I started playing around with time dilation in 4.20 as a way of validating some ball rebound physics that happen too fast to be easily seen in my level. I noticed something straight away: physics at a time dilation of 0.96-1.0 gives a completely different result to physics at a 0.1-0.95 time dilation.
I’ve massively increased physics substepping to get the best result from it I can, with up to 64 async sub-steps. But regardless when I shoot a ball relatively fast I just get totally different results at that arbitrary threshold. In fact, the 0.95 results are more stable and repeatable, so I’ve taken to dilating my game time to 0.95 by default.
Has anyone else experienced this?