Friday, October 29, 2010

More Curried leftovers

Bit of a backlog of blogging to do, but now I'm back in Japan with hardly anyone to talk to, I've got no excuse...

First up, I noticed Judith Curry continuing down her bizarre rabbit-hole. Luckily mt got there first, and I don't have much to add except my broad support for what he has said. Note that in the very first premise of her argument, she only assigns 70% probability to the fact that surface temperatures actually show a warming at all! This is the warming that the IPCC famously called "unequivocal" in their 2007 report. As far as I can tell, at this point she is simply so far out of touch with mainstream climate science that her analyses aren't worth the time it takes to read them. End of story.

If you want more detail, then yes, I agree with mt that her approach to probability is pretty dodgy too. I consider myself reasonably ecumenical in my approach towards the more esoteric probabilistic ideas such as Dempster-Schafer theory and imprecise probability, and have no real objection to them - I mostly take the view that we should merely try to do standard Bayesian probability a bit better before deciding it is inadequate for the task at hand. However, I don't think the "Italian flag" analysis - at least JC's interpretation of it - is a useful or even coherent contribution. The rot sets in right at the outset, where she apparently conflates the concept of evidence for and against the proposition "most of the observed warming was very likely due to the GHG increase" with an estimate of the proportion of warming that was due to anthropogenic vs natural factors. This seems like a rather elementary point to get confused over. Far from being the claimed synthesis of much detailed thoughts regarding the failings of the IPCC, most of what she has written reads to me like a stream-of-consciousness blog post that hasn't been properly thought through at all. But hey, "very not the IPCC" is all it takes for a stream of admirers and press attention, irrespective of whether there is any there, there.

It's not as if I'm the IPCC's greatest fan, either. But I try to base my criticisms on actual failings, rather than just bandying about terms like "corruption" and hoping that something might stick. Since I've been officially threatened with increased ostracism if I dare to say anything nasty about them in public I won't bother re-hashing any of that again. At least not right now :-)

9 comments:

Dikran Marsupial said...

The problem with the Itallian Flag argument seems to me to be that it doesn't reflect the fact that the "evidence for" or "evidence against" the hypothesis in probabilistic arguments is never unequivocal unless you are talking about something that is a logically certainty or logically impossible. Thus in reality the red and green parts of the italian flag ought to be infinitessimally narrow, and the flag should be all white (insert your own joke about military history here ;o) reflecting the fact that all of the evidence is uncertain, at least to some degree.

A more sensible aproach would be to have a flag with a gradual transition from red to green, reflecting the plausibility of different degrees of certainty regarding the truth of the hypothesis. In other words, a Bayesian posterior distribution.

EliRabett said...

Ain't no damn Rabbit hole, more likely an attic. The motivation for her crazy aunt act escapes Eli.

James Annan said...

Well it looks like she did a moonbat when the CRUtape letters came out, and didn't have the self-awareness to row back when it became clear that she called it dead wrong. Of course she was already trending in that direction anyway...

David B. Benson said...

Whatever.

Steve L said...

I don't get the reference to ostracism. I think I get everything else, though, and I don't know why I don't visit more regularly. Good post.

Steve Bloom said...

Er, yes, "threatened with increased ostracism"? Have I missed something?

Anyway, much though I'd like to hear all the dirt on the Ulan Bator's roads, public transit system and flight connections, do they even have a climate research center in Mongolia? :)

But if you mean *scientific* rather than geographic ostracism, hasn't it been made clear already that the powers-that-be have decided to adopt your ideas while giving you no credit whatsoever? :) :)

Brian said...

I'm with the Steves on ostracism. Any chance for more info?

James Annan said...

There's really nothing to add beyond what I said. At least, nothing I can add without risking increased ostracism :-)

Though it did occur to me to wonder what they think they can actually do, beyond not allowing me to attend meetings where I actually would have had something useful and original to say...(which seems to have been what precipitated the threat).

Steve Bloom said...

Get back to campaigning, you. Middle of ther night? So what! If you have a spare moment you can always go out and, um, suitably relocate some of your oppnent's campaign signs. Anyway, nose to the grindstone! Hup, hup! You can sleep Wednesday.

(Good luck, Brian!)