Hmm, I have tried numerous possible solutions but nothing solid. I’m afraid this might not be possible without a fixed time step version of Box2D. It would solve problems with box2D overcompensating when a Codea is paused too as far as I can tell.
Without any idea of the integration of box2D in Codea it’s hard to gauge exactly what it is but the timestep issue seems intuitive. I was hoping Simeon or John would give their view on this.