The Superensemble isn't doing as well this year so far as it did last year, but it never does terribly well early in the season....plus the fact that we've had a good number of storms that probably thus far that probably don't have similar storms over the past few years in the training period for the model to lead to better forecasts. With more storms down the line -- time to better assimilate how the other models have changed for this season and adapt the model to those biases and tendencies -- the model will do progressively better with time.

Remember as well that about half of this season's forecasts come from Dennis, with most of the other half from Arlene and Cindy. With a relatively small sample size, any model that does really well with one storm and average with the others may stand out. The Superensemble has been very close to the ensemble mean most of the season, meaning that the historically best models aren't doing as well as they usually do...but it's still early. NOGAPS is historically one of the better models, having some troubles last year, but with very good consistency along the MS/AL border & timing wise with Dennis, it's starting to make up for that. Only time will tell which one comes out on top.