Cueing/Rega RP1w/Bias 2 cartridge

Your dealer provided half the story. Some arms disengage A/S until the stylus touches down, but on Rega's A/S is always engaged (at whatever level the slider's at). Since the arm is always being pulled outward it has a tendency to drift until touchdown. Some resist better than others depending on friction, as Jrtent noted.

Swampwalker and Elizabeth provided the rest of the story. Setting A/S much lower than "equal to VTF" will reduce cueing drift and often provides sufficient anti-bias to track real world records properly.

I strongly agree with Elizabeth's suggestion (start with the slider at zero). As Jrtent described, on many Regas even a zero setting still results in some A/S being applied. This may be all you need, but it's cartridge specific so user experimentation is required.

One clarification: don't listen just for channel balance; A/S doesn't always affect that. Listen also for R channel mistracking on difficult-to-track passages. If you hear that, nudge A/S teensily higher until it tracks both channels equally. That's all the A/S you need, using more often does more (sonic) harm than good. (Note: if BOTH channels mistrack you should increase VTF, not A/S.)

I'm sure your dealer is right. A friend's P1 (also with Bias 2) does the same thing. That doesn't necessarily mean the antiskate is set too high, just that the cueing mechanism doesn't provide enough friction to keep the arm in place against the outward force exerted by the antiskate setting. FWIW, my friend set his antiskate to its minimum value so the cueing worked more accurately, and with no drop in sound quality or tracking ability.

Instead of lowering the anti skate just increase the skate.

Should have said lower anti-skating.

Experiment w different anti-skating force.