I remember from when I did my first mildcat round between July 2008 to its first firing in July 2012 that it's not possible to dimension the radius at the neck because there is no distinct point there. The member of the bulletin board of which I was also a member at the time said such a measurement makes the drawing "constrained" or "overly-constrained." I drew things for years and years on vellum with triangles and a T-square. I never knew much about constraints.
I appreciate the effort you're putting toward this. It's amazing to me that you'd spend a small portion of your holiday weekend on my behalf...
It's just stacking error. If you have a right triangle and say one straight leg is X the other straight leg is Y the hypotenuse is always going to follow the pythagorean square. Further, the angles of the triangle are always going to be determined by the tangent of X/Y, in most cases, unless you're using a 45deg angle, the X and Y are never going to give a whole degree, or X and Y will not be whole numbers, they might not even be rational numbers. With this in mind, it's much easier to say the angle has a leg that's X long, and this angle, and let the computer determine what the value of Y is.
I'm working this weekend anyways, and this literally takes so little time, I think I've put 25 minutes in.