Let me point something out: NHC does not predict Cat 4 and 5 generally. Why? Becuase the SHIPS model is a statistical model based from a decade long survey, and there were so few Cat 4 and 5 storms during that time that, simply speaking, they don't know what makes a system go from Cat 3 to Cat 4 or Cat 4 to Cat 5, or retain Cat 5 strength. Generally, if you see any prediction of Cat 3 or Cat 4 from NHC, you should immediately say "It's a major hurricane that could be anything from Cat 3 through Cat 5, depending on exact conditions."

GFDL model does a better job predicting strong storms, but it isn't consistant - one run makes a TS a Cat 5 in 24 hours and the next run dissipates it. How can you make an educated guess as to what a system is going to do when the models are so erratic?

The time averaged GFDL intensity is generally about the same as the SHIPS intensity, and both poorly predict anything over Category 3 or weak Category 4 storms.

You can't blame the NHC - they only work with the best models out there. Storms have always had a mind of their own - the intensity is purely a guess. Ever taken a look at the intensity charts NHC puts out? You'll notice that most systems that reach Cat 2 have a chance of hitting the Cat 4+ range, though generally fairly low. It's just like track: don't just follow the centerline.

--RC