Tuesday, August 23, 2011

The Crapshooters Dance

Some, not Eli to be sure, think that you can spot a Rabett post by the presence of snark. Now Eli will admit to a fondness for the stuff, but nothing like Ethon who likes his with liver, and has been snacking over at James' Empty Blog

The subject of all this, is as the proprietor over there puts it

But principally, I think it's important to realise that Roger's blizzard of posts is a very straightforward smokescreen to bury the car crash of his original claims. His goal is to get everyone to agree that it's all far too complicated and even experts disagree. But they can't disagree on whether his original idea is credible, because it's obviously nonsense to think that the correctness of a probabilistic claim can be determined solely on the observed outcome. Which is where I came in :-)
Steve Scolnik discusses the Red Queen
"I daresay you haven't had much practice," said the Queen. "When I was your age, I always did it for half-an-hour a day. Why, sometimes I've believed as many as six impossible things before breakfast."

RPJr is clearly going for a new world record
Mo gets a bit impatient
Concerning JA's last paragraph: Question: what is worse - having somebody tell you your fly is down, zipping it up, then berating that somebody for "snark"? Or, doing the same, except then stomping away petulantly with your fly still wide open?

Answer: I don't know, but today I learned that in the intersection of impenetrable high self-esteem, low competence, and a low standard for argumentation, there exists a hero.
but Bob Grumbine had the nutshell version
.* Scientists should be forthcoming about their uncertainty.
* To be less than 100% certain is to be wrong (by the amount of the difference).

Hokay. A little more subtle than 'heads I win, tails you lose', so I suppose points for creativity
Still that's Ethon's organ meat, Eli is more interested in some science, which goes back to Roger's original claim that
How many claims of the IPCC AR4 are incorrect, answer 28%
based on the idea that the "average" confidence the IPCC panels assigned to their findings was 72% ( it is really not clear how Lunch got this, but no, never mind), rather let's look at the ranges assigned to the measures of likelihood
Virtually Certain (>99%)
Extremely Likely (>95%)
Very Likely (>90%)
Likely (>66%)
More Likely than not (> 50%)
About as likely as not (33% to 66%)
Unlikely (<33%)
Very Unlikely (< 10%)
Extremely Unlikely (<5%)
Exceptionally Unlikely (<1%)
There is, of course, a basic nonsense here, because, as James points out something rated exceptionally unlikely has a 99% probability of being wrong in Roger world and contributed to that 72% mystery, but there is also something interesting in these rankings which Eli has never seen commented on.

Take the ranking of Likely. This does not mean that the IPCC expects that of 1157 such predictions a minimum of 764 will be observed.

What, you say?

Look at the rankings, Likely is for a prediction that has a likelihood of > 66%, Very Likely is something that has a likelihood of > 90% . If the panel thought something had a likelihood of more than 90% it would have assigned a Very Likely rating to it, if it thought the prediction had a likelihood of less than 66% of the time it would have assigned a ranking of More Likely than not.

From this we conclude that the panel thought the likelihood of something the labelled as Likely was between 66 and 90%. Moreover, it is UNLIKELY that a substantial number of panel members or whoever did the rankings would have assigned a likelihood of exactly 66 or 90%, otherwise they would have down or up rated the likelihood, which leaves us most conservatively with a uniform prior across the interval between 66 and 90%, and an average likelihood of 78% over a huge number of instances, or much more likely with some sort of bell shaped distribution centered in the interval whose width is somewhat less than the interval at worst. It is unlikely that a substantial number of panel members would have been holding out for lower or upper limits. It would be pretty silly to rate something that was right on the dividing line.

What's not to like? But you could, if you were desperate, get a paper out of this. Tell them Eli sent you.


David B. Benson said...

Yes, RPJr is clearly trying for the Red Queen award.

Steve Scolnik said...

Update: Andy S has pointed out that it was actually the White Queen. Full details at the snark mine:


Anonymous said...

Actually, given the flurry of communications over this ... misunderstanding, it seems to me that he should be nominated for the White Queen and the Red Queen awards!

Cymraeg llygoden

badger badger badger said...

No need to hunt for it here.

Anonymous said...

Has Pielke read Tristram Shandy? Sterne beat Carroll to the punch in his comedy about logic and language. Fittingly -- and almost presciently -- one of the high points comes in a discussion of a white bear.

A White Bear! Very well. Have I ever seen one? Might I ever have seen
one? Am I ever to see one? Ought I ever to have seen one? Or can I ever see one?

Would I had seen a white bear! (for how can I imagine it?)

If I should see a white bear, what should I say? If I should never see a white bear, what then?

If I never have, can, must, or shall see a white bear alive; have I ever seen the skin of one? Did I ever see one painted?--described? Have I never dreamed of one?

Did my father, mother, uncle, aunt, brothers or sisters, ever see a
white bear? What would they give? How would they behave? How would the white bear have behaved? Is he wild? Tame? Terrible? Rough? Smooth?

--Is the white bear worth seeing?--

--Is there no sin in it?--

Is it better than a Black One?


And since that merges with our previous discussions about Charles Monnett, I'll retire on Sterne's glory.

Jeffrey Davis

Anonymous said...

I will point out that I raised the question of whether Roger used the lower limit or the center of the range: (at http://www.blogger.com/comment.g?blogID=9959776&postID=2850025249697796529)

"1) "very likely" should be 90 to 99% - or, on average, about 94.5% - does Roger assign it 94.5% or (as somehow I suspect is more likely), 90%?"

JA has also speculated that Roger has inverted the unlikely probabilities, such that the IPCC only gets dinged if a very unlikely event actually happens - if so, Roger was very unclear about it.

Finally, I'm still amused by Roger's statement that "The IPCC SRES went to some length to explain that its scenarios were considered equal probability," when, in fact, the IPCC went out of their way to note that the scenarios were _not_ equal probability. I don't expect Roger to get statistics right (eg, Megan the undergrad), but to claim that the IPCC "went to some length" to do something that was the opposite of what it actually did is taking confusion to new lengths...


John said...

Crap shooters OR crap slingers?


John Puma