It would be interesting for someone to do a case study on the accuracy of model forecasts for these shortwaves that originate in areas with lower amounts of sample data. My totally half-assed ignorant theory is that as the model resolution has increased over the last few years, paired with the faster flow we've had, it causes more run-to-run inconsistencies on the operational models for these types of storms. If the flow was a little slower, there would be time of the physics to catch up to the shortwaves as they develop and provide a more consistent forecast. We often do not get consensus on the info we need to determine snowfall amounts until a day or two before because many of the storms that have been happening over the last few years have been fast movers. Not a lot of systems stalling and/or phasing, and when they do, they are usually modeled better.
The older lower resolution models may have been less sensitive to this problem, though may have not generated a more accurate forecasts, just one that was more consistently wrong until go time. And as usual, we are often talking about differences in tracks of 50 or 100 miles to determine snow amounts, which is more difficult to pin down compared to just plain rain during other parts of the year.