Junk Russian mods are not evidence of anything. If you can replicate the problem on a vehicle that comes with the game, then it will be worth looking into.
There are 2 effects at play. One is due to slip angles. A basic diagram of a cars' steering has the car turn about an instant center in line with the rear axle. In reality, due to both front and rear tires' contact patch flexing (ie slip angles) the car actually turns with an instant center further forward near the center of the vehicle. When you look from a camera locked behind it looks like the butt is coming out a little bit and that the car is turning about its center. It will look this way even if the car is not oversteering. Second, the player is using the keyboard which throws the steering right to lock way faster than happens in real life. A sudden jolt left or right at the front wheels causes a torsion in the z axis about the cars center of mass. Front turns left, rear gets thrown right. If you're playing with a keyboard, throw your "feelings" about how a car handles out the window, because nobody drives that way in real life. Play any decent racing simulator and rapid keyboard steering will have you understeering or oversteering off every corner.
Here is what I think is wrong with the tire traction: It's too unresponsive. My conclusion is made this way: LFS has awesome tire physics in my opinion when driven with wheel, BeamNG is undecided at this note because I have yet to test it with wheel and FFB. LFS with keyboard-steering and similar steering rates as BeamNG is SUPER sensitive to spinning out. Even the normal XRG car. It doesn't take much for the car to lose control and end up in a wall (even when driving straight, once you try to correct at all you lose all control). So this leads me to believe that the problem is within the shape of the traction curve and its response. The maximum forces seems good. What could be done is to make a traction curve tester within beamNG and reference that to a real traction curve and most of the "unrealistic feel" causes should reveal themselfes right there.
Well, we're very aware of the "too unresponsive" problem. The issue is fixing it. It seems to be mechanical, and not having anything to do with the friction code or traction curve. We do our best to improve it, but it's really hard to figure out what works and what doesn't.
V1 Increased wheel nodes: from 24 to 70 = much smoother sliding/friction, and no more wobbly wheel when burnout...with rear wheel drive cars, just try to burnout with bolide GTR original wheel model This has allowed to get higher G-force in the curves from 0.95G to 1.5G with Covet race, on Industrial Added different toe to Bolide, Front: -2% Rear: -15%/+2% It was not much time to work on this mod....please test this out and answer how you like and what can I change to get better wheel mod! V2 1. New groundmodel for better and more realistic friction. 2. Added soft steering, test out, and compare. Please select every time the soft steering when you change to other vehicle... 3. New cameras GoPro style on Gavril Grand Marshal and Ibishu Covet. More later... Some video when I get time to it.. Test and compare!!! Have Fun! Wheel-Handling-improvement!!
You chose wrong factors to compare. The problem is that pretty much none of the roads on the maps included are of a race track quality you've experienced in real life, these are worst grade, worn out tarmacs that haven't been repaired for decades. Potholes, cracks, badly made patches and sand scattered around everywhere galore. And how often do you ride your real car at frantic BeamNG.drive speeds on such crappy roads like these? Let me answer that for you: never.
Some notes: Above youtube video is from the the older version before the latest update. A lot of things have changed since then. Also if you want to see the weight transfer happening in real time, add the per wheel weight app by pressing the "+" on the lower-right. - - - Updated - - - Very interesting experiments . Some questions. Did you also adjust the wheel weight (node mass) when you went from 24 -> 70 ? If you didn't, your wheels might have unrealistic mass. Even so, i find it very interesting that the red G-meter dot is so stable. Also the 1.5G with Covet race, seems to me to be a little excessive. A real vehicle wouldn't be able to pull that much Gs. Thank you again for your experimentation .
he mentions in his thread that he did adjust the node mass. when i was driving with an earlier version of the mod, i checked the weights myself because the wheels felt heavy and more unresponsive to direction changes - but the weights given by the debug menu looked very reasonable. maybe there is another factor that would make the car 'feel' heavier simply by having additional nodes?
Yes, its not just wheel weight that really matters, its where the weight is. All those new nodes on the outside *may* have changed the moment of inertia compared to the old wheel. A funny thing to do is get some wheel nodes with a lot of mass on the hatch, change to zero grav, spin up the wheels and steer. GYRO EFFECT! lol
It happens with the default cars also. You can replicate this for example with Civetta by removing the body and placing the camera above the car. Use a map with grid, drive straight about 50-80 km/h and start making small turns to right and left. It's like the car is turning around a vertical axel placed on middle of the front axel (marked red in the picture below). (imported from here) I don't know if it's the driving model or is it just that rear tires loose their traction with even a tinyest amount of lateral force affecting them.
The chase camera is fixed to the center of the car, so it looks like the car turns around that point. Try the same in free cam.