One of the nice things about using simulations to illustrate a point is that they always do exactly what you expect, well almost always. Today I challenged my first year students to draw st, vt and at graphs for a bouncing ball, a pool ball and a pendulum. They first sketch what they think is the answer, I then do them on the board and prove that I'm right by showing the motion and graphs with interactive physics. Everything went well until the pool ball. I specified the velocity was constant across the table so acceleration was therefore zero. Not according to IP, there is even a vector clearly showing that there is a constant acceleration. This was quite worrying but a quick check with Algodoo showed zero acceleration with sharp peaks every time there is a collision restoring my faith . Still don't know what is going on in IP though, maybe its just getting old