|Aug. 7 — While Pentagon official John Poindexter was busy last week trying to defend his agency’s plan to set up an online terrorism betting scheme, another site was taking virtual wagers of a different kind.|
AT NEWSFUTURES.COM, the question was whether
Poindexter—a national-security adviser during the Reagan
administration—could keep his job at the Defense Advanced Research
Projects Agency following the outcry over DARPA plans to set up a
futures market in Mideast terror acts. The bettor’s answer: Not likely.
They were right. The Pentagon killed the market the day after it was publicized; by the end of the week, Poindexter had resigned. But the idea behind the futures market was neither as outlandish nor as unique as it might seem. Market systems can be astonishingly accurate about future events. The 15-year-old Iowa Electronic Markets lets people buy and sell “shares” in presidential candidates. Its final share prices—measured on midnight before Election Day—have consistently outperformed large national polls in predicting election results. And at least three Web sites—Tradesports.com and betonsports.com, as well as NewsFutures.com—offer the chance to bet real or virtual money on current events. Probably the best-known, NewsFutures, which was founded in 2000, had a new market this week with the announcement that Arnold Schwarzenegger will run for governor of California. (By midday Thursday, a slight majority had bet on victory for the muscular movie star). Site visitors can also bet on everything from the probability of catching Saddam Hussein to the chances that J. Lo and Ben Affleck will marry.
NEWSWEEK’s Jennifer Barrett spoke with NewsFutures cofounder and president Emile Servan-Schreiber about online betting, terrorism, the 2004 White House race—and whether futures markets can accurately predict a news event’s outcome. Excerpts.
NEWSWEEK: Last week, you opened two virtual futures markets on the probability that the Pentagon’s proposed futures market on terrorism would start registering traders as planned and that John Poindexter would keep his job. How were the odds early on?
Emile Servan-Schreiber: We
created, within one hour of the news breaking, a market on whether the
project would actually open on Aug. 1 as planned. That market collapsed
within about two hours after [Deputy Defense Secretary] Paul Wolfowitz
showed up in Congress and decried it as a stupid idea. The other market
was whether Poindexter would remain at DARPA. That ran for a couple
days and converged pretty rapidly to about a 60 percent probability
that he would lose his job—this was before he announced he’d resign.
Once he announced that, the probability went up dramatically, of course.|
What was your feeling on the Pentagon’s Policy Analysis Market?
That was terrible public relations, even though the guys behind it were incredibly smart.
Would you ever consider having a market on the probability of terror acts?
No, we stay away from things like how many people will get killed this week. But anything else is newsworthy. We didn’t shy away from the [Washington, D.C.] sniper case [last fall]. It was on everybody’s minds already.
You had virtual wagers on when the war in Iraq would begin. How many people correctly picked March?
We had the markets since October of last year on whether the war would start by the end of that year. Come December, the possibility that the war would start in February was at 70 percent. But the contract for whether it would start in February then collapsed quite quickly after the French started obstructing the idea in the United Nations. The contract for a March start went up in early February when it became apparent it would not happen in February. Most people knew by then it would be March.
It wasn’t so much the probability that was really the interesting thing, though, but the reaction we got on each side of the Atlantic. The French thought it was totally indecent that we were having this market—betting on a war in which all these people would be dying, whereas the Americans had no reaction to it.
Have there been any markets on news events that you removed because of protests?
When the sniper case occurred, we had a market on it—not about how many people he would kill but when he would get caught. So each week we had a market asking, will he get it caught this week. And the reaction was very strong in America this time. People were very upset. Some of our staff almost quit because they live in the area, and it was too close for comfort. But we kept it up. And on the French side, they just enjoyed a good market.
How many registered users do you have?
We have about 50,000 to 60,000 registered, but about 15,000 are really active users.
Where are most of your users from?
There are mostly Americans and French. There are some Hungarians—we offer a Hungarian version, too. Anyone can play for free. You get play money when you play. Then each month, there are prizes to be won. So if you have enough play dollars, you can bid on it. If you are ranked No. 1 in being right, you can bid this month or wait and bid for the next month’s prize.
How accurate are your markets in forecasting events?
That is the thing about the accuracy of this. Of course, if you analyze a large amount of markets, you will find, for instance, that [most] of the events predicted to have 80 percent probability of happening actually do happen. Remember when the Russians said the Mir space station was going to be brought down? Month after month, they kept pushing it back. Then finally, they said Feb. 15. We decided to give them an extra month and bet on March 15. And it happened almost exactly on the date we selected [it came down March 22, 2001].
What about the sniper case?
When you do a market on the sniper case, all you could react to was whether the police were making progress or not. Then the interesting question is not accuracy, but it is measuring the reaction of people to a particular issue. It was more about whether the people feel the police are making progress. When they could see on CNN that the white car was found, then the numbers shot up on minute-to-minute basis that he would be caught.
So the movement in your markets can be used to gauge how closely people are following the news and how they are interpreting it?
Yes, the market on Osama bin Laden’s capture is a good example. Just after the September 11 attacks, we opened a market on whether Osama bin Laden would be captured by the end of the year. That was the central aim of the campaign in Afghanistan. By following the markets, you could follow the progress of the war. You could see how the markets reacted to the start of the bombing, for example, or to the fall of Kabul or Kandahar. Each time the probability for his capture went up. People take into account many different things beside hunches.
If the market shows no reaction, does that reflect on the credibility of the news reports or sources?
Well, there was a story about the troubles of J. Lo and Ben Affleck circulating on the Web this week. We have a market on whether they will get married before Demi Moore does [75 percent probability that they will, as of Thursday] and the number has not moved at all after the reports. I think people dismissed that as gossip.
You don’t see it as, well, a bit pointless to bet on the probability of these sorts of things?
To me, it is the future of journalism. This was a quasi-journalistic venture. I was a journalist before I cofounded this company. With markets, people have a voice to say what tomorrow’s news will be. If you think about it, you [the media] keep feeding these great masses of educated, well-trained brains with the best analysis and information and then you don’t get the feedback, the interaction.
Oh, trust me, we get feedback.
But it’s not aggregated—and not enough to make it measurable.
Fair enough. How do you decide what stories are worth betting on?
It’s pretty much like you decide what to write about. We ask: do we think it has legs? Is it worth the trouble of creating a market? If you create a market and no one plays it, then you feel like a fool. We try to find the stories that can capture people’s attention. We look at things on the front page or when someone [famous] makes a statement that something will happen or that they are going to do something, we make a market on whether it is true.
Do you ever hear from people, like Poindexter, whose words or futures are being bet on?
Yes, in France, we had market on a [Parliament] member who wanted to impeach President Jacques Chirac. We created a market on whether [the politician] would get enough signatures, and it was a pretty hot issue for awhile. He actually called me up and said that it was great we were doing this.
Was there a high probability in your market that he would get the signatures?
[Laughs.] Actually, the probability was not so high, but accurate. He failed. But he came close.
Any predictions on the Democratic presidential candidates?
At this point, it looks like it is really between [Howard] Dean and [John] Kerry, and [Joe] Lieberman is behind them ... Everyone else is way behind. An interesting thing is that the number of people—the volume—on Dean is much higher. It is as if many people are waiting for Kerry to hit his stride.
And what are the chances that any of them will unseat Bush?
It is too early to make any real predictions there on whether the race will be like the one in 1992 [when Democrat Bill Clinton won] or the one in 1984 [when Republican Ronald Reagan won].
We’ll check back in with you next year.
Well, one thing we know is that it will be very interesting. Right now, Bush’s probability is at 55 or 56 percent, so the people’s analysis is that it’s his election to lose.
© 2003 Newsweek, Inc.