Prediction markets front and center at relaunched tech magazine

Remember The Industry Standard? It was one of a spate of magazines that first exploded, in the sense of growing very quickly, and then exploded, in the sense that they blew up in smithereens, during the dot-com boom and bust.

The International Data Group announced today they were relaunching the magaine as a web-only publication. it's hardly the first to do so, but there's a twist.

The Industry Standard, once known as The Bible of the Internet economy, with a new publishing model that includes editorial content, a community-driven prediction market and social networking components.

Dovetailing with the editorial content is the prediction market, a way of betting on the outcome of future events, [Derek Butcher, the online publication's vice president and general manager] said.

The prediction market uses community input and proven algorithms to forecast events in the technology industry, according to the statement. Registered users can use mock currency to place "virtual bets" on the outcome of these events.

"For example, a prediction might state, 'Apple will ship 10 million iPhones by the end of 2008,' or 'High-tech venture funding will decrease by 15% in Q2 2008,'" according to a company statement. "As the community members place bets on a given prediction, the resulting market price of the prediction represents the community's consensus as to the probability of that event occurring."

Butcher said he doesn't know of any other media sites that prominently feature a prediction market operating in conjunction with editorial content.

We're interested to see how prediction markets do there, because we at Spectrum have a high regard for them.

Indeed, we touted the virtues of prediction markets just last September ("Bet On It"). It's clear that they do a good job of collecting the wisdom of the crowd, which often is collectively wiser than any individual poll or pundit. (So much so, in fact, that they are being increasingly adopted by corporations, which was the focus of the article.)

Dan Gross, a senior editor at Newsweek, seems to have ignored all that in favor of the cachet of offering a contrarian opinion in a long but uninformed segment of the national public radio show, On The Media.

Discussing the premier prediction website for political wagering, TradeSports, in Dublin, Gross said of the $40 million bet there, "when you compare it to the activity in the real stock market," it was "a tiny amount." It's hard to know what motivates such a comparison. The $130,000 you'd pay for a Jaguar XJ Convertible pales by comparison to the total annual revenue of the Ford Motor Company, but that doesn't mean it isn't an expensive car.

Gross's most egregiously wrong thoughts about prediction markets, though, are contained in a single soundbite:

"It's clear they are just reacting to conventional wisdom rather than setting it."

Prediction markets aren't supposed to set the conventional wisdom, they're supposed to exceed it by, in effect, giving the smartest opinions the greatest voice. People who know best back their knowledge with their dollars. Uninformed wagers that move the market away from its best guess merely attract more smart money, swinging things back again. Unfortunately, as is increasingly common these days, On The Media host Brook Gladstone asked about none of this in a 6 minute, 35 second radio segment.

Referring specifically to wagers on the U.S. presidential primaries and market predictions of the eventual party nominees, Gross said:

"What you see is the action of the prices really following what happens in the polls and what happens as the tallies are counted."

This is the "just" in Gross's "just reacting to conventional wisdom," and it's the most wrongheaded thing he said in an interview where that wasn't an easy choice to make. In a word, as the philosopher Homer Simpson would say, "Doh."

Of course experts are going to use all available data in making their assessments, and as new data is available, they're going to recalculate their predictions. What Gross needs to show is that the markets move in lock-step with the latest information in a some completely mechanical way. Otherwise, they're reacting, but not merely just reacting.

For example, if a savvy political observer, considering Sen. Barak Obama's impressive win there, looks deeply into the South Carolina exit polls, she might notice that the senator did particularly well in certain demographics that are not as strongly represented in the next race. She might, then, shade her prediction away from the conventional wisdom, which doesn't look deeply into the exit polls, and bet that Obama won't do quite as well as others expect.

Gross looked at the way Obama's share price - a wager that Obama would win the Democratic nomination - went up after he won the Iowa caucuses, and went back down after his main rival, Sen. Hillary Clinton, won in New Hampshire. This is a surprise?

Gross acknowledges how well the prediction market InTrade did in 2004 state elections, getting nearly all of them right. But he says that the same claim can't be made about polls, "because of their margin of error." Well the fact is, every poll got more than a few races wrong, even beyond the margin of error.

"They're more accurate than the polls on the day of the election after 8 or 9 months of campaigning... they are completely inaccurate if you want to know, today, who's going to win the election in November."

But that's a classic straw-man argument. The issue is not, can prediction markets make perfect predictions today about an election 9 months down the road. The issue is whether they can do so better than the polls and better than the pundits. Gross sidesteps this obvious question, and the obvious answer - yes they can - and On The Media host Gladstone unfortunately lets him get away with it.

Luckily, few seem eager to join Gladstone and Gross in embracing the contrarian stance. Rather most seem interested to see more of them, such as the ones at the resuscitated Industry Standard. We wish them good fortune.


Tech Talk

IEEE Spectrum’s general technology blog, featuring news, analysis, and opinions about engineering, consumer electronics, and technology and society, from the editorial staff and freelance contributors.

Newsletter Sign Up

Sign up for the Tech Alert newsletter and receive ground-breaking technology and science news from IEEE Spectrum every Thursday.