BINGO...it never does stop.
I think alot of the early looks ahead to winter that early (heck probably even going into late summer) stem with those involved in energy demand/forecasting.
But I have to agree...at least from the minimal work I've done with attempting long-range forecasting there is not a whole heck of correlation between fall/winter. Significant changes happen to the northern hemisphere during this period from changing wavelengths to changing correlations between teleconnectors and the overall pattern. Also, you can have significant short-term weather phenomena which can result in a significant hemispheric pattern change. For example, perhaps a significant recurving typhoon in the PAC or a monster ATL storm which completely re-shapes the NAO or a significant stratospheric event.