Peter O'Kelly boldly predicts there will be fewer bold predictions from analysts in the future. This reminded me of a BYTE article I was working on a decade ago. It was an update on Windows NT, I think. In any event, I re-read the original BYTE article and it included an analyst firm's prediction about NT penetration after three years. Being a good editor, I went back to the analyst firm and got their current assessment of its penetration. The prediction and reality were not exactly in alignment, to put it politely. Not even close, actually. Kind of off by an order of magnitude if I remember right.
Needless to say, I was told by management not to embarrass the analyst firm by publishing the graphs side by side.