Michael,
Am I wrong in thinking that it's not as mere of a thing as what you wrote here reads as? The most important thing is what the ball does and the most important factors in determining the ball's behavior is the path and facial orientation of the club striking the ball. When you get right down to it, that's it. If there's variability in (just) one input of a system that's built on the input of two variables, that system's going to experience a lot of variability and it's usefulness as a model with predictive power is in doubt, no? Maybe I'm reading what you wrote wrong and what I should really think is that you're trying to say that shaft behavior is not all that variable shot to shot and as such this variable is fairly controlled. Thanks in advance for your consideration in answering--this is probably just minutiae but I thought you wrote an interesting post and I wanted to understand exactly what you were trying to convey.