track/car co-evolution

[honnza]

I've got this idea:
Let's have a set of C cars and a set of T tracks. All cars are evaluated on each track;
Cars try to get the best score; the fitness is the average over all tracks.
Tracks try to maximize the difference between the car fitness values; I suggest standard deviation divided by the average. All tracks end at 1000 (or sooner).

[QwertyCat]

Clarification : You have x cars that go on x amount of tracks, then find the mean, or average, of all the fitness of those tracks.

1. What if one car excelled in tracks that are similar, like the hills and the peak, then lagging behind on other tracks like the raceway and the gap?
It would be equal to an all-around car.
Example: Cheap Car, 2, 2, 0, 0. All Around Car, 1,1,1,1. Cheap Car = All Around Car.

2. Rather than a limit of the distance of the track, why not limit time?

[JesicaMooe12]

Cool stuff man