"...the more off center you pull on the string the more off set you would want in the tiller."
As I understand it this premise holds. What most folks don't appretiate is that the nock point for most bows is 1 1/2" to 2" or more above center. Take the string off your bow, fold it in 1/2 and see what I mean.
Moving a finger from above to below the nock point actually means you are pulling the string closer to it's center, and consequently the bow's center. In this scenario you need less tiller, because of drawing the bow relatively less above center.
Ideally the fulcrum of the bow hand AND the center of drawing force on the string are both at dead center. Then you'd need no positive tiller. But since we can't pass the arrow through the center of the bowhand we compromise with nock point above center and positive tiller.
I think Dean's discussion of tiller had more to do with the pecularities of building selfbows, where balancing the load to the design is paramount in order to ensure longevity, viability in some cases.
While I'm not sure the implications for nock travel work in our favor, in terms of tuning out interference w/ the arrow pass, he postulates negative tiller would better match the bow design where arrow pass is arbitrarily placed 2" above center, or the so called (ironically) "even lenght limb." This because the upper is under more stress. Being inherently shorter it can't bend to the greater curvature this geometry requires without taking more set. I suppose one could compensate with more width on the upper instead.
More accurately, Dean uses this argument to support his preference for placing the arrow pass at dimensional center, rather than 2" above, which in turn calls for less positive tiller as there's less asymetry in the design to begin with. With glass bows it doesn't matter, as we don't need to work so close to the materials' elastic limit. And as I understand it having the fulcrum of the bow hand closer to center, versus the arrow pass as again they can't occupy the same space, makes the bow "more stable." Not sure what that means, or how much it matters.
I've built 'em all different ways and can't really see a big difference. Despite all the theory, the perfect tiller and tune come from trial and error, matching the bow to the archer's form. As such, I'd argue there is no constant, ideal config, rather one puts a stake in the ground somewhere (with a little positive tiller) then adjusts (nock point) for each circumstance. Or with selfbows, shooting the bow in while tillering, watching set closely and adjusting to avoid excessive set on one limb or the other.
My form is such that I might nock above or below on any given shot and still have the same arrow flight.... poor. ;-)