I think we should add supplementary material, showing how the shape of the surface is found, and explaining how the rotational motion of the ball affects the center of mass motion. I think it changes the effective inertial mass of the particle by an amount that depends on the moment of inertia of the ball. Since the gravitational mass is not affected, this means that balls with different moments of inertia won't move quite the same way on this surface, in contradistinction with the usual equivalence principle.