Gabkicks- that is an excellent example: variable grip.
So you have a track, and the grip seems to vary around the lap ... but why? If you're writing a game and the car goes unrealistically fast through one corner, you say to yourself "Well, it's probably to do with the asphalt in that corner, so we'll take a few percent off the grip level". An -hey presto- the car goes at the right speed through the corner, and everyone thinks they've done a good job.
This is exactly what we had with one circuit that wasn't laser-scanned, and we reduced the grip in one specific corner. Then along came the laser-scan, which (to look at) wasn't different from the original ... but suddenly the correlation was excellent without the need for reduced grip. The reason was that the bumps in the surface were causing load fluctuations at the contact patch, and this causes grip to drop (which is why you want as much damping on your car as you can get).
Now we start to see the difference between a game and a simulation. In the game, the car will always lose grip in that corner, because the grip-loss is built into the circuit. But in the simulation, it might be possible to fit the car with better dampers, and reduce the load fluctuations, and go through the corner faster. If you're just using your sim for circuit-learning then the game will probably be OK, and the lap-times will be about right, as will the top speed, apex speeds etc. But if you want to do any serious engineering work, you need to have a simulation that not only gives the correct results, but gives them for the correct physical reason ... so that if you modify your car, its performance will vary in a realistic way.
So if you look at the case of improving the damping, you need to model a damper in such a way that it has the same force-vs-velocity characterisitcs as the real one. So you need an actual damper and a test-rig to produce the figures. As far as I know, no game-maker has access to such data, so there's no way that their software can be giving the right answers for the right reasons. And that's just the relatively simple case of dampers ... for aero modelling you need a wind tunnel to produce the numbers! Want to make a realistic engine? You'll need a real engine and a dynamometer to measure it ... and you'll then need to incorporate a real F1 SECU to control the model, otherwise its behaviour will not be correct.
And that is why it gets my goat, when somebody asks "Is it based on a game?"
END OF RANT
)
. 
