The distinction between a ground that moves and doesn't move is actually critically important when we consider the COR of the ground in this scenario. If the ground is some immovable perfectly rigid object, it has a COR of 0. If it moves just a single Planck length due to gravitational interactions, and then back again when it bounces off the ball, it has a COR of 1.

I prefer the latter solution because an immovable object is one ideal assumption more than necessary. It's not even possible in a perfectly ideal universe, it's just a made up concept that has no meaning in physics.