Isn't the most useful model just the formla for the force of gravity? Sure there is lots of interest in understanding an underlying mechanism. But a useful model abstracts that away so you can predict behaviors.