Well, I think that that settles the question of whether NIST needed to consider the stiffeners when estimating the walk-off distance: the girder would fail anyway once its CoG was off the seat, regardless of the stiffeners. NIST probably knew that and didn't need to consider the stiffeners. Omitting them was a reasonable simplification of the simulation.
The column (and thus the seat) displacing to the east settles the question of whether the beams expanded by the length of the walk-off: they didn't need to, because the displacement of the seat added to the total distance for walk-off. And that distance was large enough for breaking connections in three other floors.
The expansion of the beams before the girder settles the question of why the girder wasn't trapped between the plates.
So far nothing changes in NIST's analysis even when all these factors are considered.
So, the only issue remaining from gerrycan's list is the beam stubs. What about them, gerry?
The column (and thus the seat) displacing to the east settles the question of whether the beams expanded by the length of the walk-off: they didn't need to, because the displacement of the seat added to the total distance for walk-off. And that distance was large enough for breaking connections in three other floors.
The expansion of the beams before the girder settles the question of why the girder wasn't trapped between the plates.
So far nothing changes in NIST's analysis even when all these factors are considered.
So, the only issue remaining from gerrycan's list is the beam stubs. What about them, gerry?