Remove your bias for better sound

Al/Doug - To me Doug's logic has a flaw.
Although the stylus is being pulled inline with the cantilever, there is a resultant force on the arm itself pulling the arm toward the centre due to the cartridge offset and overhang. Now the arm also has an inertia resisting this inward force proportional to its effective mass. Therefore there is more than likely a conflict of forces on the cantilever - inward force applied from the offset & overhang versus inertia of the arm. To my mind the amount of antiskate required is that required to keep the cantilever in line with the groove. I very much doubt that this would be 0, but it would also be influenced by the horizontal compliance of the cantilever pivot.
Intuitively unipivots have lower bearing friction than traditional gimbal arms and that would reduce the antiforce required assuming the same cartridge and tonearm effective mass is the same.
Let's face it the VPI unipivots are junk anyway, they are overdamped, the leadout wires are so rigid they must be loading the horizontal movement of the arm, the pivot point is upside down in terms of acting as a mechanical diode and every one I've heard has lacked transparency and sounds slow and turgid - this might explain why no antiskate sounds superficially better on these particular arms.

Doug/Al - I 100% agree with Doug's comment that too much antiskate is applied in most instances. The proof is that very few cartridges have straight cantilevers after a short while. The most common mistake I've seen is people using antiskate to adjust channel imbalances in the system. The other major factor is armboards that are not level. How many people 0 balance their arm and make sure there is no drift in or out before they start. This is a simple check to ensure there is no inward or outward pressure on the cartridge from the horizontal bearings in the arm being out of level because the armboard is not level. When I recently installed a new arm I had to shim the armboard by a only few micron to ensure the arm had zero drift and the resultant decrease in antiskate required was significant - 0.5gm.