David,
Appreciate your comments, and as a practical - not just theoretical - approach, think you have it about nailed.
However, check several of the above posts for other considerations. My personal preference was to use the axiom that symmetrical sections have center of pressure at 25%, with no change at different AoA ( no moments about the cp ), which led me to hinge at 20% MAC. SOME, but not overpowering, return to 'trail' tendency. Ted F's comments are valuable. What I tried was to use the smallest total tail area sufficient for the needed 'turn.' That worked, for several - but not all - conditions. It made a model that went where it was aimed, but needed to be aimed there at all times. Fun, but tiring...
For modern CLPA demands, it is at least as important that a model LOCK into the exit direction, hard and instantaneously. All-flying stabilators are not good at that, particularly in severe wind conditions.
It IS theoretically less than the best to carry tail surface area that opposes the needed turn potential - EXCEPT when tracking out of a severe turn is just as important as achieving the needed, severe turn. All aerodynamic design is compromise: the stab+elevator solution is a win/win solution...