I'm back again with yet another frustrating intermittent EM nuisance bug. I'm getting an erratic double-advance in the sequence unit. This game has 1-9 numbered bagatelle rollovers and 3 alternating top PF rollovers, each of which when lit and rolled over, advance the sequence unit by one digit and award a bonus. But in every game, one or two of them "bump" the sequence up an extra digit. It's rarely the same rollovers that do this. I've checked/rechecked/cleaned/tested the sequence unit several times, with no effect on the issue. I also checked the "E" Advance Seq. relay which I cleaned and gapped properly, as well as all 12 rollover stack switches in the sequence. I'm thinking a maladjusted motor switch might be the culprit, but I can't find anything amiss. This is a 3-ball game, and all other functions work okay. As always, any ideas or advice are greatly appreciated.

I actually don't know for sure, but I will check!

Answer is yes: the Advance Sequence unit activates twice/moves 2 positions when this happens.

Thanks for your observation, Rolf. I have checked relays C, D & G also for proper gap and operation, but I will closely inspect them once again. Let me update you further. I have analyzed every score motor switch operation in every radii of turn, with special attention on the switches you mentioned. Yes, I agree these were the likely suspects, but I did confirm they all gap and close correctly on every cam post and dwell position.
Your notion of residual magnetism is intriguing. I will certainly perform the procedure you mentioned on all the noted relay armatures. (I LOVE banging things with a hammer).
To answer your other question: the "sometimes" extra sequence advance fault ONLY OCCURS when the ball makes contact with one of the 12 rollover "advance sequence" switches. This is why I suspect possible faulty operation of the advance sequence stepper unit itself. I'm not happy with the manual operation of it and it's hard to say what truly happens when it's energized. So... I need to take it all apart, clean out the sludge, check/repair wear parts and replace the coil and 5 springs with new, then run test games. We tend to look to the schematic for electrical solutions when sometimes the issue is pure mechanical. I must rule this device out first.
Regarding your advice about using insulating tape on relay armature plates -- I do not do this. On one of my Ballys, I discovered a previous owner used electrical tape on one armature and it had melted from heat and vibration over time which caused even worse sticking problems! So if I can't bang out the magnetism, I'll replace the armature! Regards.

Update: been traveling (Australia & New Zealand) for some time and am now back over jet lag. I disassembled and checked the E, C, D, G, R relays for residual magnetism. None existed. Ordered stepper springs for sequence unit and installed them. Also disassembled and cleaned entire sequence unit itself. The unit works perfectly, except that the same sometimes "sequence skip" issue still existed. I spent hours looking for a visual flaw. Then, I discovered a hidden solder tab on the #2 rollover switch that was bent close enough to its other tab and probably causing an intermittent arc. I corrected it and checked every other solder tab under the playfield. Then I fired up the machine. The sometimes fault was gone. Sheeez do I feel dumb. Thanks to everyone for the help and concern. I hope someone can learn from this one.

But now I have another issue with this Atlantis that has me stumped again. But I think this will can be isolated by an experienced schematic reader. I've generated a new thread for it., entitled "Atlantis special won't add free game."