Your question was how I "think the collapse should have progressed," and I answered that question directly. You hadn't asked me to recount my reasoning for that opinion before now, and it's weird that you imagine you did. As for how I arrived that that answer, I'm not going to bother detailing it, as I'm not claiming to have done a thorough enough analysis to have proven as much, but I did answer your original question of how I "think the collapse should have progressed."
Probably the most unrealistic assumption in their model is the way they applied all their assumed temperature and damage, as can be found in
NCSTAR 1-9, 12.3.2. I'd quote if it wasn't for NIST locking their PDFs, but here is a graph from that section detailing as much which I'd found uploaded elsewhere:
[qimg]http://img44.imageshack.us/img44/9919/43690010.jpg[/qimg]
Put simply, they took their estimate of what happened to the building over the course of hours and applied it in a matter of seconds, with all the fire damage kicking in at once just to get the model to come down as quickly as it did in what little they've shown of it actually collapsing.