Wednesday, July 22, 2009

Nate Silver offers a sucker bet (updated)

(NOTE: I have retracted most of arguments in this original post. See my update below.)

Nate Silver, political pundit and sabermetrician, has challenged climate change skeptics to a bet. An unfair bet.

Nate starts by quoting a conservative blogger, John Hinderaker, who scoffs at climate change by claiming his hometown of Minneapolis is abnormally cold this year. So, Nate says, to Hinderaker or anyone else, put your money where your mouth is. He'll bet $25 daily, on a continuous basis, that your hometown will be warmer than trend that day. You bet your $25 that it will be cooler than average. Bets settle at the end of every month, and are renewed until one side cancels. The canceling party's loss will then be trumpted on the other's blog.

I'm arguing that Nate is a little bit of a hustler here; the bet is tilted heavily in his favor.

Those skeptical of global warming aren't claiming that temperatures are getting colder; they're assuming that the climate is staying pretty much the same. With both parties assuming cooling has zero probability, that leaves only two possibilities:

1. temperatures being above average;
2. temperatures being about the same.

Which means, according to the rules of the bet: If there's warming, Nate wins. If there isn't, then, on average, he breaks even.

I respectfully submit that "Heads I Win, Tails I Tie" doesn't exactly sound like an equitable wager. Nor does it sound like the kind of offer you'd expect from someone who's confident in his position.

For the bet to be fair, Nate should state his expectations for temperatures. Perhaps he thinks that, because of climate change, the weather should be warmer than average, say, 60% of the time. The skeptic might say it should be 50% of the time. So, in that case, 55% would be a fair bet, equally appealing to both sides. Silver pays $5.50 every cooler day, and the skeptic pays $4.50 every warmer day.

Now *that* is a bet where both sides have something at risk. And, on those terms, I'm willing to accept if Nate is. Officially, I don't meet Nate's requirement that I be a resident of the United States, but I do have access to US dollars. And the stakes are a little high; I'd be willing at the $5.50/$4.50, but warier about $27.50/$22.50. (But if the stakes are the only thing holding Nate back, I'd be willing to reconsider.)

Nate, are you interested in this fairer bet? Are you confident enough that at least 55% of days should be warmer than trend?

And if you're not willing to bet at 55%, because you think the true number is lower, then what do you think it is? What kind of bet are you willing to make that puts you at the risk of actually losing -- instead of just tying -- if you're wrong?

P.S. Please do NOT argue climate change here. I know little to nothing about science of global warming, and I do not consider myself a skeptic. My point is about Nate's proposed bet, not about climate change itself.

Hat tip: Freakonomics


UPDATE/RETRACTION: In the comments, Doc suggests that Nate's bet really is directed at global cooling advocates. I assumed it wasn't. But I think Doc is right. Nate's original post quotes Hinderaker:

"... the current low level of solar activity suggests that the cooling trend could indeed be universal."

I missed that originally. I should have caught it.

Not to excuse my error in not reading carefully enough, but Nate doesn't explicitly mention cooling anywhere. In addition, his post is entitled "A Challenge to Climate Change Skeptics." And I'd argue that's overly broad; in my unscientific exposure to skeptical arguments (mostly from the National Post newspaper, which frequently publishes "denier" views on climate), only a small minority of skeptics argue for cooling. A more reliable count comes from That site, which seems to be devoted to debunking the skeptics, has a empirical (but not scientific) list of the frequency of arguments it's found against the consensus view of global warming. A generous estimate is that cooling theories seem to comprise about 20% of the skeptical arguments.

So I'd agree that those 20% should be willing to accept Nate's wager; the fact that none of them has does, indeed, signal something about the confidence the cooling advocates have in their hypothesis. My confidence in the global-cooling theory is lower than it was before Nate's bet, and I thank him for the evidence.

But it's too bad the post didn't emphasize the cooling issue. I googled "nate silver climate bet," and randomly clicked on a few of the finds. None of them explicitly mentioned that Nate was betting only against cooling. And many commenters, on Nate's site and elsewhere, interpreted the lack of interest in the bet as reflecting on skeptics in general. But they shouldn't. An advocate of the "sunspot" theory would have no reason to expect colder temperatures, and therefore no reason to accept Nate's bet. Technically, Nate should cause my estimate of the chance the sunspot theory is correct to rise, not fall; since I am less confident about cooling, I should be proportionally slightly more confident about the remaining hypotheses -- since my intuitive probabilities still have to add up to 100%.

To me, and many others who are trying to make sense of the debate, the willingness or unwillingness to bet on one's views is a very strong signal of one's beliefs. Indeed, it's one of the strongest arguments to those (like me) who don't know much about the science. The climate debate is highly charged politically, and any apparent evidence will get trumpeted far and wide by those on that side of the political debate. That's why it's so important to get things right.

Labels: ,


At Wednesday, July 22, 2009 1:51:00 PM, Blogger Phil Birnbaum said...

I should probably make it clear that I don't think Nate is deliberately trying to sucker anyone ... it was probably just an oversight that he proposed a bet riskless to himself.

I should also disclaim that I don't know Nate at all, don't know his political work very well, but am an admirer of his sabermetric work.

At Wednesday, July 22, 2009 6:22:00 PM, Anonymous Matt said...

I tend to think your proposal is more "fair," although I seem to remember that a climate change skeptic posted a graph on his blog demonstrating that most of the warming of the last couple of decades or so has been cancelled out by recent cooling.

For the record, I live in Denver and the cooler-than-average days have out-numbered the warmer-than-average days this month sixteen to five.

At Thursday, July 23, 2009 11:50:00 AM, Blogger D. said...

I probably shouldnt try to put words in Nate's mouth, but I think you are missing part of the point. From the beginning of his post, I get the sense that he is not trying to bet against the global warming skeptics as much as he is the people who blog/appear on Fox News/rant that "global warming cant exist because it is cooler this summer than normal".

At Thursday, July 23, 2009 11:57:00 AM, Blogger Phil Birnbaum said...


I agree with you that that's what he's trying to do ... but he's attacking a straw man.

The skeptic argument is, "the cool summer is evidence against global warming." The skeptic argument is NOT "cool weather will continue," which is what Nate is assuming, and asking them to bet on.

Nate's rebuttal should be, "a cool summer doesn't disprove global warming." That would be correct, in my opinion, but it requires a statistical argument and doesn't lead easily to an opportunity for a bet.

At Friday, July 24, 2009 10:32:00 AM, Blogger Don Coffin said...

I want to argue that Nate's offered bet is aimed at the advocates of global cooling, so it is, in fact, a fair bet. And are there such advocates? Yes:,2933,333328,00.html

That's just from the first page of a Google search on global cooling.

So a global cooling believer--of whom there are apparently many--should find Nate's bet a fair one.

At Friday, July 24, 2009 4:30:00 PM, Blogger Phil Birnbaum said...

Doc, I think you're right. I may have been fooled by the title "global warming skeptics". The guy he quotes does seem to be an advocate for the cooling theory.

I'll prepare a correction ...

At Sunday, July 26, 2009 1:03:00 AM, Blogger Brian Burke said...

This bet seems in line with a general confusion about what warming skeptics are skeptical about. Anyone can read a thermometer. The debate is over the cause of climate change.

I think some on the left who advocate major government intervention want the debate to be centered on whether warming is happening, which over the past century is generally undeniable, despite a slightly cooler period over the past ten years. But that doesn't mean the cause is man-made.

The current cooling phase is evidence against the man-made theory. Theories make predictions, and reality is not conforming to those predictions at the moment.

At Sunday, July 26, 2009 9:56:00 PM, Blogger Jeff J said...

BB: There is no such debate. Not to the extent you think there is, anyway. There is a very strong consensus among people who actively work on this problem.

In anticipation of the canned libertarian response that the source of said consensus is bias resulting from the desire for these academics to continue to suckle from the teat of big government:
1/ These academics, like all humans, are not close approximations of homo economicus.
2/ Even if they were, the money's in denying, not crying about the sky falling.
3/ They earn their living by looking skeptically at own work and that of their peers. When doing their peer reviews, these chicken littles turn into bigger AGW skeptics than any paid shill from the Cato inst.

It's unfortunate such a clever post has to be marred by these standard internet blatherings. I apologize, Phil. I just can't help myself.

At Sunday, June 13, 2010 12:36:00 AM, Anonymous Careless said...

It's a year late, but Nate's bet was indeed a sucker bet, and not in the direction you meant. Global warming climate models are really, really inaccurate in any small time period, no matter what you think of their long term accuracy.

Weather forecasts for the next 1-3 months are orders of magnitude better. I wrote at the time that I'd love to be able to take the bet up with him, but didn't qualify (I don't even have a blog), and I would have won something like $750 from him. Why? Because there were large parts of the country expected to have well below average temperatures in that time period, and those forecasts were correct.

And if he wanted to extend the bet for another month, it would have wound up well over $1,000 in my favor. Even smart people can write ridiculously stupid things, and that's what he wrote here.


Post a Comment

<< Home