I remember in 2004-ish, some highly-paid analysts predicted the Nintendo Wii would come in third place with the PS3 dominating the majority market. All they were doing was looking at what had happened in the past and extrapolated it to the future.

And remember a few months ago when there were analysts claiming Nokia had dropped Maemo in favour of Android?

And for literally years before that there were analysts claiming Nokia was going to imminently drop Symbian in favour of Windows Mobile.

Really you have to wonder what these people are paid for. The only predictions most analysts seem to get right are the ones that are so short term and predictable that no one could get them wrong. Anything long term (i.e. anything actually useful) and they just seem to spout any old random cobblers.