MM, my only qualm with your thinking is this: how do you account for the QuikSCAT passes that showed non-rain flagged 35 & 40kt winds within the circulation on multiple occasions? The apperance of the storm was very similar to that of Franklin and Harvey -- among many others in the past -- at times, storms that we know to have been of tropical storm force from recon observations. It certainly did not have the classic look of a lowly-sheared tropical storm, but it had the winds, convection, and circulation to support the TS intensity. There have been many studies done on sheared TCs that show them to be more resilient to shear than originally thought -- and as has borne out this season. Just because there isn't a CDO or persistent region of convection -- e.g. a better satellite apperance -- doesn't mean the winds aren't tropical-storm force. Remember -- the Dvorak technique can be as much as 20kt off in the long-term mean -- even higher in some isolated spots (source: discussions with forecasters from the NHC; NHC seasonal reports; storm discussions). QuikSCAT has some error inherent to the measurement system, but not that large.

Just my two cents.