"Col" wrote in message
...
Richard Dixon wrote:
On Feb 15, 10:11 am, Dave Cornwell wrote:

Nevertheless real forecasters don't have the luxury of missing out
days that are "difficult" as there is an expectation upon them,
hence the occasional poor forecasts.

I've tried saying this in the past. His statistics are not a like-for-
like comparison with any forecasting house statistics. Find me a
forecasting centre that forecasts when it feels like and maybe then a
comparison is worthwhile !! Dressing up a small proportion of
forecasting days when the wind's in the right direction as having 70+%
accuracy is entirely misleading, as far as I'm concerned - heck maybe
I'm wrong.

Nothing that Dawlish hasn't always been open about of course.
What would be interesting though is not so much the % accuracy but
the % number of days the forecast is attempted. And that can't
be very many, perhaps once every 1-2 months by my reckoning.
Keep maintaining that accuracy rate, whilst increasing the frequency
of forecasts would be more impressive.

If you don't like calling it a forecast then it's an indication of
inter-model correlation at 10 days. Interesting as an excercise
but a world away from being a 'real' forecaster who has to take
a judgment call on the tricky stuff as well. But I'm sure Dawlish
has always acknowledged that.
--
Col

Bolton, Lancashire
160m asl

Col, knowing how models behave and how NWP works, I'd still say that Dawlish
is picking out times when analysis differences are small in areas critical
for the UK. In the interests of research Dawlish could have an interesting
database of times when this has occurred. But unfortunately he does not seem
interested in that, otherwise he would supply links to all his forecasts and
times and possibly write a paper in a respected journal! he must have kept
all his forecasts in a private note - surely to goodness?

Will
--