Ralph, I am saying that when the cantilever/stylus is tangent to the groove, the friction force generated by the stylus to groove resistance has a vector directed tangent to the groove, too. But because the headshell is offset with respect to a straight line drawn from the cantilever to the pivot, there will still be a skating force due entirely to the headshell offset angle. Interestingly, if you use an "underhung" tonearm (no off color pun intended) with zero headshell offset, then there can be only one null point on the surface of the LP with respect to tracking angle error, but at that one null point, skating force will also be zero, because the forces will line up with the pivot.
Tangency (null points) is not enough to eliminate skating force. Lew says it quite well, but maybe it needs to be said a different way as well.
Skating force is the vector sum of the forces of the record "dragging the stylus in a generally forward direction". Forces sum and cancel each other.
The amount of drag and its direction is a function of groove friction and geometry. The friction part is what makes it tricky, because this in turn, is a function of record velocity (changes across the record), the dynamics of the musical passage (varying resistance to larger "wiggles" in the groove), the condition of the vinyl, the shape and polish of the stylus, and ... I'm probably forgetting something, but you get the picture.
Here's a visualization that might help (and spare you the vector math):
- Hold your left arm out in front of you (horizontally) with your palm facing toward the right.
- Bend your wrist so your fingers point further to the right, so it resembles the headshell/cartridge offset.
- Have someone tug on your fingertip in a direction parallel to your bent hand.
So, even at the null point, there's some skating force.
As Raul and others have correctly stated, there is no single correct anti-skate setting for your rig because of the frictional factors I mentioned in the third paragraph.
Cheers,
Thom @ Galibier Design