While it is tempting to go on about how to calculate the loads and stresses on a tower supporting a wind turbine, I have to step back and admit that I shouldn't do so publicly - there's a ring on my finger that reminds me not to give too much advice outside my field of expertise. Rich Hagen pointed out above that there are a lot of variables, which is very true. Any attempt to simplify the analysis to a single line of algebra will be an estimate only. There are estimates, and then there are estimates. Some over-estimate the number, others under-estimate it.
I want the estimate that over-estimates the loads.
I did say I've done the math, before, and it falls into the "estimate" category, and the same goes for what Dan and Flux did. We, in fact, do get the same numbers, but via different means, and by adjusting variables that Dan and Flux didn't use, I can show that a windmill's drag could be 50% higher.
I'll never be able to design a windmill that captures all available energy within the Betz criterion through its disk, but I will design a tower that can handle it, plus the safety factor (4 sounds right, but I'd look that up, too).
On to furling...
Has anyone tried changing the angle of incidence of the tail? The plate on the tail, whatever its shape, is generating a force to one side or the other, depending on its angle of attack. You could change that angle of attack by, say, shimming the plate away from the boom.