Trendsurfing: Prediction markets (The Times)
By David Rowan
You wouldn't help us out writing next week's news pages, would you? We were just wondering which supermodel you thought would be next to confess her Drug-Fuelled Sex Shame. While you're thinking, see if you can predict Peter Mandelson's next diplomatic gaffe and Lord Birt's latest unlikely job application. Oh, and here's a little request from the sports desk. What's the score going to be in today's England-Austria World Cup qualifier?
Don't mock - between the lot of you, your guesses will be far more accurate than you would believe. With a couple of million of you reading today's Times, your collective predictions stand an extraordinarily good chance of hitting the mark. That, at least, is the premise of a hot forecasting trend currently sweeping corporations and internet communities. The trend stems from the notion that a crowd's wisdom is greater than that of its individual members. So if a crowd predicts how various events are likely to turn out - well, it ought to be taken seriously.
The idea - popularised by the economist James Surowiecki, in his book The Wisdom of Crowds - has led to a boom in "prediction markets", where groups of speculators bet their cash (or their reputations) on forecasting certain outcomes. Whether it's the price of oil or the General Election result, these "markets" have shown an uncanny ability to beat pundits and pollsters. The markets work, says Surowiecki, because participants are driven by pride or profit to think effectively. "They are extremely efficient at tapping into the collective wisdom," he says.
One of the most famous, a university-run online exchange known as the Iowa Electronic Market, has thousands of members betting up to £280 that their instinct is correct. In the past five US presidential elections, its "market price" for the winning candidate's vote share, based on bets placed, averaged at just 1.3 per cent out. Professional opinion polls were roughly half as reliable.
Big business has jumped on board. Hewlett-Packard and Siemens have used in-house prediction markets to forecast sales; Goldman Sachs and Deutsche Bank have run them to predict economic indicators. Last month, Google announced its own internal prediction market - one where employees place bets on the timing of product launches and other things "of strategic importance". There is no cash involved, just a guesstimate of the chance of something occurring by a certain date. According to Bo Cowgill, the project manager, the process is producing "decisive, informative predictions" that are proving remarkably accurate.
The game can also be played for fun. The Foresight Exchange, an online "fantasy prediction market" (at www.ideosphere.com), allows users to stake their reputations on predicting the likelihood of anything from Prince Charles becoming king (currently at around 89 per cent probability) to World War Three breaking out by 2050 (a more reassuring 43 per cent). Or there is the Hollywood Stock Exchange (www.hsx.com), where you can predict film box-office receipts and stars' ever-changing market values.
Here, then, is a final challenge for Trendsurfing readers. Can you guess what fascinating trend I'll be writing about next week? E-mail in your thousands: I predict my deadline angst will magically disappear.
(The Times Magazine, October 8 2005)
. . . AND READ MORE TRENDSURFING COLUMNS HERE . . .
You wouldn't help us out writing next week's news pages, would you? We were just wondering which supermodel you thought would be next to confess her Drug-Fuelled Sex Shame. While you're thinking, see if you can predict Peter Mandelson's next diplomatic gaffe and Lord Birt's latest unlikely job application. Oh, and here's a little request from the sports desk. What's the score going to be in today's England-Austria World Cup qualifier?
Don't mock - between the lot of you, your guesses will be far more accurate than you would believe. With a couple of million of you reading today's Times, your collective predictions stand an extraordinarily good chance of hitting the mark. That, at least, is the premise of a hot forecasting trend currently sweeping corporations and internet communities. The trend stems from the notion that a crowd's wisdom is greater than that of its individual members. So if a crowd predicts how various events are likely to turn out - well, it ought to be taken seriously.
The idea - popularised by the economist James Surowiecki, in his book The Wisdom of Crowds - has led to a boom in "prediction markets", where groups of speculators bet their cash (or their reputations) on forecasting certain outcomes. Whether it's the price of oil or the General Election result, these "markets" have shown an uncanny ability to beat pundits and pollsters. The markets work, says Surowiecki, because participants are driven by pride or profit to think effectively. "They are extremely efficient at tapping into the collective wisdom," he says.
One of the most famous, a university-run online exchange known as the Iowa Electronic Market, has thousands of members betting up to £280 that their instinct is correct. In the past five US presidential elections, its "market price" for the winning candidate's vote share, based on bets placed, averaged at just 1.3 per cent out. Professional opinion polls were roughly half as reliable.
Big business has jumped on board. Hewlett-Packard and Siemens have used in-house prediction markets to forecast sales; Goldman Sachs and Deutsche Bank have run them to predict economic indicators. Last month, Google announced its own internal prediction market - one where employees place bets on the timing of product launches and other things "of strategic importance". There is no cash involved, just a guesstimate of the chance of something occurring by a certain date. According to Bo Cowgill, the project manager, the process is producing "decisive, informative predictions" that are proving remarkably accurate.
The game can also be played for fun. The Foresight Exchange, an online "fantasy prediction market" (at www.ideosphere.com), allows users to stake their reputations on predicting the likelihood of anything from Prince Charles becoming king (currently at around 89 per cent probability) to World War Three breaking out by 2050 (a more reassuring 43 per cent). Or there is the Hollywood Stock Exchange (www.hsx.com), where you can predict film box-office receipts and stars' ever-changing market values.
Here, then, is a final challenge for Trendsurfing readers. Can you guess what fascinating trend I'll be writing about next week? E-mail in your thousands: I predict my deadline angst will magically disappear.
(The Times Magazine, October 8 2005)
. . . AND READ MORE TRENDSURFING COLUMNS HERE . . .
<< Home