Prediction Machines
Dean
One of the most proven methods of making predictions about future events is to set up mock "stock markets" wherein yes/no questions are asked, and people are asked to buy shares in the "yes" or "no" stock. The theory is that while individuals will vary in their individual accuracy, they'll tend to be more honest about what they think will happen (rather than what they want to believe will happen) if they are putting small amounts of money on the line. It doesn't have to be a lot, a few dollars only. Coupled with this is the theory that while any one individual may not be particularly accurate, in large groups, the consensus wisdom, the group intuition, is usually correct.
Such mock stock markets have proven over and over again to be extraordinarily accurate at predicting events. Even fairly small ones work well. For example, the folks at StrategyPage run a fairly strong predictions market that's well worth checking out. They publish all their results, including both when the market's been wrong and when it's been right, and boast an impressive 84% success rate over market lifetime, and 94% accuracy by the time any given stock "closes" (i.e. when trading on any question stops). You can see all the predictions their market has made, both when they were right and when they were wrong, right here. Some of them are rather startling in their specificity.
Sadly, while science has proven repeatedly over the last couple of decades that these market games are astonishingly accurate, our government still is not using this tool to help them make predictions in the War On Terror. The Pentagon announced plans to carefully integrate a tool like this into their war planning and their efforts to protect Americans, but irrational political backlash forced them to shut the nascent program down. It's a real shame. James Dunnigan has an excellent article about how and why the fledgeling Pentagon program was shut down. It's worth a read, if nothing else as a lesson in what happens when people's political views become kneejerk and reactionary.
Related Posts (on one page):
- Manipulation
- Ranking The Pollsters
- Prediction Market Update
- Predictive Markets
- Prediction Machines
There are very easy ways to structure something like this so that no one's making any real profit, and to make sure there are no conflicts of interest. To throw away a potentially lifesaving tool just because it makes you queezy.... does that even make sense?
Of course, they're generally always more willing to conclude that some particular life saving measures are worse than death when it's not their life on the line and the whole thing is purely hypothetical, but that's just human nature.
This has been accurate in fourteen elections since 1988. It is run as an academic project by the Business School at Iowa University.
The public is invited to participate.
It is not gambling, it is trading. They're called forward contracts.
Predictive markets are used by business to predict market changes. If DARPA had been allowed to run their predictive market, the intelligence in Iraq and other places may have been better.
http://www.tradebetx.com