Jazzgene
you said
The thing is if the anti-skate force differs from the beginning to the end of an LP, how can a constant weight compensation like the VPI weight on a string work correctly on the entire LP? It can't, I would think.
If your antiskate is the JMW 9 then the force will increase as the arm moves towards the centre if the little arm with the weight is horizontal when the arm reaches the run out. if it is literally a weight on a string, like the old SME, it can aso be set to apply an increasing or reducing force depending on the angle it makes with the arm, though this doesn't apply if the string run round the diameter of the pillar and the radius it is acting at doesn't change.
Check the output wire and make sure it is not counteracting or adding to the antiskate. You can use headphones and a test record to check for distortion on one channel more than the other. If it distorts on the right, you need more antiskate.If the left, less. If it still distorts on the left with no antiskate, the wire is pulling outwards.
Using a greater VTF stops the distortion by nailing the stylus in the groove, but it also increases the skating force because friction is proportional to VTF. If this was not so, our cars would go faster as the braking force increased...