I think, though, that a mixture adjustment cannot substitute well for a worn-out pot. Since that adjustment is a calibration for the mechanical action of the airflow meter on the fuel distributor, it should need to be made if the airflow meter body (i.e. the MB part) is replaced, or it there has been enough mechanical wear. The procedures described on this site involve measuring the EHA current to determine how much it is compensating for an incorrect mixture.
I was surprised by how easy replacing the worn pot was. I was careful, and pre-adjusted the new one to be as close as identical as my DVM could resolve. That done, it was plug-and-play. No mixture adjustment should be needed.
I don't know why the bad pot was causing the idle speed fluctuation and decel surging, but I have a theory:
[WAG] At steady throttle, and demand of the engine is essentially constant, the pot sensor return to the ECU is a fluctuating voltage where the wiper hits the worn spot. The ECU interprets this as 'leadfoot' and tells the EHA to go rich. Since this wasn't actual demand, the Lambda sensor tells the ECU 'very rich', and since we are signalling 'throttle closed' too, it compensates by opening the idle control valve more. [end WAG]
I stress - I really have no source that tells me this is what happens, but it seems reasonable for the symptoms.