2011-05-03 12:36:21Critique of basic rebuttal of 'models are unreliable'
John Cook


Got this email:

I've found an overstatement or oversimplification re climate models, though, which needs fixing:
"If they get the past right, there is no reason to think their predictions would be wrong." - http://www.skepticalscience.com/argument.php?a=15&p=8 & http://www.skepticalscience.com/print.php?r=21

Not sure that it's actually, technically, in error, but it ignores the problem of overfitting - "when a statistical model describes random error or noise instead of the underlying relationship. Overfitting generally occurs when a model is excessively complex, such as having too many parameters relative to the number of observations. A model which has been overfit will generally have poor predictive performance, as it can exaggerate minor fluctuations in the data." - Wikipedia.

I would hope that climate models in use today are robust and checked for overfitting. But the pages linked above are currently inaccurate. Google the sentence I've quoted and you'll see doubts expressed about the claim - and as it stands, it's a weak point waiting to be exploited by a clever opponent.

A fair criticism. Anyone have any thoughts on how Graham's rebuttal could be worded more appropriately?

2011-05-03 20:53:06Old rebuttals
James Wight


Yes, I think there are quite a few old rebuttals that could do with revising.

There are even a few that consist of only a one-liner – eg. “Extreme weather isn’t caused by global warming”.

2011-05-09 09:16:26


GCM are not statistical models and global temperature is not the only quantity that needs to be reproduced.Tuning a GCM byt varying some parameters is different from fitting.