CM: I hear you.
Modifying the TO waypoint does mean
- not only doing a DCT-to some other place (2 keystrokes with (L)NAV still engaged)
- but also adding a intermediate new waypoint along the active leg.
For the latter, which you brought up, some of the modifications require pulling HDG, otherwise FMS refuses accept the change with NAV mode engaged.
Use cases? 20 NM before IDESI, 10 MIN before VESAR from the top of my head.
I would understand the "wiggle" to get back to the center line, if the green line moved away from actual PPOS during the reprogramming, but that is somewhat unlikely. At the same time, I do not recall such behaviour (to the extent I would be displeased enough to rembeber).
What FMS verison are we talking about?
Is the wiggle always towards the same side? (*)
Are you using onside FMS and onside AP (the normal way: FMS2 + AP2 and v.v.) (**)
* = if it is just a mathematical inprecision, the behaviour should be reasonably random, i.e. 1/3 left, 1/3 right and 1/3 nothing
** = presumably programming via FMS 2 with AP1 engaged could create a tiny discrepancy that would need to be compensated for?