Flieger
Yes, in theory the angle of the arm on a torsion bar system relative to force will change the spring rate. However, why make it overly complicated. At the angles and range of movement we are dealing with this rate cange is probably very small.
There are also some small changes in the angle of the rear shock to the rear arm that will have some effect on the wheel rate but again, this is going to be small.
Anything with in about 25 lbs per inch should be usable as we end up balancing the system further with sway bars.
Another thought. I have seen 21mm front bars run w 26, 27, 28, 29, at 30mm rear bars. Most tuners used to recommend 22/28's as a solid set up. Now many are recommending 22/29's. That is about a 30# difference. Start looking at higher spring rates and the gap is larger.
fredmeister
I picked up what I believe to be torsion bar wheel rates from two sources noted here some where and some time ago. I guess there could be some error in them but I have seen them quoted by others before. One I think might have been from a Wil French or some such and the other was from someone that I believe tested the rates at the wheel but not 100% sure. If my numbers are incorrect I would love to know this.
Here is one reference as to percentage differences between torsion bars and something on calculations.
http://www.pelicanparts.com/techarticles/911_torsion_bars/911_torsion_bars.htm
Again, I realise this is a bit simplistic but getting overly complcated may not yeald any more usable information.