« Markets in everything: randomized pizza prices | Main | The new html edition of Tyler Cowen's Ethnic Dining Guide »
99 10 Red Balloons
Earlier this week DARPA, the Defense Advanced Research Projects Agency, moored ten, 8 ft red, weather balloons in undisclosed locations across the United States.
The DARPA Network Challenge offered a prize of $40,000 to the person or group who first identified all the locations.
The MIT Group which won the challenge used a clever pyramid incentive scheme. Each balloon was worth $4000. The person to identify the location earned $2000. The person who invited that person to join the MIT group got $1000, the person who invited the person who invited the person who located the balloon got $500 and so forth (any money not distributed in this way was given to charity.)
The incentive scheme meant that contestants not only had an incentive to find balloons they had an incentive to find someone who could find balloons (or find someone who could find someone who could find balloons and so forth).
Incredibly, the MIT team located all ten balloons in just under 9 hours! The challenge may seem frivolous but in fact is a great example of how prizes and network technologies can combine to collect and use highly dispersed information--a problem of very general interest and relevance.
Posted by Alex Tabarrok on December 10, 2009 at 07:37 AM in Economics, Science | Permalink
Comments
The counter-insurgency and law enforcement applications are obvious. Substitute "Osama bin Laden" or "James 'Whitey' Bulger' for "Red Balloon" and you can solve a case in 7 hours (give or take).
Posted by: MostlyAPragmatist at Dec 10, 2009 8:36:29 AM
I thought the same thing MostlyAPragmatist. Though why wouldn't a simple bounty be better for bin Laden? I could see maybe trying to get a picture of insurgents operations by using such a scheme, and paying multiple independent searchers and searcher-finders to get a picture of things as they happen for future use.
Posted by: mike kenny at Dec 10, 2009 8:56:34 AM
So, maybe Osama Bin Laden is going to be classified as a balloon and we will be setting up a pyramid scheme in Pakistan to find him?
Posted by: Bill at Dec 10, 2009 9:06:29 AM
The key thing is the publicity that the DARPA challenge got. Without this much publicity, it is unlikely that sufficient people would have been there to solve this problem. That's the key issue.
Posted by: sa at Dec 10, 2009 9:06:41 AM
What they probably also need to do is flood Pakistan with cellphones to make the network work.
Posted by: Bill at Dec 10, 2009 9:09:11 AM
Would have been a more realistic experiment if they kept moving the balloon.
Posted by: Bill at Dec 10, 2009 9:11:40 AM
they should try this with neutrinos!
Posted by: babar at Dec 10, 2009 9:19:54 AM
So if the MIT group lost they would be out quite a bit of money. I doubt this would work on a large scale with many different projects if individuals had to front the money. (The difference between this and the Netflix challenge, etc. is that there people invest their time mostly and it is recreation for them. Here presumably they have to pay the guys who find the first few balloons even if they don't win the competition. I doubt universities, etc. would start funding a lot of these competitions which may come up empty all the time.).
I know this is similar to any first-past the post competition (elections, scientific research, etc.), but I don't think you are going to be able to scale this up too much without building an organization committed to funding these risky, non-profitable endeavors. If you are doing that why not just open a government agency like the DARPA.
Posted by: Mo at Dec 10, 2009 9:46:36 AM
Mo makes a good point, I checked the MIT rules again and see that all payments were also conditional on the MIT team finding all 10 balloons first so in fact they were not risking 40k.
Posted by: Alex Tabarrok at Dec 10, 2009 10:01:06 AM
So really, Mo made a bad point.
Posted by: Gabe at Dec 10, 2009 10:07:41 AM
The Military Industrial Complex doesn't want to find Osama. You'd have to be ignorant and naive to believe otherwise. You are truly a denialist if you can see the statements made over the years and still think we are desperately trying apprehnd Osama.
http://www.youtube.com/watch?v=nb3fesK-4S4
Posted by: Gabe at Dec 10, 2009 10:12:20 AM
I like Alex's use of pictures in his posts.
Posted by: Millian at Dec 10, 2009 10:50:33 AM
The problem with finding Osama bin Laden with this is that your enemy would have the incentive to give you false information, depleting your resources in the search, and probably ambushing you.
Nice try though.
Posted by: Bill at Dec 10, 2009 10:53:55 AM
The other problem with this is that it relies on a friends and family network to assist in the search. I am sure the MIT folks communicated widely, and to trustworthy links. Think about the Pakistani who takes a risk in asking a friend to help him find Osama. What if the link that you seek to enlist is untrustworthy.
Could be a bad whipping, or worse, for you.
Posted by: Bill at Dec 10, 2009 11:11:05 AM
The balloons had no incentive to move or rehide themselves.
Posted by: CK at Dec 10, 2009 12:02:36 PM
I like that they solved the distribution problem at the same time as the incentive problem.
Posted by: Rick Weber at Dec 10, 2009 12:35:04 PM
There is little to be learned about the hunt for Osama, but plenty to be learned about the hunt for plainly visible, stationary or slowly moving objects. Pilotless drones, for instance. I bet DoD is rethinking just how secure their fleet of them is, based on this experiment.
Seems any insurgent group with wide enough popular support could locate all of them in the air over their territory at any given time. They can be taken out from the ground with laser-guided weapons. Not too much of a stretch, given what seems to be available in the arms bazaar these days.
Posted by: Bob Knaus at Dec 10, 2009 12:39:38 PM
i think this doesn't offer much against ObL. I, personally, think this is far scarier for citizens of a high-tech, wired country. say, americans. Replace "balloon" with "outstanding warrant: 1999 mazda 3 plate # such-and-such"
Posted by: farmer at Dec 10, 2009 1:13:00 PM
The entire purpose of this exercise is PR so DARPA can keep their budget stable in an era of flat or declining defense budgets. It has become a zero sum game and all the services and departments have to fend off each other for money. Operation and, especially, personnel costs are rising (as Dems and Repubs try to outdo each other with ever more generous benefits for veterans) - so even a stable overall budget means declining R&D and procurement spending. DARPA is fighting to protect their rice bowl.
Posted by: jim at Dec 10, 2009 3:13:44 PM
Since Osama's followers would not knowingly disclose his location, and to employ the social network approach of MIT with a twist, what you should do is create an urban myth that whoever sees Osama will be blessed in his presence and will be forever saved. And, don't forget the virgins to spice it up.
Now see if he can hide.
Or, we could paratroop in some Hollywood paparazzi.
Posted by: Bill at Dec 10, 2009 6:45:35 PM
There is little applicability to finding Osama himself, because Osama would represent one great big very well-hidden balloon rather than many little balloons. However, this would be highly applicable to taking out the "middle management" of al Qaeda or the Taliban (more needles, less haystack).
Something similar was used once upon a time to flush out Pablo Escobar: he personally was very well protected and supported, but his associates and entourage were not. The latter were gradually "neutralized", starting from the lowest levels of the pyramid and working upwards, until the whole structure was undermined and Escobar himself was defenseless and on the run and finally killed. Evaluating the roaring success of the larger objective of actually ending the drug trade itself is of course an exercise for the reader.
Finding Obama would be like winning the lottery; finding dozens of smaller fish would be more like an Easter egg hunt. The incentives for participation are different, so the prize contests have to be set up differently. The security precautions taken by the targets are also different, amounting more to safety in numbers and crossing your fingers in the latter case rather than truly high security.
Posted by: anonymous at Dec 10, 2009 8:28:57 PM
I can't believe I mistyped "Obama" in the last paragraph. My deepest apologies.
Posted by: anonymous at Dec 10, 2009 8:32:33 PM
I accept your apology.
Posted by: obama at Dec 11, 2009 1:22:10 PM
Bill - the problem isn't that your links are untrustworthy, it's that the counter-incentive is very high. Weather balloons aren't going to get revenge on you and your family for ratting them out.
Posted by: Anthony at Dec 11, 2009 4:14:16 PM
Wow, surprised Amway wasn't all over that.
Posted by: anon at Dec 11, 2009 5:20:43 PM
Thank you for another great blog. Where else could I get that kind of information written in such a perfect way? I have a project that I am presently working on, and I have been on the look out for such info.
copyright Ukraine
Posted by: Derk at May 5, 2010 8:45:00 AM
