Friday, August 18, 2006

Steve McIntyre's bet offer

Via Backseat driving, I see that Steve McIntyre wants to bet over tree rings. His contention seems to be that they are a very limited or perhaps even useless indicator of temperature:

Do any of the warmers want to bet that European tree rings in the very warm year of 2003 did not show very wide rings such as predicted by the MBH assumption of a linear relationship between temperature and ring width?

Or that Sheep Mountain bristle ring widths in the period 1990-2005 were as wide or wider than projected by a linear model - we can define the model, but essentially it’s the linear assumption of MBH.

I’ll bet either.

Regular readers will be unsurprised that I consider this bet idea to be in principle a good one, although no doubt there's a debate to be had over the precise details. I don't know anything much myself about the reliability of tree rings as an indicator of temperature myself, and (especially given the confusion over the NAS definition of "plausible" etc) I would be very interested to find out the real opinions of the experts in the tree ring circus, as demonstrated by their willingness to put their money where their mouths are. Furthermore, the ultimate test of scientific theories is their ability to make predictions: validation of a real prediction of tree-ring growth would go a long way towards demonstrating that these analyses are credible. In contrast to my bet on future temperatures, the data are sitting there just waiting to be measured and so there is no problem with a long time scale between bet and payoff. If the stakes are high enough, it could even perhaps pay for the fieldwork, and climate science would benefit from some free data courtesy of the oil barons who fund M&M or the environmental pressure groups who pay MBH etc, depending who loses :-) (Please, that's just a feeble joke. Don't get your knickers in too much of a twist. I realise that there are others apart from MBH who ought to be prepared to front up to this challenge, but they must be first in line.)

I'd be happy to play any part I can in negotiating reasonable terms for a bet based on what the various protagonists have stated in the literature and elsewhere. Eg, I presume Steve M would expect (next to) no correlation between temperature and ring width, hence no trend over the last 20 years (or whatever). Whereas those who do the reconstructions presumably expect to find a linear trend of a certain size, as determined by their previous calibrations. So there should be plenty of room to split the difference and define a bet that is highly attractive to both sides. Bring it on!


Anonymous said...

Why would the warmers want to bet that European tree rings in the very warm year of 2003 did not show very wide rings


Shouldn't climate audit be full of takers for that though they wouldn't be warmers :)


EliRabett said...

How about cause while it was was hot it was dry. OTOH, European covers a lot of territory, and I suppose you can get a ring of anywidth whatsoever somewhere. for example Switzerland and Western Norway were both hot and wet.

“‘Son,’ my father told me, ‘there will come a time when you are out in the world and you will meet a man who says he can make a jack of hearts spit cider into your ear. Son, even if this man has a brand-new deck of cards wrapped in cellophane, do not bet that man, because if you do, you will have a mighty wet ear.’”

— Damon Runyon

Brian said...

I don't have any objection to this bet (assuming it's fair), but I don't think it's as useful as a bet over global warming. The odds we give to global warming should signficantly affect our real-world policy positions. I think we've shown that most denialists are willing to risk other people's lives but not a few thousand dollars of their own money, which indicates to the rest of us how seriously we should take their policy recommendations. The value of tree ring proxies to actual policy is far more removed.

But as to Eli's quote, if you don't trust the bet offer, you can make up your own offer, and buy your own deck of cards.

James Annan said...


Well, this bet doesn't attempt to address the same question. The predictions of continued future AGW would hardly be affected by any outcome. However, it certainly represents a clear independent test of the ability of tree-rings to represent climatic variations (and eg to what extent they measure temperature as opposed to precip, as Eli mentions). That is, I think, an interesting scientific questioon in its own right, and one on which people have expressed widely differing views.

Dano said...

StatsMan, though, would insist you'd need a sufficiently randomized sample to be able to process the signal:noise to eliminate bias and...and...zzzzz...

IOW, by the time the sample set is to those folks sufficiently analyzeable, it'll be too late and who will care by then?



James Annan said...


Even eliciting the (possible) bet can be interesting in itself, as I found with "as likely to cool as warm, so long as I'm offered 50:1 odds" Lindzen :-)

You may have realised that SM wasn't really very sceptical in his bet offer - even to a true believer, the probability of the rings being at all lower than the linear prediction is 50%, same as being higher. If he was confident that the rings showed no relationship, he should be prepared to make a much better offer. That's what I tried to allude to in my last paragraph.

But even if his "offer" was more showmanship than meaningful, there are still no counter-proposals from the ringers so far. I'm disappointed....

Dano said...

I agree, James.

What would also be exciting is if SM were to actually go out in the field and core some trees to prove his point. Since he hasn't cored anything, he doesn't have any data to back his claim, so he is [IMHO] counting on the continued dispersal of his contructed dialogue to carry the day.



DWPittelli said...


A year later it seems that SM has indeed gone out into the field and taken some tree cores, in the U.S. Note his October 12th, 2007 post on Climate Audit.

empedocles said...

how did this bet work out?

James Annan said...

Um...I don't think it went anywhere. Which, if true, is a shame.