The 'fun', or is it that other F word (frustration) with modding is figuring out how the game works, while most of us aren't Adrian Newey ourselves either. So who is wrong and why? I don't know! And NEVER trust someone who says they know!
Almost all the time I blamed the physics engine for something I was wrong. That goes for almost all people in the past. There was a front engined Porsche 911 for rFactor 'because physics engine bad' .. Karts weighed 400kg because 'physics engine bad'. Meanwhile this was all down to incorrect numbers and you can make a pretty good 911, and you can make a 2kg remote control scale car. People thought the tire model was bad, and while it might still be, having seen and copied the source code, the pure and combined slip behavior, with the right numbers, seems completely decent.
The rigid body solver seems very clever and I haven't seen any weird things with the suspension kinematics yet. And having been wrong half a dozen times, I tend to trust Terence to have come up with mostly very solid code back in the early 2000s.
There are numerous weird decisions made regarding how certain things work, but blatant bugs while I'm sure some must be in there, I haven't put my finger on.
You seem to have many (valid) questions on the origin of suspensions etc, so it is very likely that at least a PART of the problem is you're not 100% sure if what you think you put into the game is actually going into the game as expected. Then you also can't trust the outcome and you get a growing number of question marks.
And you come from another sim if I read into it correctly, which while probably also made by smart people just like ISImotor was made, could have its own odd decisions and/or bugs in it. So you can't be 100% sure if your reference is just what you're used to working with, or if it is actually the 'truth'..
Again this is the fun and frustration of sim physics work, there are just so many question marks that I don't think anyone can be truly confident of answering. Those who make the boldest claims (you don't, you have fair questions!) should be doubted.. a LOT
The more sure people are, the surer I am they're full of ....
Variations of the ISImotor engine, I hope AMS since we added probably actually some improved things, are still used by racing teams and for driver training. I have made perhaps half a dozen pretty good cars that matched telemetry quite well and also subjectively felt good enough balance wise. But you're in extremely subjective territory there as well. Perhaps a vastly less oversteery version might be preferred by another driver, or vice versa. Even good drivers might prefer the sim to feel quite different. All based on what they're used to, what they expect, what they enjoy.. And you can get the basic telemetry (speed, G's) to match very well with quite a car balance range of course.
I've never deliberately added say 10% grip to make something work. A big ish mistake I made was with the aero. In my spreadsheet a road car had 50/50 aero if it had 10kg front and 10kg rear downforce, while producing a ground tire 'driving force' to counter the drag. Such cars, when coasting, subjected to basically only aero forces, could have like 50kg front downforce and 50kg rear lift, enough to cause a huge amount of high speed oversteer. Thankfully Reiza also has less stubborn people (Renato) who would go over the car pre release and would have his own methods (setup, grip and aero tweaks) to bring it back to a better state.
I stuck with my spreadsheet, believing it, and believing I understood aero, not realizing this was happening and causing a larger error, especially on non aero cars. Aero is surprisingly important for balance, even if it is just a bit of lift or downforce as 50kg can be 10% of the weight of a Formula Ford or something, enough to completely change the balance. (it wouldn't be 50kg in that case but you get the point)
One problem this gave cars was instability off throttle and trail braking, mostly done from high speed. Now that I found the problem with the aero, cars behave better and don't need as stiff of a front suspension which I might've ended up with before to stop excessive oversteer. Of course solving an aero problem with a stiff front only creates understeer (and often snap oversteer) at low speed..
But there are still issues I don't understand. Kerbs don't work on stiff racing cars. I think it may have something to do with the tire contact point being only vertically flexible, and also the tire stiffness being the same 300N/mm or whatever, even when the tire hits a sharp kerb, which would make the spring rate a lot lower than when trying to compress the tire on a flat road.. But I am not sure what it is.
Another issue is that there might indeed be too much oversteer on RWD cars and perhaps too much understeer on FWD cars, with how I make them. A car with not a lot a HUGE amount of power (1300kg 250hp) can easily be made to go sideways mid corner with some throttle, where the track car I drove in real life would most likely understeer from this. But I had zero data from the real car, and it isn't hard to make a sim car do this as well. But then I'm not sure if I matched the handling correctly, or if I had to put in say a too stiff front for understeer, to fix perhaps a tire model issue, so I'm just breaking the handling elsewhere.
A more recent thought I have is that the lack of compliance starts to add up. If I would go from 0% to 100% throttle in a millisecond, then one millisecond later, whatever the engine torque maximum is at that RPM, arrives at the rear tires. Way before load transfer has taken place. In reality engines don't respond this fast, flex in the drive shafts and tire also absorb this sudden torque. Perhaps compliances like this smooth out the torque delivery enough.
It would be a decent explanation that a car goes sideways mid corner because it is applying full driving force on the ground say 0.5 seconds before the load would've transferred. And if this is enough to get say a slip ratio of 0.3 going, well there goes your lateral force..
Geez I've just written a book again..
