I honestly don’t know if we should really expect a model to nail a winter storm from 6+ days out, given the technology we have today. I think a general pattern, maybe. But winter storms are so fickle around here, that it seems a little outlandish to expect we’re anywhere close to being able to get that kind of detail nailed down that far out.
With the FV3, I’m interested to see if it gets the general H5 look 5-10 days out better than the current GFS and to see what kind of biases it has, if any. So far, my understanding is that its scores are not all that great and that it has a cold bias and a bias of overdoing the strength of highs at longer leads.