The models are struggling to predict a solution out 4-5 days. What I look for are the big global trends (i.e. mainly the position and strength of the Atlantic Ridge or if there are any mid-latitude troughs predicted to drop down). I posted earlier about the building ridge after 96-120 hours (4-5 days). If this verifies, and its a big if still 5 days out, then the storm is either going to head west if its low enough in latitude or it'll get trapped within the ridge - in which case it'll stall and meander until some big change happens like the ridge moving or weakening. i try not to focus on the individual model runs but look for trends over several days. The global models like GFS still have problems 4-5 days out - in this case, the NCEP ENS means (lots of runs averaged together) indicate building ridging after 4-5 days - how this translates to track is all a question of timing. The newest NOGAPS builds the ridge like the GFS & UKMET, but it races Irene fast and beats the building ridge.