Coeffecient of friction is a number used to calculate the amount of force necessary to overcome the resistance of a standard material to sliding over a test material. (Times area of contact, downward force, etc, etc,)
The force (whether in pounds or grams) produced by a moving record groove "dragging" on a stylus is a result of the downforce on the stylus times the area of contact between the stylus and the groove times the sum of the coeffecients of friction of both the stylus surface and the convoluted (wiggly) groove wall surface.
Using a grooveless shiny vinyl record surface to determine antiskate requirements is utterly useless because:
1.) The surface of the blank vinyl has a coefficient of friction approaching zero,
2.) The surface of a polished diamond also has a coeffecient of friction approaching zero.
3.) The area of contact between the (bottom tippity tip of) the stylus and the blank vinyl surface is also nearly zero.
In this situation, and using normal VTF, almost no inward skating force can be developed because virtually no frictional force can be produced. SO, you ask, how come when I set my stylus on a blank record without any AS applied, it goes RACING toward the center?
Well, that's because most tonearms today are very sensitive, have a healthy inertial mass combined with almost frictionless bearings, and internal wiring that presents virtually no torsional resistance. Since the stylus is not CONSTRAINED in a groove, but sitting UNRESTRAINED on a flat polished surface, even a tiny amount of torsional force is going to "fling" that tonearm toward the center of the record!
In addition, if the record is not dead flat but the slightest bit concave (like from a record weight or clamp), and if the platter is not dead level, but sloping the slightest amount down, away from the tonearm side of the turntable, these factors only add to the phenomenon. In fact, I've watched a tonearm slide over a blank record without the platter even turning!
So don't be fooled. The object of antiskating compensation is to "zero out" the clockwise torsional (rotational) force developed by a cartridge/tonearm system in PLAY, in a REAL GROOVE that has a MEASURABLE coeffecient of friction.
Also, a word to those who intuit that quiet, high frequency groove modulations don't present as much (frictional) drag against the stylus as loud or low frequency grooves: you need to re-think that notion using real physics. The (observational) fact is simply that those loud, low frequency grooves, although having a larger side-to-side displacement than quiet high frequency grooves, have their undulations stretched out over a much longer (groove travel) distance, and do not present significantly more friction to the stylus than the (tightly etched) surface of a high frequency groove. The reason that, back in the day, loud passages got worn out faster than the rest of the record, was due to the poor tracking of the very low compliance (bordering on NO compliance!) cartridges of the time. And to make matters worse, often installed in the very lightweight tonearms of record changers!
Note: proper antiskating force, as Warjarret points out, contributes only a small amount towards the actual "trackability" of a given cartridge/tonearm combo. More important to that part of performance is the VTF (of course) and most important, the relationship between the compliance (springiness) of the suspension and the effective mass (i.e. natural resonance frequency) of the tonearm. In other words, a correct amount of antiskating compensation can help get the last drop of performance out of a good tracking arm/cartridge combo, but antiskating alone can not compensate for a mismatched arm/cartridge combo that simply won't track properly.
.
