Prediction Markets = Crystal balls…

…as in they’re smoke and mirrors designed to make people think you can predict the future.

Listen, I’ve got a lot of respect for economists. They find ways to turn human behaviors into variables for a math equation (witness a popular one, “willingness to pay” how do you quantify that?), and their conclusions often yield pretty convincing insights into how human civilization works.

But the boundaries of this field aren’t well-defined. Somewhere between trying to figure out how the stock market works and building a behavioral model for how Vans were once among the most popular shoes in America (see Malcolm Gladwell’s The Tipping Point), there arose the prediction market.

The prediction market first became famous thanks to the pen of James Suroweicki, the financial columnist who found time between writing middling financial articles for The New Yorker to put together the “Wisdom of Crowds” in 2004, in which he told us that large groups of dumb people are smart.

The basic idea of p-markets is that by if people who are knowledgeable in a given field (say video gamers or game developers) trade futures on the likelihood of a given outcome—say that Wii will outperform PS3 in console sales—that group of people will be better than any one expert at predicting that outcome. The idea has caught on like wildfire, so much so that the business leaders of America have seen fit to bring him in for myriad motivational speeches designed to make their employees happy and them lots of money.

Here he is looking a lot like Patrick Swazye in “Donnie Darko:”
Fear prevents you from embracing prediction markets! Swayze, as motivational speaker and child porn aficionado Jim Cunningham.

If you fold your hands like this, you see the future…Surowiecki, as himself

At any rate, real economists and academics and whatnot have spent a lot of time thinking about prediction markets, and now a law professor from George Washington University believes they’re going to change the face of modern American politics. He’s so convinced of this he wrote a book about it called Predictocracy.

But here comes the odd part. In a recent post to the blog “Overcoming Bias” that same author, one Michael Abramowicz said, and I quote:

    While recognizing the contributions of experimental economics, I doubt that laboratory experiments will be of much use in persuading skeptics that prediction markets can be useful in real-world institutions.

If experiments can’t be done to convince skeptics of something, then as far as I’m concerned, that places it among on a very low, very crowded rung of the credibility ladder. Right down there with the teachings of Scientology and this rant from Kirk Cameron’s pal about how the banana proves that young Earth creationists were right all along.

One Comment

  1. Jeremiah Messer
    Posted January 29, 2008 at 9:50 pm | Permalink

    What I like about this article is your choice of the Wii vs PS3 as an example of a potential prediction market outcome test. It’s a great example because I bet, if there had been a prediction market before the Christmas season of their release, it would’ve been completely wrong about the winning console.

    I work in Market Research, and we did political polls right up to the New Hampshire primary, and their predictive ability is pretty limited. We did well; our results for our poll right before the election were spot on for the Republican primary and within in the margin for the Democrats (of course, the pundits were too busy pinning the blame for their pre-primary baseless assertions on polls to concede that maybe some were right). But that particular poll came out on Monday before the primary. Why not just wait ’til Tuesday and see the real results? I think some prediction markets had Giuliani with the greatest “share” a while back, and look how that turned out. They may be real time metrics, is what I’m saying, but not really predictive at all.

Post a Comment

Required fields are marked *


%d bloggers like this: