1000x this. When winter weather doesn't pan out, we often hear "the models suck." Do the models suck in June, when the high is 85 instead of a modeled 82? Or in October, when 0.52" of rain accumulates at 48 degrees, instead the prior day's 18z GFS, which said it would be 0.30" of rain at 51 degrees?
No, no one cares! We care most when the errors matter most. I guess the non-winter corollary is the exact track of tropical cyclones.