Mr. Spock -- I'll concur with Jamie's reply a few posts up from here on his comments to your post. I jut don't agree with the GFS solution myself, either.

What we are seeing over the east central US right now is called trough fracture. Essentially, three days ago, there was a very sharp trough located across the region, moving towards the subtropical ridge. However, the recurvature of Gaston and track of Frances served to enhance the ridge (in addition to an amplified pattern leading to ridge enhancement as well). You can see historically a number of cases where recurving tropical cyclones strengthen the ridge beneath them for some time.

Now, that trough has split into two pieces, as it could not make it through the ridge. The northern portion is going the way of Gaston and making little or no southward progress. The southern portion, with a weak ULL in Alabama/Mississippi, is moving every so slowly northward around the periphery of the ridge as well. There is no further energy to drive into this ridge until the system in the Pacific NW makes it across the country...and it is at a fairly high latitude now as it is.

A strong storm like Frances is going to gradually turn a bit to the north with time -- Coriolis and other dynamical effects say it will -- but I'm not sure I buy it slamming into the ridge and stalling like it has forecasted to date. What worries me is that the GFS has done the best of all models through yesterday with the 120hr forecast, with errors between 100-150mi. It'll either no longer be there in a few days, or it will have solidly beat the other models.

Note about the Eta -- the storm only recently has entered it's grid domain (it only covers part of the Atlantic). It gets its boundary conditions (what is going on at the edge of the domain) from the GFS. Thus, until just the past couple of runs, it has been heavily influenced by the GFS model output -- as have a number of the other models. We'll see how long that lasts.