Mesael
Sorry.
I said
The angle changes towards the centre not varying much, but under 7cm radius, the closer to the spindle it gets, the greater it becomes, until at 0 it is a right angle..
What this means is that the inward force trying to rotate the arm stays more or less constant within a few percent reducing slightly following the tracking error curve of the arm (not the tracking distortion curve) which varies across the record. Depending on the overhang used and the arm effective length, at the inner null the force starts to increase rapidly.
So it would appear that the ideal would be to follow this curve, reducing slightly then increasing.
However, it has been shown that there are other factors which increase the friction force as the radius decreases, which would skew this result and make it more likely that the force should at least be constant then increase, or even gradually increase all the way across the record
When I designed my arm that's what I did, as I thought that, like tracking distortion it seemed to be worse towards the inner grooves, and less of an issue further out, and should be weighted that way. But it was all "suck it and see". So, really, a tonearm should perhaps give the user the option of trying different things.
The actual amount of antiskate needed is very much dependent on stylus profile and downforce. With enough downforce the stylus will never meet a modulation large enough for it to lose contact with the outside wall, or even throw it up out of the groove. However there will still be unequal forces on each side of the groove unless there is some kind of compensation.
As Dertonarm says, a longer arm has less skating force than a shorter one.
But it is still there, and it has to be dealt with somehow - you can't just sweep it under the mat....