Wouldnt put my guard down NW of the fall line (ie @Qtown Snow), despite some sobering (for weenies like me) output from key models.
Historically, our more significant ice storms occur with less than ideal timing and placement of surface features.
We have an extension of the primary high pressure, located north of the UP of Michigan, starting in a good spot for cold air draining in Quebec, but retreating NE as the low approaches. This develops some CAD, however, it's short time in a good position limits the depth of the cold pool, as noted by dew points in the teens and twenties, rising as the high retreats and low approaches
Many ice events have primary lows riding up into the WV area. As the primary rides up the western side of the Apps, and warm mid-level air rides over the Apps, and the shallow pool of cold air to the east, we typically see an inverted trough form along the coast. A difference between the colder NAM and warmer Euro are the depiction of the inverted trough placement.
The EURO essentially indicates the CAD is so shallow/weak it's not feeling enough of an influence, so the cold is scoured more readily, thus reflecting a farther north inverted trough/baroclinic zone (where warmer ocean air meets CAD). This has merit as the cold pool was not well established.
NAM on the other hand holds onto the CAD longer and shows a more pronounced inverted trough and farther south baroclinic zone. To the south of the inverted trough are warm southerly surface winds. To the north are cold NE surface winds, that typically lock in colder temperatures. See the difference in the inverted trough plavement and effevt on surface wind direction:
In these marginal setups, PHL typically turns to rain. Far NW locks in cold longer than expected. Im not sharing anything new here, I wanted to highlight the importance of the inverted trough placement. If NAM thermals are closer to reality, that NE surface flow will bring quite a mess, here's a sounding at the height of the precip to reflect it