Yeah. Engine running at impact sorta removes the obvious cause. Although running and developing useful thrust are two different things. A prop going flat or feather would leave the engine running, but developing negligible thrust. As would the engine rolling back to near idle due to some engine control malfunction.
If they care to, NTSB can probably derive something close to actual airspeed from either ADS-B or video analysis. Which will rule in or out a simple stall spin due to inattention after engine partial failure or airspeed indication malfunctions. Or just inattention: pilot drops something in the cockpit, or gets focused on something on the ground and suddenly he’s on his back. Or distracted by some other malfunction like an electrical failure (or fire!) that would leave rather little physical evidence in such a hefty post-crash fire.
After that you’re sorta left with some manner of control jam, perhaps from somebody in the copilot’s seat. But that doesn’t seem to have been occupied.
I noted they put more fuel in right than left tank. Suggesting they’d burned from the right more than the left on earlier flights. In any case, the side taking the heavier fuel onload was the wing that went up, not down, as they lost control. So if there was a fuel imbalance favoring left roll, it would have been worse before refueling than after. IOW worse at the previous landing than it was at the mishap takeoff. And obviously the previous flight landed with no excitement.
Hmm. Not straightforward at all.