Anti-skate applies an OUTWARD force to counter the inward skating force of pivoted arms. It varies with position and groove modulation. All anti-skate mechanisms are just bandaids. My FR 29 and FR 54 arms have no anti-skate and sound fine!

Minimal outward force is needed to properly offset the skating force that exists due to the overhang of the cartridge.  Your description of the outward jump indicates one of two things (maybe three).  First is that your tracking force might not be adequate and worth rechecking.  I would place the Anti-skate weight as close to the pivot as possible - first notch.  Double check that the TT itself is level.  If the VTF is correct, antiskate is set to its minimal setting, deck is level, and this backwards “jump” continues to occur there is a defect in the arm to address.  

@drrsutliff is correct. The problem is not your antiskating. There is something else wrong. Check to see if you can swing the arm over the spindle. If you hit a hard stop something is blocking inward travel of the arm and it is hopping out of the groove. Once the arm loses contact with the groove it is always going to hop backwards. 

@jasonbourne71 And If I were you I would throw those tonearms away or put them up on a shelf for display. There is no such thing as accurate antiskating, but having it is WAY better than not. Without it you will constantly mistrack the left channel and prematurely wear out your records and stylus. I have also seen that approach permanently ruin cartridges within a year by causing a permanent cantilever deflection towards the left channel. 

ANTI-Skating is VERY IMPORTANT.

The makers and re-tippers say the MAJORITY of used stylus they evaluate are worn on one side only.

Inward skating usually cannot be seen (except on a blank lp). Anti-Skate also cannot be seen (except on a blank LP)

Forget the TT’s built-in indicators for both tracking weight and anti-skate:

I use the blank side of this LP (copy/paste in Amazon search):

Hudson Hi-Fi Turntable Cartridge Alignment Protractor Mat (Single-Sided Print, no Strobe Included)

1. Anti-Skate to Zero

2. Tracking Force Applied (digital meter) copy/paste in amazon search:

Neoteck Digital Turntable Stylus Force Scale Gauge 0.01g/5.00g Blue LCD Backlight for Tonearm Phono Cartridge

3. Blank LP, still no anti-skate.

AVOID stylus tip falling off the outer edge, and pulling too far in when manually spinning, i.e. be ready with your arm lifter.

3a. Lower arm onto the blank lp, about 1-1/2"" in from outer edge. Spin platter by hand, see the natural inward skating, you don’t want that.

3b. Now adjust anti-skate bit by bit, spinning platter by hand, you will see the natural inward force reduced, avoid too much outer pull, you want the stylus tip to play with essentially balanced in/out forces. You rarely get it ’perfect’, just ’generally staying where you place it. Check it also about 1-1/2" in from the label, make the best compromise from 1-1/2" from each edge

3c. IF you had too much anti-skate, the arm would move to the right as you lower the arm down onto the LP. Move your lifter up/dn/up/dn, does the arm move?

4. End of Play Safety Lifter

Check to see if you have room to place this safety lifter. (not enough room: denon and jvc vintage TT’s wide platter edges restrict the space for placement. And your tonearm base will effect available space. Thus you want return privilege.

You adjust it’s placement, height and orientation so trigger wire is just inside of the grooves after play. Arm hits trigger, safety lifter moves up just a bit, enough to get the stylus up/out of the groove, arm stays there. You move arm back to your arm rest. Reset/push down the safety lifter in down/cocked position so the tonearm can pass over it while playing.

Audio-Technica AT6006R Safety Raiser

For the Nth time, a blank LP surface does not replicate the condition that gives rise to the skating force in actual practice, which is friction between stylus and groove walls. The stylus tip cannot engage the blank surface in the same way it engages the groove, because the contact patches will be different. However, since some AS is better than no AS, Elliott’s method is probably about as good as any other, so long as it doesn’t result in too much AS, which is as bad as no AS.