Here's what we've been playing with for our club. It is a dynamic system based entirely on driver times -- the cars aren't classed initially. Two people in the same car could end up in different classes based on skill.
We are proposiing five classes -- "STIG", "WILD", "SCORCHIN", "SPICY", and "MILD".
FTD is always scored at 100 points. FTD score contributes to STIG class score.
One "outlier score" can be used if the differential between FTD and the second fastest time is greater than 1.4 seconds, and the differential between the second and third place score is greater than .400 seconds.
Outlier score is added to the WILD class, therefore there could be two 100 point scores in the WILD class.
With the exception of the STIG class, and the MILD class there is a minimum of four cars per class.
A new class will start when the minimum number of cars is reached and the time differential between two cars in an overall ranking of fast time is equal to or greater than .500.
Once all four classes are assigned, all of the remaining cars will be scored in the "mild" class, regardless of time differential.
A car is classed by its fastest raw score -- penalties are then applied to move that car within the RAW time classification.(the penalties do not allow the driver to drop a class).
Fastest car in each class is scored at 100 points, and then every car within that class earns a ratio score of their time against the fastest time in the class.
One thing that we are currently working on is an overall bonus score so that faster cars are always earning more points than slower cars, despite the fact that a class score always earns 100 points. I'm still trying to figure out the best way to do that -- I'm looking to NASCAR for ideas :)
This works for our club -- we generally have a small number of cars compared to bigger clubs, and a much greater variety of types of cars (ranging from a mid-fifties Morgan race car to a new Porsche GT3). Our members don't generally "build" for a class, and as a result it's been hard to find a class-based system that works for our demographic.
I've run sample data from last year that illustrates the idea well. If you are interested in seeing it I could email it to those who want it.
Rob