It doesn't really change anything because you still have to start your database with the fact that your first storm was in May. If the next storm is in June or August, you only have 5 historical events to work with - and 4 if the next storm is in July. Those numbers are probably way too small to be considered a useful statistical sample. Even the 14 events in May is a mighty small sample. Of far more interest to me was that no matter how small the sample, if a storm had developed in May, later on in the season the U.S. mainland had a tropical storm or hurricane landfall (so far). However, all of the numbers are there in the main blog, so you can develop the new percentages yourself if you wish, but the 'landfall' reality would still be there from a climatological point of view.
Cheers,
ED