The formula that was given assumes that you're starting your cut at a 90 degree angle (perpendicular) to the top edge (should the top edge have been parallel to the ground)...
Just to set the record straight, the formula linked has nothing to do with anything surrounding the circle (or radius), just the circle itself. The cut coming into or going out of the desired radius has no effect on the radius itself or the formula used to calculate it, so what was quoted above is incorrect.
Oh, and that's quite an interesting approach you have there Quarters...
Sorry for any confusion I might be causing, but what I'm describing
is a concern... although a fairly small one...
Hopefully the attached picture will help my explanation... 'h' and 'c/2' are given as they would need to be for
the equation in question. The problem is that the derivation of the equation relies on 'h' being a part of the radius of the circle while a radius will intersect the circle's circumference at a right angle. This would mean that the angle 'A' shown in the attached picture would have to be a right angle which it is not (even very near where 'h' and the circumference intersect). Given that 'A' is not a right angle, the equation would lead to an incorrect value for 'r'.
as stated before, to get an angle of 'A' that is a bit less than 90 degrees, increase whatever 'r' the equation gives you by a little bit and you should be good.
then again, maybe I'm not "looking at the same picture" you guys are: maybe you guys are using a different 'c' and 'h' and I'm just confused

I would also use the flexible ruler idea... you'd just be using a big protractor then
