September 19, 2004

THANKS, ADMIRAL (via Tom Morin):

The Prediction Machine that Could (James Dunnigan, September 17, 2004, StrategyPage)

In early 2003, the Department of Defense proposed establishing a Futures Markets Applied to Prediction (FutureMAP) program. FutureMap was intended to provide some accurate predictions of events in the war on terror. This was to be done by allowing many people in the government to place "bets" on which way a future event would turn out..

FutureMAP is pretty much identical to our Prediction Market. In fact, we decided to create our Prediction Market after media and political opposition caused the cancellation of the FutureMAP program. It was cancelled mainly because some members of Congress, and the media, considered it trivial and doomed to failure. Well, that was certainly a self-fulfilling prophesy. Unfortunately, most people in the Information War and financial markets business agreed that FutureMap would probably work.

So late last year we put together our Prediction Market. It’s been in beta test for nine months. And it works. So far, over 150 events have been submitted. 45 of them have reached a conclusion, and 42 of them were accurate. That’s 93 percent accuracy. Actually, it’s a little lower than that. If you take into account accuracy over every day successful item was in play, the accuracy rate is 89 percent for the ones that were correct in the end, and 84 percent overall (including the items that were not correct in the end). Sure beats flipping a coin. We still have over a hundred predictions waiting for either an event to occur, or a time limit to be reached. Some of them are quite interesting. Anyone can propose a prediction (after registering with a valid email address, which helps keep out the trolls and bombers). A StrategyPage staffer checks each proposal for spelling, grammar and logic, and then lets it go live.

The Prediction Market has accurately predicted the outcome of terrorism, military and political events all over the planet.

It's long past time to shut the CIA's doors.

Posted by Orrin Judd at September 19, 2004 9:15 AM

Thanks for the pointer, OJ (and Tom). I visited the page and there are _very_ interesting market predictions, most of which agree with my own estimates.

Posted by: Bruce Cleaver at September 19, 2004 10:09 AM

As a big fan of Asimov's Foundation, why won't any of these people admit that (AGAIN) science fiction was as effecively predictive as their "MAPS"...

which are, in effect, Hari Seldon's psychohistory.

As an aside, if you mix a chess master's capabilties with a well informed mind - Drucker/VD Hanson type - you don't need as many people.

Posted by: BB at September 19, 2004 11:00 AM