Interesting to see the various GFS-based models shift way back to the right in the 06Z runs. The 12Z runs will be interesting to view. The last SHIPS run is more aggressive with intensity, taking the storm up to 90 knots in 96 hours. The fact that Wilma finally strengthened a little probably has something to do with that... it seems like recent intensity change is one of the inputs to that model, based on how it behaves.

The fact that Wilma is drifting south now, rather than west, could make a big difference in the eventual path. If that persists for any length of time, that means the system will potentially be further removed from any weak westerly steering current that may try to develop to its north in the near-term, so it will be further east when the stronger northerly steering current is expected to develop in a few days. Of course, if it gets far enough south, it may end up getting too far south to be picked up and instead drift around for awhile until likely impacting Central America at some point. I don't know of any models that have a solution along those lines, but when there is quite a bit of model divergence in track, you shouldn't throw out the persistence (extrapolated) track out of hand.