example, if a virtual GM team combined the points from Banner in G-A and Corvette Racing in GT1 (or say Audi got a DP engine approved and combined results from DP and P1) it doesn't seem fair for them to rack up so many points in GT1, being nearly unable to finish less than fourth (or even second), relative to a team (possibly through proxy) competing in G-A GT and GT2.

Perhaps teams racing against little opposition are penalized too much under the system as it currently is, but I think they are certainly rewarded too much under a "top-down" points system as is.

[Danske]

Okay, for a little more transparancy, here's the ALMS-only results under the system (omitting the P* class; may also contain errors, no warranties expressed or implied, use at your own risk):

Pts Cls #               Fin             Pts             # entries           
86  P2  7   Penske      1   1   2       89  90  78      8   9   8   
80  GT2 71  Tafel       7   1   1       56  92  93      15  12  13  
75  P2  9   Highcroft   4   2   1       56  80  89      8   9   8   
74  P1  2   Audi        2   1   1       67  75  80      5   3   4   
71  GT2 46  FLM         8   2   3       50  85  79      15  12  13  
70  GT2 45  FLM         1   9   2       94  31  86      15  12  13  
67  GT2 61  Risi        3   5   6       81  62  57      15  12  13  
64  P1  1   Audi        1   2   2       83  50  60      5   3   4   
64  GT1 3   Corvette    1   2   1       75  50  67      3   3   2   
61  GT2 44  FLM         2   7   7       88  46  50      15  12  13  
53  GT1 4   Corvette    2   1   2       50  75  33      3   3   2   
50  P2  16  Dyson       3   4   7       67  60  22      8   9   8   
49  P2  6   Penske      8   3   3       11  70  67      8   9   8   
46  P2  26  AGR         6   5   4       33  50  56      8   9   8   
32  P1  37  Intersport  3   3   4       50  25  20      5   3   4   
19  P1  12  Autocon     5       3       17  0   40      5   3   4

Perfect win/DFL record in class:
P1 = 79 / 21
P2 = 89 / 11
GT1 = 72 / 28
GT2 = 93 / 7