It is hard to understand how an onlooker could see clearly enough in the dark to discern stationary rotor blades, but discerning motion of the whole helicopter is much more plausible. But in the dark what would have been seen most clearly would have been its lights, and they could very easily have given a misleading appreciation of its gyrations. A "tumbling end over end" motion as I understand the journalist believed he saw is hard to explain. But imagine fro a moment a helicopter exhibiting a very high yaw rate, and at a significant bank angle in the dark. In aircraft axes, the helicopter is yawing, but in earth axes that yawing will appear as a "cartwheeling" motion and could easily be described as tumbling end over end, especially in the dark with the observer not realising he was seeing the aircraft in somewhat of a "plan view".
How would a helicopter get in such a condition - well the loss of drive to the fenestron could generate a high yaw (depending on airspeed (yaw stability and yaw damping) and rotor torque (as the engines try to spin the cabin in opposition to the torque applied to the rotor). And the bank angle - various mechanisms - aerodynamic rolling moment form the fuselage at very high sideslip; immediate lateral stick input in response to sudden yaw rate; even perhaps lateral stick input due to the inertia of the stick and hand on it in a very aggressive yaw acceleration. Although entry to autorotation after power loss can be and is trained and practised, the loss of yaw control and resultant sudden yaw rate when the reaction to engine/rotor torque is no longer restrained by a tail rotor or fenestron cannot be, and would be very disorienting, especially in the dark. I have a faint recollection of once seeing video of a helicopter loosing its tail rotor drive and the gyrations were sickening to see. Nevertheless, if the aircraft can be got into autorotation so there is no engine torque being applied to the rotor, there will then be no torque reaction back onto the fuselage and the yaw rate will damp down and - with enough height and a measure of good luck - enough control regained in pitch and roll (though not of course yaw) to have some chance of a survivable arrival. But add in the hazards of darkness and a very obstruction-littered urban environment, and the challenge facing the pilot would be enormous.
Incidentally, an early observer of the fairly-recent San Francisco 777 crash described that aircraft as "having cartwheeled", but all the pundits on PPRuNe immediately dismissed that as impossible and inconsistent with the condition and layout of the wreckage. Some time later video emerged and was posted. It clearly showed the aircraft *yawing* through ~360 degrees but while at a roll attitude of of perhaps 30-40 degrees. The observer's description of "cartwheeling" was not at all an unreasonable description of what the video showed (allowing for the fact he only saw it once, in real time, and was not "primed" for what he was about to witness), but his account was simply interpreted too literally by the pundits and rejected as "impossible". We should not too lightly dismiss eye-witness reports that do not fit our own preconceptions. They may need some careful interpretation (and they may be wrong in some respects - forensic recollection of sudden, rapidly-evolving events is not something humans are good at), but we would be well advised to strive to see if there might be significant insights concealed within them.
The accident at Glasgow was a truly horrific event both for the crew and for those in the bar. God rest the souls of all who perished.