I have a paradox here. Look at this diagram of colliding ellipses (they might be elliptical prisms in 3D). Now if you stretch the image (for example looking at the image from an angle) it becomes two colliding circles. Therefore you would expect by that argument that the colliding force would be applied through the centres of the ellipses (yellow line). But another argument says that looking close up at where the ellipses collide it is like two colliding planes and the force should be perpendicular to that (orange line). And this would cause the ellipses to rotate. So which is right? And will the ellipses be rotating after the collision? I have not found any equations for colliding ellipses as there are with colliding circles.