Have experienced the same as you Albatross while sitting on the end of a longline for extended periods holding a load in position with no change in power.
Normally did if by default anyway but used to move the machine around to stop VRS/settling from developing as we never had the luxury of height for the recovery. ( normally moving around s bit anyway)
I wonder if a prolonged stationary hover gave the induced flow a chance to accelerate to a higher level than normally would happen and then a little decent would start due to reduced A of A, a small pitch pull to hold the height and then VRS.
Essentially what Nick said, settling first then VRS.