Over the past 48 hours, the medium-range guidance has done for the 4/1-4/3 period almost exactly what it did for the 3/21-3/23 period at this range: strongly advertised a major pattern change with significant potential for severe weather, then backpedaled dramatically to the point that today's 12z runs look like more of the same. Predictability in that 7-10 day range seems to be even worse than usual so far this month. From experience, it's not that common to see complete flip-flops on the longwave pattern over the CONUS inside Day 10 (when looking at multiple models and their ensembles); that's more what you'd expect at Day 12-16.
In any event, it seems that every single wave has amplified while passing over eastern North America for weeks, and this destroys any chance of a favorable pattern developing. Every time the medium-range models suggest a pattern change, they do so by allowing a wave to finally pass off the east coast without amplifying (at least too early and/or strongly), but then trend toward more amplification as new runs come in. Note the H5 height field differences between the 192 hr. panel from today's 12z runs vs. the 216 hr. panel on yesterday's 12z runs for both the GFS and ECMWF. Even though it's still over a week out and a reversal of the trend is possible, persistence certainly favors the newer solutions.
