

I measured the model and the layered original video (same size, angle etc.) separately just to see what I get. I measured the fire and antenna of the video and the model 5 times each, averaged the 5 results (thin curves) and additionally averaged the results over 9 frames (fat curves). That's what I got:

Large:
http://img163.imageshack.us/img163/3738/graphinitialdrop.png
As you can see (e.g. the "stairs" in the thin pink curve) the failure due to pixel edges corresponds to about 0.06m.
At about frame 180 the top floors started to tilt and at about frame 230 the north face (the hinge) started to descent. Considering the relations the first sagging of the "real fire" curve (dark pink) is probably just caused by a flaming up of the fire in the first moments of movement.
I wouldn't interpret that fist sagging of the fire as a real movement that stopped for about 50 frames. Imo that fire curve is pretty close to the model and therefore to Case3 - " the immediately response of the south face to a vertical core collapse. The only difference to the "virtual" case - the model - is that the real antenna appears to be a tiny little bit lower than the model. In other words, the real antenna either dropped a little bit earlier or a little bit faster for a short period of time.
...or should I say, compared to the model the south face / SW corner indeed responded 5 frames later? How ever, there is NO CLUE of a perimeter collapse that overloaded the core. There is NO CLUE of a row by row core collapse. There is NO CLUE of any core resistance e.g. causing a convex roof deformation or just a delayed descent. There is NO CLUE that the tilting of the top caused any damage at all. Instead it appears that the core went straight down, overloaded the south perimeter, the bowed south perimeter responded almost immediately and collapsed, the perimeter collapse in the south caused a trapeze deformation of the outer shell (but not of the core = no additional hat truss tilting = the core was already at the same or a lower elevation compared to the SW corner), that deformation of the outer shell overloaded the corners and the east and west face, the south half of the west face bowed inwards between 92 and 98, the top started to tilt towards south, at a tilt angle of about 0.22° the north face started to descent, at the very same time and due to tilt and deformation of the outer shell the south side already dropped for about 0.9m ... global collapse.
Go to NIST and get your money back!







