Place your bets!

[Jim F.]

This is the dumbest thing I have ever seen the government attempt to do...

http://www.cnn.com/2003/US/07/29/terrorist.market.ap/index.html

Apparently, starting this Friday, the Pentagon is going to open a futures market for world events. Have a feeling that this is the year that Saudi Arabia's ruling family is finally overthrown? Place a bet and see what happens!

Even the theory that they're using to justify the program is idiotic.

[Timemaster Tim]

How outlandish. Has the United States become so stock market centred that even intelligence gathering must go through a commodities and futures market? I can imagine this as a Star Trek episode where they come across an alien race who spend all their time doing nothing but betting.

[Shawn Metcalf]

The web site (http://www.policyanalysismarket.org/) for this program is fantastic. Always wanted to legally wager on the continued survival of the Jordanian monarchy, and be paid for it by the U.S. government? Now you can!

[Mark Asher]

I thought this was a parody at first.

Who would want to make money off this? If you successfully predict an event, you'll likely be investigated by the FBI and CIA.

[Ben Sones]

This is surreal. It's like the Dead Pool (Remember? With Dirty Harry?), only sponsored by the US government. Which has apparently been taken over by producers from Fox.

[Rywill]

My God. Who could have possibly thought this was anything even vaguely similar to a good idea?

[Tyjenks]

Wait a minute! This comes from our present administration and, by the theories set down in Political Timology, since we are supporting efforts in Iraq we are supposed to support this as well. All of you hawks pipe down! Pipe down, I say!

[Anders Hallin]

Making terrorism etc. profitable for more people? Sounds like a great idea to me!

[Brandon Clements]

The plan's now cancelled, according to Jim's link.

[Jim F.]

I'm glad they cancelled it. It's just amazing that people at DARPA sat down, formulated this plan, and considered it a great idea.

If I didn't know it was true, I'd think that this was some joke that someone was circulating through the Pentagon e-mail system that accidentally got added to the budget.

[DennyA]

I wonder if they were hoping that terrorists would be tempted to bet on their plans?

IE: An increase in predidctions of terrorist attacks alerting the gov't that something was up.

Or Predict that Quiraq leader Mohmmar Jones would be assassinated Tuesday... After he is indeed assassinated on Tuesday, the CIA/FBI/NSA shows up on your doorstep.

[Anders Hallin]

I wonder if they were hoping that terrorists would be tempted to bet on their plans?

IE: An increase in predidctions of terrorist attacks alerting the gov't that something was up.

Or Predict that Quiraq leader Mohmmar Jones would be assassinated Tuesday... After he is indeed assassinated on Tuesday, the CIA/FBI/NSA shows up on your doorstep.
It might very well work, for predictions. The question is, however, if terrorism and the like would increase because of it. The risk of which is "slightly" too high to go through with it.

[bago]

Think about it. Markets are quick at finding information, especially when there's many varied people with money on the line. Aggregate information filtering and valuing. The downside of the idea is that it would be easy to make false positives, causing runs and shorting certain events. That and panic mentality.

[Jason McCullough]

I'm glad they cancelled it. It's just amazing that people at DARPA sat down, formulated this plan, and considered it a great idea.

If I didn't know it was true, I'd think that this was some joke that someone was circulating through the Pentagon e-mail system that accidentally got added to the budget.

It's not completely crazy, just incredibly tone-deaf PR. You'd think the "experts" would have noticed, though, that the kind of things that make the Iowa Electronic Markets so good at predicting political results - tons of people, everyone's a price-taker, etc. - aren't present here.

[James Gutierrez]

I agree with Jason. It's not that crazy, just a bit bizarre that it was presented like this complete with the public web site and everything. There's actually merit to the idea of using market structures as tools for aggregating and filtering massive amounts of information. Just do a Google search for "decision markets" or "information markets". Here are
2 articles:

http://hanson.gmu.edu/infomkts.html

http://www.newyorker.com/talk/content/?030324ta_talk_surowiecki

Frankly, this sort of unconventional, long shot research is exactly what DARPA is supposed to be doing. As Jason mentioned, there are problems with the approach, but I suspect that this research will continue with a much lower profile and in a few years may result in another tool for vetting intelligence data.

[Mark Asher]

Most markets results are driven by profit. The problem I see with the now-cancelled DARPA market is that people might try to manipulate results for other reasons, especially if people thought the government was being responsive to results.

[Shawn Metcalf]

Amazingly enough, Instapundit (www.instapundit.com) is still gung-ho about this, and is decrying the "idiots" who managed to get this thing killed.

True to form, he's mentioning only the two Democratic Senators who first made a stink about this. Last night on the radio I heard a wide array of Republican Senators like Bill Frist, Pat Roberts, and others verbally beat this thing like a pinata.

[Jim F.]

Amazingly enough, Instapundit (www.instapundit.com) is still gung-ho about this, and is decrying the "idiots" who managed to get this thing killed.



After reading the links provided on the Instapundit site, I'm even more glad they cancelled this.

The idea of using markets to determine world events is that, at its heart, it assumes that everyone is an expert on the subject and has all the necessary information to make an informed decision.

If the price of North Korea Nukes South Korea before July 31st options shoot up, what does that mean? Does that mean that the North Korean inner circle knows it's going to nuke the South, so it is buying futures so it can cash in on it? Or is it more realistic to assume that the recent report on Dateline that the U.S. is refusing to make a deal with North Korea is causing the general public to pick up those options?

Even the oil futures market that is pointed to as being a great predictor of events is more grounded in CURRENT news than future news. Oil prices didn't shoot up the week before the World Trade Center bombings... they went up the days after. Why? Because people didn't predict the event, they predicted the following events that were triggered by that event. Prices shot up because it was assumed that the bombings came from an oil producing Arabic nation. Had the U.S. acted on this "knowledge", Iraq and Saudi Arabia would have been the chief targets instead of Afghanistan.

The system is so inherently flawed, it just doesn't make sense to me to use it as a basis of, well, anything when it comes to world events. Even after reading the suggested literature, I still can't see it as a viable and reliable source of predicting political events.

Until the information is publically available, it doesn't hold much sway on an options market. The only time private information influences public prices is when a known individual does something and it shakes up the market. If Warren Buffet buys a bunch of corn futures, people will notice and assume he knows something. Before you know it, corn futures are worth a fortune. Self fulfilling prophecy and all that jazz.

Oh well, I'm obviously not an economist, so I'm sure I'm missing why PAM is such a good idea.

[Daniel Morris]

There's actually merit to the idea of using market structures as tools for aggregating and filtering massive amounts of information.

And there's certainly merit to the idea of creating a market structure that invites the aggregating and filtering of information regarding the prospects of terrorist action, particularly by "geographic market."

This is simply because globalization has brought so many investors into an intimate awareness of regional geopolitics. OmniCorp does all kinds of due diligence before risking billions of dollars in a new market -- it's conceivable that they might hear something someday that should rightly set off alarm bells...bells that might be reflected in DARPA's decision market.



Frankly, this sort of unconventional, long shot research is exactly what DARPA is supposed to be doing.

And you have such far-fetched projects as the Internet to show for it.



I suspect that this research will continue with a much lower profile and in a few years may result in another tool for vetting intelligence data.

Count on it.

[James Gutierrez]

The idea of using markets to determine world events is that, at its heart, it assumes that everyone is an expert on the subject and has all the necessary information to make an informed decision.


Actually, it doesn't. That's kind of the whole point behind decision markets: collective information aggregated from a large number of relatively uninformed individuals is often more accurate than the opinion of any single expert.

From Slate: (http://slate.msn.com/id/2086315/)


Oddly, the hope of the PAM might have been the ignorance of investors, rather than their intelligence. Policy market day-traders who don't speak Arabic or have access to classified information wouldn't necessarily make worse bets than the professionals who spend their days sifting through Al Hayat and humintel reports. This is a seeming paradox called the "dumb agent" theory—one of my esteemed predecessors in this box explicated it a few years ago. Walk up to a traveler waiting in an airport, ask how many minutes late the plane will take off, and you're likely to get a wrong, uninformed answer. Ask 75 more of your fellow passengers the same question, and you'll get 75 more similarly wrong, uninformed answers. But throw them all together and take the mean, and you're likely to get something pretty close to the right answer.


Here's another Slate article (http://slate.msn.com/id/1001932/)

[Jim F.]

Once again though, the thinking doesn't make sense to me on predicting events. Pridicting prices, yes, I'll agree with that completely.

If you put 100 people in a room with an item in a sealed box and ask them "How much would you pay for this box?" you can get a good estimate on how much the box will sell for. But ask them "What's in the box?" and they won't be able to predict it without relevent information.

The examples people are giving are too general. DARPA is looking for specifics here. Taking 100 people and figuring out how late a plane will take off is one thing. But the stuff DARPA is looking for is stuff like "Will the Saudi Government by overthrown in December?". Put 10,000 uninformed people in a room and you're going to tell me that they can predict that the Saudi monarchy is going to be overthrown in the Summer of 2004?

[Jason McCullough]

Weren't they going to use just a small group of experts to do the trading? Though I guess that fixes one problem by creating another.

[chet]

Small groups of qualified people, trading next to no actual dollar amounts. So it shoots down most of the objections raised here, but does show people's knee jerk reaction winning out over a response based on the actual material.

Chet

[Jason McMaster]

This idea was actually quite fascinating and I'm pretty sad that it didn't go through.

[Shawn Metcalf]

Folks who want to try this out on a private level can do so here (http://us.newsfutures.com/home/home.html). This has all kinds of future speculation, including whether or not John Poindexter will lose his job over this (http://us.newsfutures.com/group/group.html?groupId=3209). The site seems to be doing a good job testing out the theory, without the ghoulish prospect of directly cashing in from an assassination or terrorist attack.

[Jason McCullough]

It's truly amazing that Poindexter is still able to get a job in Washington. A felony conviction just doesn't cut it anymore.

[Daniel Morris]

Chet wrote:


Small groups of qualified people, trading next to no actual dollar amounts.

This is correct. The DARPA market was/is intended as a "wargame" and not any kind of serious economy.

[MikeJ]

Folks who want to try this out on a private level can do so here (http://us.newsfutures.com/home/home.html).

Their current odds for various democratic candidates add up to 122%. Does that mean you can't lose if you bet against all of them? One would hope there would be opportunists in these markets who would enforce a little consistency.

[Woolen Horde]

What's kinda sad is that I've been reading a lot of editorials bashing (justifiably so) this idea, but that DARPA gets smeared along with Poindexter, when in reality, it should be Poindexter who should be canned. The guy is just Numero Uno in coming up with creepy ideas.

And all of the news stories I've read don't mention some of the truly great and revolutionary things that DARPA came up with, you know, like that little itty thing called The Internet and that little itty technology the Air Force uses called Stealth. So DARPA comes off like some evil organization whose only purpose is to spy on Americans and take bets on Middle East turmoil.

[bago]

The visceral reaction against markets is really quite suprising.

[XPav]

And all of the news stories I've read don't mention some of the truly great and revolutionary things that DARPA came up with, you know, like that little itty thing called The Internet and that little itty technology the Air Force uses called Stealth. So DARPA comes off like some evil organization whose only purpose is to spy on Americans and take bets on Middle East turmoil.

Lately DARPA has been pushed toward short term useful projects that directly help current problem, rather than the long term projects pie in the sky stuff like the internet.

[Jim F.]

Small groups of qualified people, trading next to no actual dollar amounts. So it shoots down most of the objections raised here, but does show people's knee jerk reaction winning out over a response based on the actual material.

Chet

My 10,000 people number came directly from the PAM website. The site has ben wiped clean, so there's nothing to back me up on this though.

The members weren't going to be recruited. It was going to be an open signup with 1,000 people allowed in between Aug 1st and Oct 1st. After Oct 1st, the number was going to be increased to 10,000 until Jan 1st 2004. After that, there were no details on how many people would be allowed in. I attempted to click on the "Become a member" link, but since signups weren't being accepted until Aug 1st, I wasn't able to see if they were asking for, you know, qualifications.

The entire system was prediction for profit. PAM was going to be the bookie with the other members being the betters. No government money was going to be used to support the system, it was all money from the losers being used to pay the winners. That just sounds sleazy to me.

The only thing I can think of that they were shooting for here was a way to distill down 10,000 opinions into 1 unified opinion. "OK, 63% of our experts think that Saddam is hiding in Syria. We should start searching there". If that's the case, why make it based on a monetary system at all? When you introduce money into the equation, you realy make it seem, as I said, sleazy.

[Kalle]

Hey, was it only going to be open for Americans?

I guess my desire to earn money at the expense of other people's suffering would never have come true even if they hadn't shut the thing down. :(

[bago]

Kalle: I take it you've never heard of Life Insurance then?

[Kalle]

Yeah, sure. But the only way I'll benefit from life insurance is if people I know end up dead. And that would be mildly disturbing. Nothing that heaps of cash wouldn't fix, but still.

This DARPA plan could have me earning money off the suffering of complete strangers, buffering me from all the emotional and moral considerations.

[Woolen Horde]

Good News! NBC just reported that Poindexter is OUTTA THERE! Even Rummy was apparently miffed at the PAM.

Poindexter's Total Information Awareness plan that would have kept files and dossiers on every American citizen and then the PAM was enough to get his ass canned. Rejoice citizens, rejoice! Now if only we can get Ashcroft's butt canned too. Then our civil liberties might breathe easier.

[Jason McCullough]

Whew, first good news in forever.

[bago]

I still haven't heard a good argument as to why the PAM is a bad idea...

[Anders Hallin]

I still haven't heard a good argument as to why the PAM is a bad idea...
That depends if you mean the limited, non-money method (or very little money) or something large scale.

[bago]

Shares paid out a dollar per.

[Anders Hallin]

Shares paid out a dollar per.
Seeing how markets can be tampered with, and indeed generate certain events, allowing a market like that to be too profitable might not be such a good idea.

[bago]

Well, you can only buy shares on events that are created and listed on the exchange, which is run by the intelligence service. It's not like you can make up your own events.


It's just formal bribery of people with their ear to the ground.

[Shawn Metcalf]

Shares paid out a dollar per.

What exactly does that mean? An analyst couldn't have bought something for a dollar per share and sold it for $99 per share? And couldn't they have bought multiple shares in an offered exchange?

The details on how PAM would have worked are sketchy, but either the amounts of money involved would have been a pittance, in which case it's hard to see how that would have provided much incentive, or it would have been serious money, in which case the original objections fully apply.