Rob wrote:

... sometimes you just have to rely on zen.

Zen works if you know where to give something the english, so to speak. With wing twist I couldn't figure out what variable to tweak.

But I think I licked it.

Aft of the mainspar on most wings the curvature of the airfoil toward the trailing edge is so small as to be virtually flat. The plan view of such a section would be three sides of a rectangle with the fourth side being the angled trailing edge forming a trapezoid. With curvatures close to nil, simple trigonometry can be used to calculate what the true shapes would have to be to achieve a desired amount of twist out toward the tip (wash-in or wash-out).

The key to the whole thing was recognizing that the trapezoid can't actually be a trapezoid, but has to be two joined triangles, with the joint being a "fold" that allows the tip chord to be set to its degree of twist WITHOUT lifting the trailing edge off a base plane. That "fold" runs from the top root corner to the trailing edge tip corner.

Assuming no twist, the front triangle has the tip chord flat on the base plane, and tilts up off the base plane until its apex is as high off the base plane as the root thickness at the mainspar. The trailing edge triangle joins that triangle, with its root chord edge tilting back down to meet the base plane, and the trailing edge itself flat on the base plane.

Use trigonometry to calculate the true shapes of the joined triangles and voila - you've got the true shape of the wing pattern aft of the mainspar (at least close enough for the tolerances involved in paper).

Once the formulas are derived, you can start to add twist and/or dihedral to desired amounts and work out the true pattern shape every time.

I've applied this method to three different wings already, and it's been working perfectly.

Without pictures it's a bit difficult expressing the concept clearly, but I hope there's enough here that it helps anyone who's been struggling to prevent their wings from twisting.