VicMel
ANDRASZ's answer covers this matter as being remote. My own flight testing on the -500, we were required to do baseline stalls of the aircraft at its limit cases, fore, and aft, at high and low weights. Those we had to meet by having a couple of tons of drinking water from Sams Club and moving from aft hold to forward hold, we also had movable ballast inside the aircraft so that the TO and LDG were done within the basic envelope of the aircraft. longitudinal stability was adequate at all times. The only notable curiosity in the plane was a clean stall where the slat rigging could result in a minor amount of roll but it was a non-event.
On the autothrottle musings, the info available so far doesn't match well to a thrust change occurring leading to a yaw/roll/APLT disc/uncommanded open-loop roll departure. it is a simple scenario, but with a relatively steady state before the upset, a thrust change is not evident. That may be the data granularity, or the scenario doesn't fit. It is unlikely that the asymmetry was there for minutes undetected, so a change in rates immediately before the upset would be expected in the ATR clutch scenario, and that doesn't seem to be there.