Ladies and gents, I offer a nomination to the myth-makers hall of shame:  Dr. Judith Curry.

I took apart a paper she published in June in this potential blog post:  Mythic reasoning about uncertainty.

Let me know what you think.

Oh cripes, so Curry fancies herself a psychology expert now?  Unbelievable.

I think it might be a good idea in the intro to emphasize this - that it's much more of a psychology paper than a climate science paper (and that Curry has no expertise in psychology).

The section that begins "We find instead that this 'invisible hand' pushes those who self-identify as 'skeptics' to become 'deniers'" - it might be better to phrase it like "Curry's comments are a perfect description of denial..." so that it's not quite so aggressive.  And rather than "deniers do this and that...", "denial causes people to do this and that".  Or where being specific about things deniers do, I'd specify "climate deniers" so that at least they can't whine about being compared to Holocaust deniers.

I went in and made some formatting tweaks.  Overall, I think it might help to tone it down a bit.  Try to be a little less directly critical of Curry and deniers, and stick a bit more to describing what Curry's done and what's wrong with it.  I certainly understand the frustration with Curry, playing the role of condescending psychologist, while projecting her own faults onto the entire climate science community.  But at the same time, we think our posts are more effective when we limit the emotional content and stick to the facts.

I think there's a special issue coming up on uncertainties in Climatic Change, there's plenty of other uncertainty papers in their preprint queue (including a paper by Pielke). There you might find some different approaches (than Curry's) to the uncertainty issue to use as examples of better analysis of uncertainties.

ANY POST THAT QUOTES FROM JABBERWOCKY IS A WINNER!!!

Roger shows in his blog article that he is either dishonest or doesn't understand uncertainty.  He says "How many findings of the IPCC AR$ WG1 are incorrect?: answer 28%".  However what he doesn't mention is that this is based on one paragraph from his paper, and even that doesn't support his conclusion.  What it would support is a statement along the lines of "Assuming the IPCC fundamental scientific position is correct then we would expect about 28% of their testable projections to be wrong".  However that is entirly uncontentious.  The whole point of the IPCC explicitly stating the uncertainities in their projections is becase they don't expect all of them to be true!  Roger however has made a post that deniers will latch onto as saying 28% of IPCC science is wrong - it clearlt doesn't.  See coverage at James Annans' blog (who clearly does understand uncertainty - unlike Pielke).

Annan eviscerates Pielke.  Also here and here.  Would it be worth discussing Pielke in this post on Curry?  Neither seems to have any understanding about how uncertainty works.

FWIW, I am engaging with PIelke here.  It seems to me that he had difficulties reasoning about probabilities of probabilities.  I am not unduly surprised by this as it isn't that straightforward - it is all too easy to confuse the probability (or expectation) of a probability with the probability itself.  It is easier if you are from a Bayesian background rather than a frequentist.

My comments seem to have suddenly started staying in moderation for a very long time, even though posts made later have already appeared.  This one has been waiting a particularly long time:

=========================================================================

    I think I may be getting there; I think the problem may have something to do with having a p.d.f. for the value of a probability. Talking about probabilities of probabilities often causes difficulties as it is easy to mistake the probabilty of a probability for the probability itself. Which I think may have happened here.

    In computing the "maximum probability", you have taken a "worst case" p.d.f. with point masses at 0.05 and 1.0, and then computed the expectation. But this is not the "maximum EXPECTED probability" after having marginalised over the uncertainty expressed by the p.d.f.

    Likewise for the "minimum probability" you have taken a "best case" p.d.f. with point masses at 0.05 and 0, and computed the expectation. But again there it is the "minimum EXPECTED probability", not the probability itself.

    Thus you have formed bounds on the expectation of the p.d.f., but that is not the same thing as a bound on the actual probability itself (which as I have pointed out can be anywhere from 0 to 1).

    To demonstrate that the probability can be anywhere from 0 to 1, is there anything in the original statement that definitively rules out any particular value of x? No (hence comments about unfalsifiability), so the range can't be any narrower than [0,1].

======================================================================

However roger has magically inserted the word "expected" into his reply at 36 into his reply to my comment at 35 to his question at 34, neither of which mentioned expectations! ;o)

If further of my posts fail to appear on Roger's blog, it isn't because I haven't been posting them, I shall persevere!

Back to Curry.  I am reading the paper at the moment.  It seems to me to be mostly word salad with little useful to say (if if it does say something useful does so in language that makes acting on it very difficult).  It would have been much more useful had shee given some suggestions on how she thinks AR4 projections should have been worded.

Anyway, it looks to me as if there is an error on page 4, where she says "In the presence of scenario uncertainty, which characterises climate model simulations, attempts to produce a PDF for climate sensitivity (e.g. Annan and Hargreaves 2009) are arguably misguided and misleading".  Surely estimates of climate sensitivity are based on *hindcasts* of climate and hence use the scenario that was actually observed, rather than forecast.  I would have expected there to be only "scenario uncertainty" for forecasts?

Thanks for the comments; I've cleaned it up some and made it less personal.  Dikran, I only hit what I considered the three most glaring parts.  To think that BAMS would actually publish 'the uncertainty monster' is astounding.