That is a thought I had too.
But something is going on, I mean we can even extend what seems to be poor skill back 3-4 years now, especially in winter. There have been some extremely large jumps within guidance inside of 72 hours and it seems just about every major model has had its share of failures.
Part of me wonders if its just related to how complex the weather patterns have been or if its more human related. There is just so much garbage thrown around on social media for storm potentials several days out (sometimes even 10+) and there is so much focus on just one or two graphics, and forecasts start getting chucked out 4 days in advance.
There is too much with taking a model at face value and very little critical thinking or skill applied. Its just take the model which shows what the user wants and then try to spin it into a forecast. Then when the forecast doesn't pan out, the models get blamed.
At the TriState Weather Conference in October, a person asked a question relating why it seems like models were doing worse. I believe it was Dr. Tuell who responded, but he essentially said one problem is there is lack of responsibility taken and models are essentially used as a crutch as to why a forecast is wrong. I wish his answer was recorded, I thought it was a great statement.