The model has been doing it's usual track flip/flop with the 0Z/12Z vs 6Z/18Z runs - I think this is a product of the age of initialization data. The 0Z/12Z runs have been further right.
In general, though, the trend has been very consistant about developing this but turning north before reaching the islands.
Here are the surface plots of the last 3 runs, valid at 12Z on the 10th:
Note that the following wave is looking interesting as a potential Caribbean threat.
The runs are consistant in turning the system NNE at this point and deepening it - although a later turn back to the NNW seems possible:
(valid 12Z on the 15th)
The following wave shows here south of Cuba.
Jan





