It took me a long time to understand as well. Now I'm grappelling with mass and it's relationship with damping. If you are like me, you think about this stuff all the time, and things come to you, (obviously, look at whats come back at us).
This part I think you might be missing- you mention that all (most) your records are not off center. While there are quite a few that are obviously off center, what we are talking about here within the cantilever/coils is microscopic. All records are off center, evan if just a fraction of a groove width, they aren't manufactured to the kind of tolerences we are striving for. If you dismiss this, then you shoot for the arm staying put at the cartridge. Really, what we want (in theory) is the arm to stay put directly over the cantilever. The key to this, with fluid damping, as stated by Z above, is the viscosity. Control the resistance selectively at selective speeds (traditionally, the speed being the frequency of the warp/wow). To see it, you definitely need to get in and look at the cantilever at the cartridge to see how much it is moving in relation to the cartridge. to measure it, you definetly need test records and the like. ( I don't have these resources, maybe you do).
There is more to controlling frequencies then just keeping the coils centered, though. There are, I believe and suspect, differences and simularities between fluid damping and mass. We are not done with this yet.
Last night, I put my immedia back on to test some of this. I drained some fluid out until the arm could tilt easily to one side but come back without overshoot. Then I weighed some pennies, came up with 5 pennies on each side to be about 12 grams with electrical tape, and slung them out with a paper clip. They are about 1.75" out on each side, just below the record height, slung out at the same angle as the cartridge, and just taped on the top with a piece of electrical tape, over the center of the bearing, and they swivel front to back because of the tape. archaic, but they increase the horizontal mass considerably with minimum effect on any other parameter.
What I thought I heard was a wider soundfeild with more separation between intruments. But the induvidual instruments seemed more truncated, less air and detail around them. The high end seemed more recessed with a loss in detail. I didn't adjust damping at all between the two. I did notice more movement of the cantilever, at least I thought I did. Sounded similar to adding to much damping, and of coarse this experiment is extreme, I didn't play with it any further.
One resource I don't have lately is too much time. I am currantly getting more enjoyment in the tweak factor by hearing your results.
The immedia has damping, and it is easy to control and adjust/change, but it is also much heavier than the rega, I plan on starting another thread if you don't to see if anyone has tried fluid damping with the rega, and if any is readily available.
This is a new tonearm you are designing, isn't it? You are changing the parameters of the design. Seems to me you are somewhere roughly as far from the rb-series as the ol 750 is.