Hehe, thanks. You should be in for a pleasant surprise since the new version behaves a lot better than the last demo. There were actually a number of fundamental flaws in that version of the physics that held back realism.
(heads up - this got kind of wordy )
The simulation model is all original, I'm not familiar with the Marco or Pacejka methods. Actually, up until a few months ago, I made an effort to stay away from reading material on car simulations and physics in general just because I wanted to figure it out myself. Looking at it now, that was kind of dumb, since I quickly got answers to some problems that had been bugging me just by reading a few light articles on dynamics (the problems were mostly regarding torque). Making everything on my own (during more than a year) did give me a lot of insight on what I was actually doing though, so it's probably a good thing to do as long as you don't go to extremes. You should always think about a problem for a while before heading online to find someone else's solution (which might not even suit your application that well).
Anyways, as for general physics tips I guess you should read some stuff about mechanics/dynamics (relationship between linear force, torque, mass, velocity etc) and maybe some articles about simulation. Then develop a program that simulates a sphere or box interacting with a simple plane (collisions and friction). If you want it to be realistic you shouldn't aim for simulating a car right away, the back bone of my model is a generic rigid body that could be anything really. It has no notion of shape, just position, orientation (which is a matrix btw), velocity and torque. The actual "car physics" are all in the simulation of suspension (springs) and tyres against ground. The forces resulting from those are then applied to the rigid body.
When it comes to getting these components right, I'm afraid it's just down to looking at the indivual problems and "working on them". It took me more than a year to get to this point, but I'm sure similar progress can be done in a few months given a more efficient approach. As for details, springs are quite simple and documented all over the place. Tyres are trickier but a simple model based on the "grip circle" will get you far (that is, forces up to a certain magnitude are applied in full but larger magnitudes are clipped to the maximum limit). A key thing is also that the car's weight should be distributed over all four wheels according to the "weight shift" of the body. I did this wrong in the old version, but figured it out in this one. You get the "weight" at each wheel by looking at the lifting force provided by the corresponding spring, and then you multiply those with the forces at each wheel. The general rule is that, at full grip and locked brakes on all wheels, the combined force should be exactly big enough to make the velocity equal zero (remember that the force is applied as velocity=velocity+force/mass, at least with euler).
ANYWAYS, I might write a tutorial series later on (I kind of started one earlier, but it was horribly flawed and has been taken down). It most likely won't happen for another few months though. You can always ask me specific questions here or via mail and I'll try my best to answer them.
Good luck with your project and let me know if you come up with anything cool
Thanks very much for your reply, sorry I didn't reply sooner - I don't always remember to check forums I've posted in ;o)
Anyway, my simulation is coming along slowly, although I've been distracted by getting my tracks up and running.
However, I'm coming back to physics now, and I'm taking your advice about applying the car physics to a generic rigid body. I've got it more or less working although I'm not looking forward to implementing collision response
Anyway, thanks for your advice, and I'll post again if I produce anything interesting.
By the way, looking forward to the next demo of Rigid Racing...