Bruce,
I agree with all your practical advice, but I'd still go for a furthest aft CG at 20%.

Just to clear up the theory, when we calculate a CG what we are really doing is trying to calculate the Neutral Point and then placing the CG a suitable Static Margin (Stability Margin) ahead of it, right?
The formula for NP contains the term Tail Volume Coefficient TVC (as V-bar usually) and also the term (1 - DWF)
where DWF, Down Wash Fraction, (the differential de-epsilon/de-alpha) is how much the downwash is compared to AoA (measured from zero lift).
So (1-DWF) is what is left of a notional angle of attack increase by the time it gets to the tail. i.e. if the tail gets an addition 1 degree AoA the tail gets 0.3 or 0.5 or 0.7 degrees or whatever.

When we calculate pitching moments, lift etc. we have to use the whole of the wing area. We also have to use the whole of the tail volume. You are quite right.

When I said I reduced the TVC, what I am really doing is reducing the (1-DWF) term because on a triplane I believe it will be about a third of what it would be on an equivalent monoplane. So I use the monoplane CG equation, but divide by 3. The (1-DWF) is an invisible built in factor, so I said I was dividing the TVC by 3 instead.
What I should maybe have said is that for a monoplane I use
CG = 10 + (40*TVC)
for a biplane CG = 10 + (40/2)*TVC
The biplane one works well
so for a triplane I would use CG = 10 + (40/3)*TVC
but I have never built a triplane or even checked it on a triplane.
Those figures are designed to give a 15% Static Margin.
On this triplane I would maybe reduce that to 10% or 11% to reduce the tail download, hence my original 16% CG position modified to 20%.
Alasdair