yblad

I'm semi-certian a value of zero is impossible and is automatically rounded up a value on runtime. If it is at zero the game would be allowed to slow the simulation all the way down to a standstill i think. Or it may be that there is a failsafe in the code to stop this in which case setting to zero effectively is no different to setting it to one step higher anyway. I may be wrong on that, but I think I'm right. Actually setting it lower helps the physics be more accurate, but at the expense of your simulation time being lower than real time (1s in the game is less than 1s in the real world), setting it higher sacrifices accuracy in order to keep your simulation time high. If you want to make the game appear to run fast you want to be increasing this value. With a low value it run smoothly but slowly. So you will get better framerates but at the expense of things happening slowly on the screen. It doesn't actually effect performance, just where the performance is spent. It depends on what type of lag you prefer, the jutter/skipping of a low framerate and a bad simulation or the everything going smoothly and accurately but it taking 5 times as long for things to happen. You can tell if the game is slowing down time by the colour of the timer. Green means everything is running at 1:1 speed, yellow means it's slowing things down.

They really use Euler's method? Jees I would've thought they'd have at least gone with a 2nd order Runge-Kutta algorithm. Much better performance/accuracy ratio. Maybe i should write to them and offer them a hand. Really don't have time right now though