Looking at that GFDL model data, I have to be reading something wrong.

If I'm reading it right, it's predicting a motion of 315/1.8kt @ 54hrs, 196/2.5kt @ 60hrs, 42/3.5kt @ 66hrs, 340/2.9kt @ 72hrs, and 85/4.1kt @ 78hrs.

That is a fun little circular loop, isn't it? I'm assuming that's a good portion of the stall pattern they are expecting over the Yucatan Peninsula (and the coords are more to the left edge of the peninsula, correct?)

These models are really having a hard time getting the pattern of motion down. I would hazard a guess and say that this storm is going to be one of the hardest ones this year to nail down a pattern for.