Kay wrote:
In article , shazzbat
writes


I read somewhere that if you forecast every day that tomorrows
weather would be the same as todays, you would have a success rate
comparable with the forecasts on TV.

Not quite. The probability of tomorrow's weather being the same as
today's is in the region of 67%. This means the forecasters have
the
challenging task of improving and getting the forecast into the
rather
narrow gap between 67% and 100% - ie, don't be too scathing, it's a
difficult task they have there. It's always easy to improve
something
if it's really bad to start with, less easy to improve something
which is already rather good.

What interests me about these discussions is people's expectations.
The daily weather forecasts are indeed pretty good, and even the
weekly ones aren't a waste of time. But the long-range jobs are at
the cutting (or should that be "blunt"?) edge of science: they ought,
perhaps, to be more often and more explicitly presented as the rather
exciting experiment they are. Recent TV brings to my mind the way in
which the Soviets kept quiet about their space efforts till they were
successful, while the Americans -- for equally good reasons --
carried theirs out in front of the world's press.

--
Mike.