The problem is that the jet stream has been behaving differently over the last 8 years or so to how it had behaved for the preceeding decades. So at first, the forecasters had no models on which to base predictions. And, unless the Jetstream is where it is expected to be for the time of year, the models they are using now have only data from the last few years, so the forecasts are of course less accurate.
And actually, reports of actual conditions from airline and other pilots and from aerodromes and weather stations are an important part of the information used by meterorologists
I can recall the day they announced that the latest Cray supercomputer could now calculate the weather for 24 hours ahead with complete accuracy. The only problem was, it took 25 hours to do so. All my pals found this hysterically funny; I was the lone voice saying, "You're missing the point. The point is that they now have enough information in the models to do the calculation and get an accurate result. Faster computing will come."
I expect that the Cray "supercomputer" which had produced this accurate, but late, forecast was about as powerful as the laptop I am using now!
(And it was 1980 or 1981, if you were interested. I know because I remember where I was working at the time, and I was there from Sept 1980 to Jan 1982.)
For making hay, we need a four or five day accurate forecast. These only got really accurate towards the end of the 20th century. The jet stream started misbehaving around 2006/7, I think - so actually, most farmers have been making hay according to their own weather sense for most of the last x00 years... ;p
