Alasdair, after looking into this stuff, I've come back to the graphical method as presented at the Palos r/c site as adequate, and figuring only the wings seen in the plan view as the parameters..
The attached Pitts demonstrates two items... I found a plan for a Pitts in Model Airplane News, March '95 which shows the c.g. at the leading edge of the lower wing.
The Palos method gives this same location for a 30% c.g.
If the "mean wing" is located between the wings per the areas, this ratio is about 97% for a typical Pitts.
Where does the "mean wing" then go?
Another method pointed out by acropilot_ty uses only the stagger between the wings to locate the "mean wing".
Using this method, which ignores area altogether, the "30%" point is always aft of the 30% point found graphically.
As I've mentioned with my biplane Kadet, the lower wing has a "setback" from the upper wing of 3", with a chord of 12" on the lover wing. Upper wing chord is 15"This places the lower wing area completely -under- the upper wing area; cannot be seen in a plan view. The plane can fly sans lower wing and no other changes.
Any method other than the graphical computes a c.g. way too far aft for stability for this biplane/monoplane.
2nd image as a biplane
3rd image as a monoplane.