John, good point.
The yaw I experienced during a feathered landing with the chief pilot was noticeable, even in a light twin. The biggest thing is the lack of drag - when selected to idle power the windmilling prop would normally be creating a large amount of drag.
The 'opposite' yaw during the flare wasn't huge, but a float could be experienced. I would pre-flight brief the student that during the landing they could reduce power on 'their' engine only - the good one. I had control of the 'dead' engine. I would ensure a small amount of thrust was being generated by the dead engine to simulated feathering.
It was also briefed that if the approach or landing turned to crap, both engines were available to the student. Since the mixture had been returned to rich when setting zero thrust, a go- around setting could quickly be achieved on both engines.
Back to the feathered landing, a review of the ops manual showed intentional feather approaches were not permitted. I wasn't impressed the chief had failed to know this, but to be honest I should have picked it up as well. With all of the risk and stress involved in conducting this sort of approach (for ATC as well) I wasn't keen to do it again.