Just for discussion purposes, I have platforms where throttle control is completely different from consumer drones. Bottom stick = low idle RPM, mid stick = hover, anything above middle is climb, with rate dependent upon aircraft weight.
Previous experience has shown that reduction of throttle from a high hover would permit the aircraft to descend at a speed commensurate with throttle reduction, which could be quite fast. The more throttle was reduced the more “wobbly” the aircraft became. However, with enough reduction of throttle the aircraft would enter a momentary unstabilized free fall before rolling inverted into an uncontrolled descent. If allowed to continue a crash was assured. If enough altitude was present, resuming full throttle would cause the FC to right the aircraft and resume controlled flight.
Those experiments are why I don’t believe multirotors experience actual VRS. What I know happens is that not enough current is being provided to the motors to effectively stabilize the aircraft, with the “wobbles” caused by motors provided only enough current to attempt stabilization, but lacking adequate current the motors induce destabilization as they cannot generate enough thrust to compensate for the weight of the aircraft. As one corner drops it is provided current to lift it back up but a shortage of total available current causes opposing motors to end up in a current shunt, starting the cycle all over again. The FC ends up fighting itself in see saw trying to provide current it is being denied through throttle position.
That’s my story and I’m sticking to it