Just thought I would post a follow-up on this. I've spent some time running different media and have come to the conclusion that the amount of liquid will differ depending on the media and the parts.
My current approach is to fill the bowl with media and parts, then pour in liquid in fixed amounts until I see the tumbling action slow, then withdraw the last amount added (basting syringe works good for this). Then write in my log all the details so when I use the same mix of components I know how much liquid to use.
Anything added to the liquid, i.e. compound, soap, etc., can change the lubricity of the liquid which can change the tumbling action, so has to be included in the evaluation matrix.
I had no idea of the challenges that tumbling offers!