New Economist: Consensus forecasting? Can one trust the 'wisdom of crowds'...
リンク: New Economist: Consensus forecasting? Can one trust the 'wisdom of crowds'....
Sunday, September 18, 2005
Consensus forecasting? Can one trust the 'wisdom of crowds'...
The aggregation of information in groups can result in better decisions than could have been made by any single member of the group. So argues James Surowiecki in 2004 bestseller The Wisdom of Crowds, drawing in part on the early work of Sir Francis Galton. Surowiecki's work has inspired three Oxford University graduates to apply its premise to financial markets. A Sunday Telegraph story, How to profit from the noise of the rabble, informs us:
Stephan Bisse, 40, a former City trader at Goldman Sachs who completed an MBA at the university's Saïd Business School last year, has set up ConsensusView.com, helped by two young computer science graduates from Worcester College, Ed Smith and Sam Brightman.
Each day members of the public are invited to vote on whether they think any one of a series of share prices, indexes, currencies and financial futures contracts will rise or fall that day. Voters do not specify an actual forecast; they simply predict a rise or fall, and have to do so between 4.30am in the UK and the opening of the market they are voting on. There is the lure of a $1,000 (£550) prize for the best forecaster each month. The Oxford three hope to test the theory that crowds have wisdom, and if it holds water, create a business out of it.
The website is up and running, but has yet to reach critical mass. As the Telegraph story observes:
The key to all these efforts, however, will be the site's ability to attract voters and, ultimately, to forecast accurately the direction of markets. At the moment, two weeks after its launch, just 52 people are voting on it, but with a larger number the theory that a consensus view will tend to be right will be tested more comprehensively.
As of Sunday evening there were 212 registered users so far, so clearly the newsaper story has boosted the numbers. I encourage readers of this blog to sign up to ConsensusView.com; it's an experiment worth encouraging.
