Bishop Hill Tamino on the Hockey Stick Illusion
Jul 22, 2010 Climate: HSI Tamino has a rave review of the Hockey Stick Illusion up at Real Climate. I'm reading it now.
A few initial observations - there is a lot of discussion of proxy selection rules in Tamino's piece. This is complex for those who aren't embedded in the nitty gritty of the science, but stand back and ask yourself this: if you have over 100 series in your database, and one of these is the fourth most important pattern in the tree rings of a couple of closely related tree species in one area of the western USA, how comfortable are you that this series should form the basis of the temperature reconstruction for the northern hemisphere? The idea that you can reconstruct hemispheric temperatures in this way is deeply unsatisfactory.
Tamino doesn't try to defend the use of the bristlecones. It's not clear why it's worth arguing about PC retention if everyone (including the NAS and Wegman panels) agrees that bristlecones are inappropriate.
The observation that "McIntyre argued that the entire Gaspe series should be eliminated because it didn't extend all the way back to 1400" is wrong. MBH had its steps starting at 50-year intervals. Gaspe should therefore have been in the 1450 step not the 1400 one. There is probably an argument that Gaspe should be excluded because the update that was taken didn't show the same shape (although it was never published and everyone seems to have subsequently forgotten where the actual location of the trees is).
There's a lot of discussion of reconstructions being hockey stick shaped. The critical issue is of course the relative warmth of the medieval and modern periods.
Remembering Matt Ridley's article about straw men? I am criticised for complaining that hockey stick shape proxies dominate reconstructions and apparently I imply "unfairly" so. I explain how the hockey stick shaped series come to dominate the reconstruction, but I don't imply unfairness as far as I remember. Of course, since Tamino doesn't actually quote anything I said that implies unfairness, it's hard to respond in a precise way.
I'm accused of quote mining re the "better for our purposes" quote. Given that I start by saying that an innocent explanation is possible, I'm not sure this is reasonable criticism. I don't think it's a "killer quote" but it was there and McIntyre raised it in his correspondence with Nature, so it needed to be discussed.
There are very few quotes of any kind from the book (go figure), but certainly some I'd like to use on the cover of the next printing:
"A narrative worthy of the best spy thrillers"
"...spins a tale of suspense, conflict and lively action, intertwining conspiracy and covert skullduggery, politics and big money".
My publisher is going to love it.
Reader Comments (229)
Fred, I have a simple take on the matter. Others please correct me if I am wrong.
If you take a set of random data series and select from it only those which match the uptick in CRUTEMP temperatures after 1900 you will always be left with a hockey stick of the desired AGW shape.
Fred did say that he was a layman ^.^ if he is then the answers he received will put him off climate blogs forever hehe.
Come back pls Fred :)
Here's something really funny. Someone posted on RC that RC had twice the readership of CA. They said check Alexa.
Well, I knew that was bull, so I checked Alexa:
[url]http://www.alexa.com/siteinfo/climateaudit.org#[/url]
[IMG]http://traffic.alexa.com/graph?&w=400&h=220&o=f&c=1&y=r&b=ffffff&n=666666&r=3m&u=climateaudit.org&&u=realclimate.org&[/IMG]
As you can see, CA regularly runs ahead of RC.
So RC removed the post claiming RC had twice the readership and replaced it with:
[url]http://www.realclimate.org/?comments_popup=4431#comment-182162[/url]
[quote]#299
[edit - please stay on topic and be substantive]
Comment by chek — 25 July 2010 @ 6:58 PM[/quote]
Do you think they would have done that if chek's post had been true?
Just more shenanigans from RC.
(I see I can't get the image to show up so just go to Alexa and put in climateaudit.org and realclimate.org comparison and you will see the result.)
Fred
I hope we have answered your question. Let's know what Tamino and Galvin say. ;-)
Yipee, I now understand what Red Noise is!
f(x+1)=f(x)+random, pure drunken walk or Brownian motion.
Feed this stuff, with increasing x, into a black box that generates a graph of f(x) against x. Repeat until boredom sets in then compare outputs to infer what the black box did to the inputs.
If the difference (distance/anomaly- ignoring sign) between the start of each run and the end averaged close to that predicted by Einsteins formula (d root n?) while the mean of the differences, using signs, should tend to zero, then the box appears to be unbiased.
Taking the classic example of a chronic and hopeless drunk, who starting at a lamp post, staggers in a random way, his finishing point would, ignoring obstacles, be randomly distributed at all points of the compass on successive evenings.
If the ground were flat that is, prevailing winds were absent and there was not a kebab shop close.
Mr M's black box, despite the cloak that hides its internal clockwork, reveals through its behaviour that it would be a most valuable asset for the sporting goods manufacture.
The perceptive reader (that pretty well covers all who step into the Good Bishop's blog) may well have spotted that I resisted the temptation to allude to loaded dice, stacked decks and rigged roulette wheels.
These are examples of deliberate fraud and I subscribe to the belief that incompetence alone underpins Mr Mann's efforts.
royfomr
If you still believe incompetence on Mann's behalf read Steve Mc's latest challenge to Tamino.
Ok KM, how about
"incompetence and accidental fraud alone.."
(I feel a Python moment may be in the offing)
Gavin has just edited my last post #323 at RC. For those interested the repy from Gavin followed by my full post was this (I have put words [following edited] where Gavin cut it:
RE: #267
I asked the question:
"From a stationary stochastic process the MBH98 algorithm detects a non-stationary signal: just how does it do that and still get described as robust?"
Gavin answers this as:
"[Response: I have no idea what this question means. 'Robust' means that the signal retrieved doesn't depend excessively on the method used to get it. ...]"
[edit]
[Response: Ok, enough. I thought we were having an actual conversation, and instead you want to play games - boring. The initial PC data reduction step is not done to define what PC1 is, it was done to encapsulate the data in the N. American network. That encapsulation requires a proper selection rule (which is *not* defined as keep PC1 and throw away the rest), and when done properly (and if you don't like Rule N, suggest something else), makes no difference to the final result as demonstrated over and over and over again, and is even admitted by McIntyre. That is the definition of robust in that very small and uninteresting context. If you want to continue discussions, please move on to something that hasn't already been done to death in the original post. - gavin]
[following edited]
MM2005 (slightly paraphrased to make it clearer for readers) says:
“Without the MBH98 transformation, a 1 SD hockey stick occurs in the PC1 only 15.3% of the time.
Using the MBH98 transformation a 1 SD hockey stick occurs over 99% of the time"
Remember MM2005 are testing both a conventional and the MBH98 algorithms using stationary red noise which does not contain a hockey stick.
So using a conventional method without the MBH98 transformation a hockey stick is spuriously identified 15% of the time but using the MBH98 transformation a hockey stick is spuriously identified 99% of the time.
The first method tested is what is generally referred to as robust: the spurious signal is found in relatively few runs. The method using the MBH98 transformation appears to retrieve a hockey stick signal from the test data 99% of the time. As you yourself say: "'Robust' means that the signal retrieved doesn't depend excessively on the method used to get it."
The signal retrieved by MBH98 appears to depend excessively on the MBH98 method and is therefore not robust.
PS ;) Glad to see you are tracking Bishophill too!
PPS I'll reply to Lazar #287 separately.
Posted on RC:
#317
"and yet the points never rise to anything constructive, never address anything other than the 1400-1450 ad step in mbh98, never look at the larger data sets now available, and almost always mistakenly assume that this has some terribly important consequence that the world must be made aware of. Yes, we are aware. -gavin"
Gavin, look at your answer to #17
Would MHB1998 and MHB1999 not have looked quite different if the results of Vinther et all were known around the time Mann constructed the Hockey Stick? And you know where the HS led to.
“temperatures during the warmest intervals of the Medieval Warm Period,” which they defined as occurring “some 900 to 1300 years ago, “were as warm as or slightly warmer than present day Greenland temperatures”
Vinther, B.M., Jones, P.D., Briffa, K.R., Clausen, H.B., Andersen, K.K., Dahl-Jensen, D. and Johnsen, S.J. 2010. Climatic signals in multiple highly resolved stable isotope records from Greenland. Quaternary Science Reviews 29: 522-538.
Posted at RC:
Re: #325
Hi Gavin,
You have edited the remainder of my post which includes the relevent points I was making. As moderator it is your prerogative to conclude the argument in this way but I will not contribute under censorship. You made several minor edits of my posts earlier which were ok - they were generally off-topic remarks by me, your editing did not affect my argument and you similarly edited opposing views to mine. This kept it fair and balanced. By editing the rest of my points, that is no longer the case.
Regarding my promised reply to Lazar #287 I will post at Bishophill on the Tamino thread - I am sure that any comments Lazar may wish to make there in response to my answers will be posted in full and received with great interest.
I would appreciate it if you could post this in full.
Thanks for the conversation while it lasted!
Best Wishes,
ThinkingScientist :-)
Thinking Scientist
You are having fun, for sure. :)
You should point out to Galvin, what while he is British, he is living in America and is therefore under the constraint of the First Amendment -- Freedom of Speech.
RC aren't having a good time of it.
Two first rate contributors (Thinking Scientist and Judith Curry) publicly (and probably irrevocably) lost in one day.
"To lose one, Mr. Schmidt, may be regarded as a misfortune; to lose two looks like carelessness."
Are they just careless over there, or is it really desperation setting in? Pissing off your best contributors is no way to run a whelk stall.
(H/T Oscar Wilde).
RE RC and Gavin,
Gavin has posted my Re: #325 response above in full at RC.
He has therefore given fair opportunity to me to close out my end without it appearing that I have somehow disappeared becuse I could not think of answers or such like. He has allowed this to be concluded in an even-handed way in my case. Accordingly I sent him the following response:
Gavin,
Thank you for posting my response in full - you have acted in good faith
and I fully acknowledge it. I don't agree with your comments about
playing games (that is not my intention) but I do understand that you have
multiple arguments/commentators simultaneously and that as both moderator
and responder that gives you quite a high workload here. We have probably
both exhausted the dialogue here for the moment and both are repeating
ourselves.
Thanks for your invitation to keep the door open for me to continue
subtantive conversation - I am sure I will!
Feel free to post or not post this - its meant as a personal
acknowledgement that you posted my final message at this time in full and
I do appreciate that. I will also acknowledge that at Bishophill where,
as I am sure you have noticed, I cross-posted.;)
Regards,
ThinkingScientist
Well, I am a nailed down sceptic of the C02-driven armageddon. However, I was really surprised by Judith's remarks on RC. They read as vague and not really from someone who is meant to be up there with the top 10% of people understanding this subject.
Judith if you are to get involved in challenging the team, you need to pull out bigger guns than you have in that RC thread.
Just bought HSI online. Too bad it's not available on Kindle or I would be reading it tonight.
From Tamino's site March 3,
Tamino (or anyone else who cares to comment): what do you make of this chronicle of the Mann vs. McIntyre story?
[edit]
[Response: I'll give it all the attention it deserves.]
The edit was a link to Caspar & the Jesus Paper, so it looks like it took you 5 months to get his attention!
Incidentally, if you go into Google's cache to just before where his posts have disappeared
http://webcache.googleusercontent.com/search?q=cache:G1b5qiPD6foJ:tamino.wordpress.com/2010/02/22/statistical-geometry/+site:tamino.wordpress.com+n-sphere&cd=1&hl=en&ct=clnk&gl=us&client=firefox-a
you find that Tamino identifies a circle as 1-dimensional, and a sphere as 2-dimensional.
Even when pointed out, he made no correction.
The n-sphere is the n-dimensional surface of an n+1-dimensional ball. The 1-sphere is just a circle; the 2-sphere is the ordinary sphere we’re all familiar with; the 3-sphere is the 3-dimensional surface of a 4-dimensional ball, etc.
Thank you. Your idea of the hands being shackled is a good one, but I think the chain should be left unbroken. That would indicate how marginalized and constrained we are, and thereby add "Omph" to the image. It would be powerful by illustrating how, despite being "outlawed," we still dare to smash the icon of settled science. It's a striking image of dauntless defiance.
As for the motto, I still like my "Gore Resisters League," because it immediately clarifies what the button is about (global warming), Gore being the major promoter of the image and the cause. (I also like its cute rhyme with "War Resisters League.")
Alternatively, how about "Warmist-Resisters League"? (The hyphen might not be needed.) That also plays on the originators of the hands-snapping image. (Or maybe "Warmmonger-Resisters League" -- but that is probably a bit to elaborate and indirect.)
Another play on words would be "Get the Puck Outta Here." But that should probably wait a year or two after the introduction and popularization of the initial button, because it's not self-explanatory.
Roger Knights
Yes, there should be a broken chain for all the reasons you give. You do have a good sense of PR from your comments.
As for mottoes, I am not very good so I leave that to the general consensus
I'm afraid I read through some of the comments in the article, as I have a strong stomach. I found this reply to a comment concerning water vapour trends from Gavin:
What interests me here is that Gavin is saying changes in technology and/or data gathering procedures have affected the trend so much it's now pretty meaningless. But wait a minute, isn't that exactly what they have such a high confidence in correcting for with the surface temperature record? I fail to see the difference. Again, if it's trending up it's Global Warming, and if it's trending down it's an error!
Hmm, it appears that topology does define a circle as 1-dimensional, so there is no error after all.
If Tamino's real name is Grant Foster, that eternal question: "Who's behind those Taminos?" has been answered--namely from a science perspective, Raquel Welch.
I think its wonderful that "Tamino" keeps advertising the Hockey Stick Illusion. I expect sales to rise as a result.
For the record, if the Sheep Mountain bristlecones are removed from the roster, Mann's program cannot produce a Hockey Stick shape as a PC1 because none of the other original proxies has such a HS shape.
If another HS shaped proxy replaces the original Graybill proxy, then up comes the HS result.
Because Mann's program over-weights series which show a strong correlation to the instrumental period, any set of time series data built from "red noise" a type of random dataset which has some form of autocorrelation, Mann's program will 99% of the time also produce a HS shaped result, because by chance there'll be one that fits the bill of spurious correlation with the instrumental record.
None of this is remotely controversial except to people like Tamino who thinks better than clear unbiased statistical expertise like Wegman and Joliffe.
I can read any posts here on Bishop Hill and even if some of the science goes over my head they still make sense.
The same goes for Steve's posts at Climate Audit.
However Gavin's posts seem to be trying to make sure that your response is "wtf was that all about? My god this guy must be a genius..
I challenge you to snip this Mr Bishop since it an honest opinion :)"
@Pablo de la Sierra: You wrote:
Thanks for the compliment, but I hope that "broken" was a typo! What I wrote was the opposite:
Could you please send a signed copy of THSI to the comedian Tim Minchin who calls you a "crank"
Me personally? Or sceptics in general?
You personally.
http://twitter.com/#!/timminchin/statuses/106888689384296448