Well that's why as I mentioned this morning, I usually just look at the individual LP centers and do a rough median/mean of those in my head, rather than look at the actual mslp. Kind of an ensemble of positions, rather than mbs.
The 12z FV3 looks almost identical to the 6z GFS at the end of its run...is that a function of the way the model works, or is it giving the GFS support (over the NAM)?