So I agree with you that you could say that "its more likely to be average or cold this winter", if you're adding the odds for both average and colder, and comparing to warmer. But I don't think that's the same as saying "above normal is unlikely".

As near as I can tell if I've got this right, the three categories are Above (A), Normal (N), and Below (B). The default is a 33.33% chance of any of the three. If that holds in the forecast, then that would give a white E rating to a zone, since above, below, and normal have equal chances of occurrence.

Saying the upper midwest is 40-50% likely to be above normal, means they're giving it 33.33% to be around normal, and that leaves a chance of 16.67% - 26.67% of being below normal. So above normal is the

*most likely *outcome, even though there is a chance of normal or even colder temperatures.

FWIW, and I assume you've seen this, but I bet most folks haven't, here's the explanation:

I think the usual interpretation that "NOAA is predicting a warmer than average winter in the upper midwest" is correct based on the map; but it acknowledges the lack of accuracy and difficulty of prediction by including some chance of "colder" even then they predict "warmer".