"Tony Cox" wrote in message
ps.com...

I claim to understand the _modeling_ process, not the
details of their models. Modeling any phenomena is a
complex business. Models that are over-parameterized can
fit just about anything. Models that extrapolate, especially
time series phenomena, are doubly suspect. Models that
deliberately choose sub-sets of data to fit and ignore the
entire data available to them better have a good justification
if they're not to be dismissed as a put-up job.

I'd say the GW models are over parameterized, attempt
ambitious extrapolation, and have a track record of
dismissing data that they can't account for (like that in
the medieval warming period).

In all the GW models, none, zero, zilch, nada... when run in reverse even
remotely models the present or past.

That's a standard test of a model.

Not one GW "scientist" has ever done a public test in that manner.

One dude, though, threw a fit when asked to show his data nad methods, that
was paid for by the taxpayer. Oddly, THAT form of politicalization is never
mentioned in the MSM.