در مورد فیزیک بازی:
Do you get to drive them?(To gauge physics/handling
We don't really need to drive them. This is the thing that people misunderstand about racing games. If we were not trying to be a sim, we might have to drive them. But because we're trying to be a sim, we don't have to drive them. The reason is, when you think about a car's 0-60 time, or its horsepower - those are all the results of physics. Those aren't 'the physics'
So we are recreating the way the car makes power, and how that power goes through the inertia in the drive shaft, and where the friction is, how long the suspension is, where the tyre patch is in contact with the road and how it move and what the resulting physics of that is. The reason that a car in the real world might understeer or oversteer is because of those attributes. We measure and model all of those physical attributes
But how can you be sure that the physics engine is working properly if you don't test at least some of the cars? Surely you need to have a baseline for the physics engine, even if you've got all the data to feed into it
So, we do research the cars. We have a whole bunch of things like 0-60 that we research. And yes, there is a test suite, that's an automation that goes through and runs all the cars from 0-60 and we're able to see that our game hits the same numbers. But we've been spot-on with that for years
The level of detail and simulation that we're into, it's about polar moment of inertia and the way the car moves around, which isn't going to show itself in 0-60. You can hit 0-60 by hook or by crook, but actually getting the way it's going to make Gs and the way it's going to add yaw by moving the weight to the front versus the back - this is all about suspension and tyre physics and inertia through the car
The only way to get that right is to measure everything on the car - really take the car apart and build a math model that allows us to reconstruct the car accurately. There's no way to test that against the real data. There's not a widespread database of cars polar moment of inertia. There's a small number. Whereas 0-60, there's a widespread database. So that's why we automate the testing of that