- Solid body mechanics are well understood and certainly qualify as physics.
- Suspension mechanics are well understood and certainly qualify as physics.
- Aerodynamics (as simulated in racing sims) basically boil down to a set of extra force vectors being manipulated in response to pitch, yaw, roll, ride-height, wind and drafts. MADNESS supports this (yes, wind too).
- Collision behaviour is well understood but expensive to simulate. Also certainly qualifies as physics.
- Driveline physics are well understood and certainly qualify as physics.
- The Tyre<->Track interaction physics is a really difficult field, yet Reiza (who have experience with both gMotor1, gMotor2 and MADNESS mind you) are on record asserting that MADNESS offers the more advanced models of the three (SETA and LiveTrack 3.0).
- FFB is part how you translate the forces and moments the tyres generate into the steering rack and part "everything else" (acceleration cues, chassis vibration cues, chassis and joint twisting cues). Then comes the part where you need to map the physics input to a wildly diverging set of hardware output devices with all sorts of compromises inherent to each specific piece of hardware. And finally, some people prefer an exaggerated response to certain physical events, though they may not even be aware of this being the case. Reiza asserts that the FFB primitives or building-blocks available in the MADNESS engine leave nothing to be desired and also notes that the track surface may play a surprisingly large role in the FFB experience.
- How the camera moves in response to the car changing direction also influences the simracer's perception of movement.
- Finally, the car/sim calibration that brings all of the above together.
Having read a lot of arguments surrounding what people like and don't like, I'm inclined to think that the combination of consistency in car calibration quality, reasonable collision behaviour, good camera/sound/FFB cues (with as low latency as possible) is what most non-engineers are really referring to when they use the term "physics".
It might not really have anything to do with the depth of simulation in various subsystems per se. As an example, note that despite AMS1 using a much simpler tyre model than PC2, people are still praising AMS1 for its "physics"'.
In contrast, people insist that pC2 is all over the map in terms of car/tyre calibration and some people find it difficult to deal with the camera and FFB settings in pC2. The end result is that their pC2 experience is inconsistent (not to mention different to what they're perhaps used to) which leads them to conclude that they dislike it and wave around the label "physics" to accompany this dislike when what they really ought to be saying is "the driving experience feels inconsistent/off for me and it's certainly different than what I'm used to in other sims with different tyre models".
FWIW, we had a guy in the PC2 WMD forums who stated he owned a Lancer EVO mkVI. He felt the in-game representaiton drove *exactly* like his car, meaning that he could make the same inputs in-sim and IRL and get the same outcome. If his words are to be taken at face value, that's a data point which suggests that the MADNESS engine *can* produce "good physics" (i.e. a simulated driving experience consistent with reality) via correct calibration. OTOH, others couldn't get their real life vehicle experience to match up with the experience PC2 served up, so it's certainly not just a case of "plug in the correct numbers and off you go to the races". And I know for a fact that the vehicle dynamics guys didn't just put random numbers into the various physics files.
But I digress. Didn't intend to derail the thread.