The pictures posted on GT Politics have been bothering me. I would have expected 21 to have been either inside the gauge or have turned the outside rail. It is possible that the inside wheels dropped first, and the outside wheels "bounced" over the outside rail before it could exert enough force to turn it. There should have been wheel marks of this evidence on the rail/ties. If not, does that mean that there could be one or more factors coming into play, as previously suggested in these posts?
The contractor did a responsible job of regauging the track to correct an inconsistent gauge, especially in the curves. Unfortunately, one of the drawbacks to this work is the necessary disruption of the existing track, and again respiking a well used tie (they did use tie plugs).
This was a "succesful" accident, in the context that there were not any injuries or serious damage.
The laws of physics still apply to any railroad, any gauge, age or operator.