oosyrag! I just was reading this again today, and your comment,
Teleporting the physics simulation ball to a separate location is a nifty idea! That seems like it would work well, as long as the physics simulation doesn't get thrown off with a sudden change in position from manually moving the object.
Struck me, because that is a problem I have already solved, in another situation. I figured others may want the information too.
Basically, like the manual warns you about, if you set the position manually, you could find the physics behavior doing some strange magic.
In my case, I found that moving the ball made the physics engines want to give it extra momentum, almost as if the movement was an impulse.
In my pinball game, I actually manually set the position of the ball a lot. And it took a zillion different tests to find a solution to that problem. But it's painfuly simple.
If you just disable the physics behavior, move the object, and re-enable the behavior in sequence, the physics continue just fine as if nothing happened. Just strait up turn it off and back on lol. In the same event block, or at least by the end of the tick.
I'm sure it probably adds a bit to the cpu, but I never saw measurable performance problems. And I do it several times every tick at some points.
If anyone else needs to know, there you go. And if there are better solutions, feel free to share.
But boy, that one stumped me for a long time. But it allowed me to add a few tweaks to how physics works in my game, as you can add events to customize physics behavior while it's disabled, and as long as you enable it before the behavior needs to run again, you're good.