And again, what the analysis of actual pitch trajectories shows is that the supposed effect of the varying forces on the ball due to the slight rotation in-flight is totally negligible! If they were not, the knuckleball trajectories would not be fittable to the 9-parameter equation, which assumes that the forces on the ball are constant after leaving the pitcher's hand. What the article suggests is really happening is that things like seam orientation do play a role in determining the forces on the ball at the release point, which in turn feed in to the trajectory. But these forces do not change mid-flight in any measurable way.
On the other hand, your description of slightly different knuckleballs that do different things is very interesting. The analysis of Dickey's pitches from last year did show two clusters of pitches with slightly different speeds. I wonder if a more extensive analysis of his pitches from this year would show more distinct clusters, indicating better control over his pitches, and if this can go some way to explaining his recent success.