Bladecrack: point taken on reports of the helicopter falling, but please consider that at 2230 hours or so, when this event took place, the eyewitnesses had what visual cues upon which to base the reported observations?
What the condition of the rotor blades implies has been previously covered.
The matter of inertia: if in a hover, or flying at a selected transit / cruise speed, rotors at flight RPM don't just stop on a dime. Even low inertia heads have a crap load of momentum at flight RPM. So the event breaks down into four rough phases:
a. Flying along/hovering along/doing mission
b. Something happens
c. Descent begins
d. Descent ends.
(* no evidence so far available that between B and C an ascent (short duration?) can be added to this sequence)
Inside each phase a variety of things can be happening, which will in all likelihood be pieced together in sequence by the investigators as they apply their skills and resources.
The transition from phase a to phase c took ... how long?
Unknown by the observers on the ground. From the post here by scooby, it appears to have been in phase c when the observer on the parking lot roof picked up his visual cues, which he then lost before the impact with the pub. For how many seconds, in the dark, did he track the object in the sky?
(As an aside, I once took my wife to watch helicopters take off and land at a training field. As I was paying attention to another aircraft, she gave a little gasp and said" It's falling!"
I looked up and saw a CH-53E doing a practice autorotation. No big deal.
To the untrained eye, it looked like it was falling ... and this was in daylight.)