Added missing prediction of rotation similar to the translative predictions

[cb3hjb]

KissMP works, the force stil works, it probably does something

[TheHellBox]

The reason rotation is not predicted is because it's not exactly needed to be done. Linear prediction is required, because with a ping of 100ms, at 10 m/s, there's a 1m lag. Rotations are much slower than that, so extrapolating that data wouldn't compensate for much errors. And compared to speed, rotations change suddenly and quickly (Snap oversteer for example), so the prediction is very likely to be working in the wrong direction and creating the error instead of compensating for it

[cb3hjb]

Would it make sense to calculate the change of acceleration in the translative direction?

Assuming that a car can rotate at the speed of 100 °/s (3,6s for full rotation) the possible error at the tip of the vehicle could be:

100 °/s * 0,1s = 10 °/s

(2*pi*r* degree°)/360° = arc length

(2*pi*2 m *10°)/ 360° = **0.349 m**

The sine ans cosine are avoidable which would make this less computationally expensive.