I get a little annoyed sometimes when people go overboard bashing weathermen. They generally do better than they get credit for. Part of that is a difference in criteria: for example, if you get everything about a storm correct except the track is 25 miles off, I count that as a win. But the people with the outdoor wedding that gets rained on count it as a fail.
I agree with this. The improved accuracy in Blizzard of Oz' chart (5-day is as good now as 3-day was 20 years ago) does not surprise me. Usually the errors for skiing are in the nature of "25 miles off the storm track," or "the storm hit 6 hours earlier/later than predicted." I remember the "fails" like Mammoth Dec. 2005 that was supposed to be 2 inches and wound up 37 inches, but those cases are fairly rare.
I also remember that the extremely rare storm track of Sandy, taking the abrupt turn west into New Jersey from the Atlantic, was accurately called by the weather forecasters a few days ahead.
The Slater article referenced by Blizzard of Oz is very consistent with what I see in my snow data.
But, to me, the thing that stands out is that temperature forecasts are somewhat reliable, whereas precipitation is not. And frankly, I've felt that way about forecasts "the day of" for a while. So, let's look at that. If temps are going to be warm, then no matter how you slice it, high precipitation or low precipitation, it's not good.
This is a regional issue. For Larry Schick forecasting in Washington State, temperature is a key factor. Also for anywhere in the East. For the entire Rockies from Bozeman south (also at Mammoth) it's nearly irrelevant because it rains in the ski areas in winter once in a blue moon. Precipitation is really all that matters most of the time to a Colorado or Utah ski forecaster. For the interior Northwest (sibhusky's region) and Tahoe temperature is somewhat important, as the very warmest events will push the rain snow line into the ski terrain.