« Trusting Sponsored Medical Research | Main | My BHTV with Tyler Cowen »

March 14, 2009


But what about those of us who almost always agree with you? Who are we going to bet with? It would be nice if you held some stupid beliefs so we could get in on the action :-)

Robin, IIRC you're not crazy about play betting markets--but token bets between intellectuals seems like a play betting market. What's the distinction? Is it play betting markets among intellectuals is importantly different than play betting markets among everybody? That seems plausible.

It seems like we could detach intellectuals from their fear of status by creating anonymous betting markets that certify the players are experts in their fields (academics, wrote a book on the subject, award winner, et c.). So an expert would just be known by some pseudonym.

I similarly think anonymous debating clubs among certified experts might be a good thing, for similar reasons of overcoming fear of loss of status (though the incentive of gaining status is also lost).

"Slowing this down to post bonds, make bets, or even just think carefully, would just not be worth the tradeoff."

An internet play-money prediction market is a mechanism that changes the trade-off into a positive one. They are so easy to set up and use, Tyler really has no excuse.

As someone who puts himself forward on the internet, he should understand his actions and statements really will reflect on his character. He speaks of "mental compartments" but his reputation knows no such. It's disappointing.

Likewise, my estimation of Bryan Caplan rises.

One has to factor in the publicity value of bad bets. If the publicity value is greater than the loss of finance and the loss of status among the portion of population who cares that a pundit predicted incorrectly, then pundits still have an incentive to public make and bet on bad predictions.

play-money prediction markets is just another term for reputation markets. Those would be good to see. But for those that care, no one has seems to have yet made the effort just to use google to create a rough quant database of predictions and results matched to pundit, politician, and/or academic. I think that's the natural starting point.

I think you're wrong about people choosing fancier hospitals in order to associate with high-status doctors. I think patients are worried about their health, don't find it easy to evaluate doctor quality, and use reputation as a proxy.

But how can we turn this into a bet?

Hopefully Anonymous raises a good point, if these reputation markets are so valuable, why don't we have any of them of any substantial size?

Regardless of the actual scoring mechanism: (if its betting, a public registry of predictions, these political bonds etc) the notion that people should be measured and held accountable for their predictions is praiseworthy. The world would be a much better place if this were the case because accountability, measurement, and verifiability are the cornerstones of the scientific method and progress.

However, I am skeptical this will happen in any meaningful way anytime soon simply because the world is not ready for it.

Firstly, I don’t think there is much demand for this type of product.

Outside of a small select group of people, the vast majority of people don’t actually care about predictive accuracy. In terms of mass media, I cannot think of a single instance where a person’s track record is of any significance whatsoever. Consider CNBC as an extreme example, none of the analysts there are held accountable for their drivel. I have never seen the track record of a weatherman even though it would be very easy to produce, never seen any such records in the newspapers other than in fun polls. In fact, in key decisions where people should care about track records, such as medical financial choices the evidence suggests that even when these track records are readily available people ignore them! Consider the mutual fund industry. Theoretically this entire industry should not exit. It’s highly likely to underperform a traditional index, but there are trillions of dollars in the mutual fund industry, and each dollar is a vote that says to me, "we don’t care about the facts and we don’t care about track records."

Secondly, I am skeptical about the supply side of this equation too.

If you are quantifying your predictions, you are open to public scrutiny and in comparison to those folks that don’t, you are going to be at a huge disadvantage. Even if you had an excellent predictive track record, these track records are optically way lower than your average person expects and so it would be a liability to publicize it. In short, what will be one’s incentive to publicly measure oneself when one’s competition doesn’t, and more importantly, when the customers don’t care. It's noble but irrational do to this.

All in, the effort required to overcome this demand AND supply inertia may be too great to overcome. It may require a really large paradigm shift that may be a long way off.


"world not ready"

Whose world? Yours maybe - how I tire of this endless Tao of No. Many interesting markets are still run behind closed doors!

@Frelkins ... Its possible that you are right and I am wrong. But if you don't mind I'll remain skeptical until you can produce more evidence. I have one question for you though, if these markets are confidential and still run behind closed doors how do you know about them?


Because everybody talks to the blonde - and she keeps all secrets!

Mike, you are putting words in my mouth; the important thing is to be betting something of enough value.

Hopefully, the question is how to get people to make precise enough claims, and then how to get others to compare these with outcomes to collect a score. Bets are one way of doing that, but hardly the only way.

Russ, I don't see how you can interpret my post as claiming there is a large demand for this; the supply would exist if there were enough demand.

"Hopefully, the question is how to get people to make precise enough claims, and then how to get others to compare these with outcomes to collect a score. Bets are one way of doing that, but hardly the only way."

Robin, you're far more expert than me on this, but I intuit that you're wrong. I think Russ is closer to the truth when he says we have an abundance of data. There's a lack of demand, but I think that's because there's not a good enough, conveniently accessible enough product.

A commenter (I forget who) proposed in your comments sections years ago that one could simply aggregate and quantify predictions from google searchable media. You were skeptical, but I think he was right. It's a mamoth task, but it's the sort of thing I could see a google or wikipedia type entity embarking on at some point. Once a decently created, decently publicized entity like that is up and running, I suspect it will create its own demand -and will occupy significant mindshare.

This isn't to take away from other ideas, such as the many innovative ones you have come up with. I just think that the data is out there and that the demand is latent. Some entity has to commit to the costs of developing and marketing the product. Then we'll all be able to track predictions and compare the status of predictors, the way we can compare the status of billionaires, itunes podcasters, and chessplayers.

By the way, Russ, if you blog please publicize it here in the comments. If you don't, I encourage you to start blogging.

Hopefully Anonymous, it sounds like you are referring to "the iSteve.com Bad Bet Service".

I am again lowering my regard for Cowen and raising that for Caplan & Hanson. But I suspect that could be irrationality on my part as I could have predicted that pretty much any disagreement between them would result in that (unless Caplan & Hanson disagreed, in which case my opinion of Hanson would more likely increase).

You say that "the supply would exist if there were enough demand." and infer that little demand exists. But surely we can hope to change this! After all, many entrepenuers suceed by inventing products which no one had previously imagined they would want.

Robin, you are being much too pessimistic. While many people do primarily want to "affiliate with a stately staccato stream of statusful statements" but some people do care about the accuracy of their pundits (and more people fancy that they do). Why should it not be possible to transform the metric by which pundit status is judged?

You say that "the supply would exist if there were enough demand." and infer that little demand exists. But surely we can hope to change this! After all, many entrepreneurs succeed by inventing products which no one had previously imagined they would want.

I think that if a few communities of intellectuals could be persuaded to adopt better, more meritocratic status norms then such norms might very well spread. It's not hard to imagine other pundits being shamed into adopting those norms. Krugman has lost status by ignoring Mankiw's proposed bet (at least among the people who heard of it).

Wikipedia has some discussion about it: http://en.wikipedia.org/wiki/Simon-Ehrlich_wager#Other_wagers

Such bets are still very rare. The main problem seems to be agreeing to terms of the wager, usually two sides want to bet on not exactly the same thing. From brief look at their blogs it doesn't seems that Krugman and Mankiw are even arguing about the same thing, Krugman is predicting some level of post-recession rebound, and Mankiw is talking about exact Obama's GDP predictions, two are only very loosely related, it's very easy to imagine that rebound will happen, but GDP predictions will be way off.

Hopefully, go ahead and try your idea and prove me wrong. I say it is too expensive for a third party to try to fairly score typical google-found vague claims as right or wrong. Claimers need to make the effort to choose a more precise scoreable claim, as people will just not trust a third party has having fairly interpreted their claims.

Jacob, yes we may well develop a niche market for a product, even if no mass market exists.

" . . . rumors that trash folks are very influential in female cultures that allow such things." Are you calling Rahm Emanuel and his media allies pansies?

I'll stick to more narrowly focusing on maximizing my persistence odds, thank you very much.
Still, I intuit that you're wrong, wrong, wrong about this "I say it is too expensive for a third party to try to fairly score typical google-found vague claims as right or wrong. Claimers need to make the effort to choose a more precise scoreable claim, as people will just not trust a third party has having fairly interpreted their claims. ... no mass market exists."

in the sense and to the the extent that I outlined in my earlier commments to this thread.

Robin, fair enough--I didn't take into account the risk to reputation that was brought up in thread.

Frelkins and Russ Anderson, you should make a public bet on whether prediction market participation or some other metric will increase by a specified percentage within a year or two.

@Khafra. I am already overly long prediction markets, but in the spirit of the post I would be willing to participate given that Frelkins seems to be a sassy blonde in the know. No problem. I guess it would be an emotional hedge of some-sort.

My basic position regarding PMs is that they have potential and merit. Obviously. They dramatically reduce forecast error. And in a perfect world they would be much more prevalently used. But they also have some fundamental challenges, mainly related to the politics of their adoption within corporate America, which suggests that perhaps their growth will take longer than most people realize or anticipate. So relative to expectation their adoption will be smaller and slower than participants expect. It is also possible that to get PMs to the point where they are perceived as more than just a cool fad, it may also require a dramatic changes in their structure and their business models. So I'm short them. Which is unfortunate. The concept is great and the science behind them sound. The primary problem with PM's is that the decision makers within corp America do not want to support them. Reminds me of the great Jack Nicholson line in A Few Good Men: "you can't handle the truth." http://www.youtube.com/watch?v=5j2F4VcBmeo

So if you can structure a relevant trade on prediction market adoption in 2009 underperforming, then I would sign up for it in the spirit of the post. Let's see if we can find some form of index or measure to trade on and go from there. I'm game. I’m also curious to know what Frelkins knows that maybe I don’t know.

Put your money down, suckas! I'd enjoy spending it on Louboutins and Champagne.

"the politics of their adoption within corporate America"

Because you don't know how to sell them, Russ, doesn't mean it can't be done. How I tire of prople telling me bumblebees can't fly - many markets exist. Even Hanson has estimated "hundreds," iirc.

@Frelkins ... You amuse me and seem to be a good sport. But I think we have taken this debate as far as it can go before we irritate the other readers here with our petty squabble. I like OB because there is little to no bickering and the quality of the responses high. Fact is I have personally been involved in selling some of the larger deals that I am aware of in the PM space, which in truth is not saying much at all. And I can assure you that it was phenomenally hard to do, which is why I made the points that I did. This is first hand real world experience my friend, not something I read about on a blog.

Secondly, those supposed luxury french shoes you linked too, I believe that they are actually made in china. And not only that, I'm further informed from a reputable source, that some of the lasts they use for the larger sizes are not even women's shoe lasts, they are men's! Oh my friend, I am giving you such an education here, I should charge you a tuition! :)


And let be noted that in the end, after tossing his gorilla dust and pounding some tree branches on the forest floor, Russ wouldn't put his cash on the line. Blonde Conquers Once Again! You're right Russ - you're done. :)

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30