Models generally have had a cold bias for a good number of years, unfortunately. All levels are affected including surface and 500 mb. That’s why there’s often (though not always of course) a NW trend in the track of storms as you get closer. They often start off in the modeling as southern sliders and become inland runners as you get closer. These cases were destined to be inland runners from the start but the models weren’t smart enough to know far in advance.
Hopefully we’re now close enough so that there won’t be more NW trending for this one.
In Atlanta, the coldest days’ lows were projected to be well down into the teens awhile back. Now they’ve warmed there ~10 degrees. Still a solidly cold period with lows well down into the 20s (MB normal) and cold enough for wintry but not nearly as cold as earlier projected meaning less room to play with.