Join 3,514 readers in helping fund MetaFilter (Hide)


"And here is a red balloon.. I think of DARPA and note its exact coordinates"
November 2, 2009 2:25 PM   Subscribe

"To mark the 40th anniversary of the Internet, DARPA has announced the DARPA Network Challenge.... The challenge is to be the first to submit the locations of ten moored, 8 foot, red weather balloons located at ten fixed locations in the continental United States. Balloons will be in readily accessible locations and visible from nearby roadways." Teams must register by December 1st and have two weeks to submit balloon locations.

Popular Science has a bit more. The Register offer some hacks. The Twitter will be here.
posted by jessamyn (108 comments total) 35 users marked this as a favorite

 
Team MetaFilter!
posted by CG at 2:26 PM on November 2, 2009 [2 favorites]


If you do a team metafilter I got an idea. It's probably shit but what the hell I'll tell you if you want.
posted by Not Supplied at 2:35 PM on November 2, 2009


Interesting challenge. I wonder what, if any, measures they have in place to prevent against sabotaging the balloons. Are they monitored to ensure that a team doesn't destroy one after sighting to prevent other sightings?
posted by Burhanistan at 2:35 PM on November 2, 2009


Dear AskMe: I need a dozen 8 foot, red weather balloons in the next 30 days. Where can I procure these?
posted by GuyZero at 2:36 PM on November 2, 2009 [9 favorites]


1. Call CNN, Fox, MSNBC, tell them "my kid's missing, I think he got into my red weather balloon.
2. ...
3. Profit.
posted by qvantamon at 2:38 PM on November 2, 2009 [13 favorites]


The prize is 40K. That'd be an awfully big Mefi Meet-up Expense Account.

and considering the usual amount of beer consumed at these things...
posted by Lutoslawski at 2:38 PM on November 2, 2009


Yep, we should give this a go. Off the top of my head to get the blue-sky session started, we produce a Google Map with every participating MeFite's location marked on it, by a certain date. Once we have all the participants marked, we draw lines on the map to mark the areas that we'll cover-we'll probably need some local or statewide coordinators to organize this process to cover as much territory as efficiently as possible.

Also: We should agree up front to donate the 40k to a charity TBD should we win.
posted by Kwine at 2:41 PM on November 2, 2009 [3 favorites]


So there are no clues... each initial discovery has to be made by brute force? I guess the whole challenge then will be negotiating disclosure of discoveries between competing individuals.

Will there be team registration? What's to stop someone from joining a bunch of teams and selling them all out by grabbing the prize on their own? Is the real purpose of this challenge to study game theory?
posted by phrontist at 2:42 PM on November 2, 2009 [1 favorite]


If I see any of these in my little corner of the USA, I'll post the coordinates here. I'm not concerned about winning this thing, personally, but it would be fun to help the group effort. (And maybe get a shout-out if my contribution's legit, or maybe even a portion of the prize...)
posted by The Winsome Parker Lewis at 2:45 PM on November 2, 2009 [1 favorite]


We should agree up front to donate the 40k to a charity TBD should we win.

That's a pretty neat idea actually. And it sounds pretty easy to organize. I vote Kwine as the leader.
posted by Lutoslawski at 2:47 PM on November 2, 2009 [1 favorite]


Dear AskMe: I need a dozen 8 foot, red weather balloons in the next 30 days. Where can I procure these?

I am in favor of making the game harder for others, planting a ridiculous amount of fake balloons. If there is a Team Metafilter, I would like to sign up for the Department of Sabotage.
posted by amuseDetachment at 2:48 PM on November 2, 2009 [19 favorites]


I would like to sign up for the Department of Sabotage.

I can't stand it. I know you planned it - Oh my God, it's a mirage. Listen all y'all, &c.
posted by GuyZero at 2:50 PM on November 2, 2009 [11 favorites]


If we're sabotaging I vote for putting a ghostly Bruce Willis besides each red balloon.
posted by qvantamon at 2:51 PM on November 2, 2009 [2 favorites]


Once we have all the participants marked, we draw lines on the map to mark the areas that we'll cover-we'll probably need some local or statewide coordinators to organize this process to cover as much territory as efficiently as possible.

Where's that Voroni diagram post?
posted by GuyZero at 2:51 PM on November 2, 2009 [1 favorite]


I guess the whole challenge then will be negotiating disclosure of discoveries between competing individuals.

Will there be team registration? What's to stop someone from joining a bunch of teams and selling them all out by grabbing the prize on their own?


In our case, we solve these problems by giving only one trusted person access to the results of our efforts. Since it's her thread and no one is more trusted around here I expect, I nominate Jessamyn. Find a balloon? Tell Jessamyn. No one else needs to know (especially no one needs to post about it on Metafilter).

There shouldn't be any posting about NOT finding balloons either. And there is, now that I think of it, a role for active googlers looking to leverage the results of online communities that aren't as careful.

I vote Kwine as the leader.

that's nice of you to say, but IMO it should be someone with a little more name recognition, and jessamyn would be a better choice.
posted by Kwine at 2:52 PM on November 2, 2009


If we're sabotaging I vote for putting a ghostly Bruce Willis besides each red balloon.

It'd be funnier if there's no clear indication it's a fake. In fact, it should have a wide assortment of indicators that its The Real Deal. A different combination of DARPA ink stamps, stickers, official looking "You've found it!" letters, etc.
posted by amuseDetachment at 2:55 PM on November 2, 2009


The rules say each balloon will be "accompanied by DARPA representatives," so best to plan on faking those as well. I'm sure metafilter can supply sufficient numbers of middle-aged men with glasses and facial hair.
posted by exogenous at 2:56 PM on November 2, 2009 [5 favorites]


I am in favor of making the game harder for others, planting a ridiculous amount of fake balloons. If there is a Team Metafilter, I would like to sign up for the Department of Sabotage.

Lol. Don't tell 4chan or the Something Awful Goons, or there is a very good chance this will actually happen. In fact, then the game will become a meta-game of infiltrating different online communities to discover their known fake/real lists in order to piece together the puzzle.
posted by PercussivePaul at 2:57 PM on November 2, 2009 [4 favorites]


The rules say each balloon will be "accompanied by DARPA representatives,"

Huh. I totally missed that on first read. Sabotaging will also involve kidnapping or worse!
posted by Burhanistan at 2:57 PM on November 2, 2009 [2 favorites]


I've been thinking about this since I first read the announcement. It's really needle-in-haystack territory, for anything except a really large team. But the problem with a really large team is that the prize money no longer functions as much of a motivator. (And if you did try to motivate a lot of people — talking thousands here — with something relatively small like $40k, you're going to get nasty competition for a bigger slice of the pie.)

I think the winning solution is for a team/organizer to step up and say what they'd like to do with the prize, and then encourage people to participate not for the money, but for geek cred and ideological reasons. This is the approach that (IIRC) has been taken with some of the cryptographic challenges broken by distributed computation. The people providing the bulk of the resources were never in it for the money; they were doing it for the technical challenge and warm fuzzy feeling of just participating.

You can't afford to buy the participation of a large team of people, but you can easily get them to donate their time and effort if you ask just right. That's what I suspect the competition will boil down to.
posted by Kadin2048 at 2:57 PM on November 2, 2009 [1 favorite]


Lol. Don't tell 4chan or the Something Awful Goons, or there is a very good chance this will actually happen.

It should actually happen. DARPA doesn't do this sort of thing for kicks, it's a sociological and intelligence experiment as much as anything. I sure as heck am up for planting false balloons.
posted by Justinian at 2:59 PM on November 2, 2009 [2 favorites]


Is anyone else shocked--SHOCKED--at how ridonkulously bad DARPA's logo is? It's like they're doing advanced research into suck.
posted by Admiral Haddock at 3:01 PM on November 2, 2009 [4 favorites]


The rules say each balloon will be "accompanied by DARPA representatives," so best to plan on faking those as well. I'm sure metafilter can supply sufficient numbers of middle-aged men with glasses and facial hair.

DoS attacks don't attempt to perfectly emulate real traffic - they just suck up cycles and prevent the victim from getting stuff done. 10 fake balloons in an area means teams need to investigate 10x as many balloons before finding the real one. If the attackers who planted the balloons already know which balloons are fake they save a lot of time. The massively distributed nature of the exercise makes this type of attack less effective though because you need to have a lot of people involved to even attempt the task in the first place.
posted by GuyZero at 3:01 PM on November 2, 2009 [3 favorites]


Also, I'm a bit suspicious. Previous DARPA challenges have been for much higher stakes. Is this just their way of having a bit of fun? Perhaps they don't even expect any group to actually locate all of the balloons, but instead are going to be hard at work analyzing traffic across various social networking platforms, and maybe even all of the public internet, to do their own data mining and analysis. They can look for actual discrete usage of terms related to the search, and also run pattern matching algorithms to see if they can determine code words.
posted by Burhanistan at 3:02 PM on November 2, 2009 [5 favorites]


You can't afford to buy the participation of a large team of people, but you can easily get them to donate their time and effort if you ask just right.

Making it scarily likely that 4chan or SomethingAwful or some similar community will actually win the competition.
posted by GuyZero at 3:02 PM on November 2, 2009


The rules also say that ballons can be moved — it won't be suspicious if it's there one day and not the next.

I suspect that if someone is willing to sit around in a suit and plant a fake balloon in a large metro area, you're going to screw up an overwhelming majority of the teams. Plant one single person in Manhattan, DC, LA and San Francisco (or other areas where people have heard of Twitter and know how to self-organize online), and you've screwed up every single team, as this game is going to require national coverage and that team probably has many people in those cities. You can also guarantee that every single team will find it by posting on twitter, &c.
posted by amuseDetachment at 3:03 PM on November 2, 2009


That's an interesting attack - simply having a group of people disclose as many balloons as possibly through a public channel like twitter. What if someone simply twittered all 10 real balloon locations within a few days? Would the competition be voided?
posted by GuyZero at 3:07 PM on November 2, 2009


Chloroprene cloud busters ON SALE NOW no minimums
posted by longsleeves at 3:09 PM on November 2, 2009


DoS attacks don't attempt to perfectly emulate real traffic - they just suck up cycles and prevent the victim from getting stuff done. 10 fake balloons in an area means teams need to investigate 10x as many balloons before finding the real one.

I'm not sure if DARPA is expecting it to work out like that, but espionage and misinformation would really be some of the best methods in this kind of competition. A naive strategy for winning would be to gather up as many people as possible, collect sightings, and share the results.

But what if a ton of people join the group and submit fake sightings (which would be fairly easy to do)? You'd have to independently verify them, and if you aren't careful about that you could accept verifications from the same group of people who submitted the fake sightings. And if you let a lot of people see the probable/verified coordinates, what is stopping them from stealing that information and combining it with the findings from other teams that they have infiltrated, to eventually beat you to the punch? It's very possible that the team to win this challenge will be the best at stealing information from other teams, and that the teams that were the best at actually finding the balloons will not end up winning.
posted by burnmp3s at 3:11 PM on November 2, 2009


DoS attacks don't attempt to perfectly emulate real traffic - they just suck up cycles and prevent the victim from getting stuff done.

The advantage of a good fake balloon (i.e., one with a convincing crew of DARPA reps) could be that it fools people into thinking they've found the nth balloon so that they stop early. But that would depends on whether DARPA confirms the submitted coordinates immediately, etc.
posted by exogenous at 3:11 PM on November 2, 2009


Lol. Don't tell 4chan or the Something Awful Goons, or there is a very good chance this will actually happen.

It should actually happen. DARPA doesn't do this sort of thing for kicks, it's a sociological and intelligence experiment as much as anything. I sure as heck am up for planting false balloons.



Giant red balloons run about $25, just sayin'.
posted by Lutoslawski at 3:13 PM on November 2, 2009


The rules also say that ballons can be moved — it won't be suspicious if it's there one day and not the next.

...

What if someone simply twittered all 10 real balloon locations within a few days? Would the competition be voided?

Guys, the balloons are only in place for 6 hours. It's in the rules. You have 6 hours, from about 10:00am to 4:00pm to find the balloons.
posted by Justinian at 3:13 PM on November 2, 2009 [1 favorite]


blarg. longsleeves beat me to it.
posted by Lutoslawski at 3:14 PM on November 2, 2009


The rules also say that ballons can be moved

I think the implication there is that DARPA can move the balloons if they wish, and either the old or new location will be accepted.
posted by Burhanistan at 3:15 PM on November 2, 2009


Oh, I see, the balloons are only in place for 6 hours but you have like 10 days after they balloons are removed to submit answers. So, yeah, it's largely an experiment in information sharing and such.
posted by Justinian at 3:18 PM on November 2, 2009


Srsly. This is a good project. 40K would buy a bunch of fistula surgeries.
posted by theora55 at 3:18 PM on November 2, 2009 [5 favorites]


GuyZero, what's the advantage to the group in broadcasting the balloon locations over twitter rather than just claiming the prize?

amuseDetachment, it's unclear why that will screw up teams-any team sufficiently large and motivated to have a chance will be smart enough to just collect all reported balloon locations and let DARPA sort out the fakes from the reals. Maybe the winning team submits 1000 locations and 990 of them are fake?

In terms of having LULZ, there's little advantage in sabotage, because there's no reason why any individual should assume that there's only one balloon (or only eight balloons) in their search area. Just keep pinging your search areas and reporting in if you find anything. In terms of trying to win the competition, there's no advantage gained in sabotage, it seems to me: those resources would be better spent looking for balloons, real or fake.
posted by Kwine at 3:21 PM on November 2, 2009


We already have the lat/long for a lot of members. What's the max visibility for an 8ft in diameter balloon moored at 1000ft? Assuming no buildings, we should be able to come up with a map of the distance everyone sees if they step outside and look around.

I have a feeling there's a critical point in which you have X amount of people looking over the continental US and you have a high percent chance of spotting the 8 balloons, which is what DARPA is going for. I would also bet that if you had an amount some arbitrary number larger than X, let's just say 2X the time it takes to detect the balloons does not decrease proportionally. Simply, even if you had a Facebook group with everyone in the continental US participating, you're not going to see better results than if you were the size of something like Metafilter.

Of course I have no idea how we'd setup an entity to register for the contest and handle the funds. The real winner of this might not be some group of spotters, but a dedicated group that watches social media and discovers the balloons other, non-registered participants have found. In that case we might really have an advantage.
posted by geoff. at 3:24 PM on November 2, 2009


Kwine: Large numbers get scary. If there are a total of 50 fakes (not unreasonable as I would assume many teams/individuals would be inclined to do so), there's a crapton of possible combinations, it's not reasonable to brute-force the problem-space of correct 10 with 50 fakes out there.
posted by amuseDetachment at 3:27 PM on November 2, 2009


DARPA are probably expecting griefing and are really looking into the ability of distributed social networks to sift through the garbage to find the real locations. The filtering is much more interesting than the finding.
posted by knapah at 3:31 PM on November 2, 2009 [2 favorites]


If I owned a business that wanted to get a bump in business I'd be sure to fly a red balloon off my roof on those dates.

I suspect that the contest assumes -

- there will be fakes
- people will subvert teams
- misinformation will be rampant
- traffic snarls if these are near cities

etc......

One of the things I'm wondering about is the level of chaos DARPA is ready to eat over this. Seems like you'd have to put them well away from towns and cities.

My best strategy would involve teams of private pilots criss-crossing the country. Shouldn't be hard to plot a grid and hand out assignments. If you could pair that with a group like MetaFilter that could confirm balloon on the ground you'd have an extremely powerful search effort.

Also, I spent a lot of time driving the "publicly accessible" roads in Western Montana. Hiding an 8 foot balloon so that's it easy to see from a mountain road in the deep wilderness, but nearly impossible to see from the air, would be trivial. I wonder how hard they're really trying to hide them. Part of the challenge may be making sure you can find out if some local hiker tells a crazy story about a guy in the woods with a weather balloon at the coffee shop.
posted by y6y6y6 at 3:32 PM on November 2, 2009 [1 favorite]


In other words.... try to think of where *we* would hide them might be a good place to start.
posted by y6y6y6 at 3:33 PM on November 2, 2009



DARPA are probably expecting griefing and are really looking into the ability of distributed social networks to sift through the garbage to find the real locations. The filtering is much more interesting than the finding.


Yeah, the more I think about this, the more I think the prize goes to whoever can recruit & organize largely anonymous participants at scale. Socially, it's Wikipedia. Technologically it remains to be seen whether this requires a structured tool or whether a group could do it with as something as simple as a wiki page.

I would not be shocked to see Drew Curtis pick up the cheque with Fark as another viable community that would do this just because.
posted by GuyZero at 3:35 PM on November 2, 2009


GuyZero, what's the advantage to the group in broadcasting the balloon locations over twitter rather than just claiming the prize?

Just griefing. There's no advantage. It would subvert an unstated assumption of the content that there's no public information shared between teams beyond the physical world.
posted by GuyZero at 3:37 PM on November 2, 2009


Next DARPA Project: $40,000 for the first person to find the top ten most wanted terrorists via Twitter.
posted by amuseDetachment at 3:38 PM on November 2, 2009


panic bells! it's red alert!
posted by sexyrobot at 3:44 PM on November 2, 2009 [2 favorites]


Wait.......... mailmen (persons?). They will be on every single road with mail service. Find a channel that is shared by all U.S. mailmen and you have one hell of a search team. Similar for truckers.
posted by y6y6y6 at 3:45 PM on November 2, 2009 [2 favorites]


I expect all 10 balloons will be found and posted on twitter in less than an hour.
posted by empath at 3:47 PM on November 2, 2009


Yeah, the more I think about this, the more I think the prize goes to whoever can recruit & organize largely anonymous participants at scale.

I will laugh like hell if 4chan wins this.
posted by empath at 3:48 PM on November 2, 2009 [3 favorites]


That's possible, empath. I suspect there will also be many thousand fake findings posted to Twitter as well. The problem because determining which are the real balloons.
posted by Justinian at 3:50 PM on November 2, 2009


aD, the rules are unclear on the submission guidelines. If you can submit ten and only ten locations, you're right-the complexity becomes totally unwieldy. If you can submit arbitrarily more than that, we're ok.

Here's an email that I sent to the contest email address to decide this:

"Hello,

I have a question about the balloon submission rules, in the context of possible sabotage of the contest by outside parties. Is there a limit to the number of balloon locations that I can submit? If I submit 1000 locations and 990 of them were fake but 10 were real, would that be an acceptable way to win the contest?

If I have ten and only ten locations that I am allowed to submit, I'm concerned about authentication. How can I verify that nefarious parties haven't bought their own red balloons, staffed them with fake DARPA agents, etc? What will separate a "real deal" DARPA balloon from an impostor?

Thank you for your help,

(Kwine's real name)"

If I find out anything interesting, I probably won't want to post it in this thread-no sense helping out another team with my legwork. We should think of a more secure way to communicate about this contest-even if I don't learn anything useful, at some point we'll want to share some kind of information without letting others in. A NOINDEX on this thread is an obvious place to start.
posted by Kwine at 3:50 PM on November 2, 2009 [2 favorites]


I will laugh like hell if 4chan wins this.

You better wash your laughing pants then because you may well need 'em.
posted by GuyZero at 3:55 PM on November 2, 2009 [1 favorite]


We should think of a more secure way to communicate about this contest-even if I don't learn anything useful, at some point we'll want to share some kind of information without letting others in.

I'm going to bet that the person that wins it, wins it by using google on the day of.

I think the best strategy is to get a team of about 20 or 30 people in irc googling like mad.
posted by empath at 3:56 PM on November 2, 2009


Mail carriers is the best option so far. I'm not sure who else has the potential to be organized effectively at a national level with that sort of coverage [rural locations will be poorly ocvered by Fark and MetaFilter].
posted by jessamyn at 3:57 PM on November 2, 2009


HA HA HA: cashforredballoons.com.
posted by GuyZero at 3:58 PM on November 2, 2009


"I expect all 10 balloons will be found and posted on twitter in less than an hour."

This assumes the balloons would be placed in areas likely to have people using Twitter. Taking Montana as an example again - Once you get out of the cities you have a a vast area where people a) don't use Twitter at all, b) mind their own business when they see something weird, and c) are much more likely to tell the story in a bar or diner than anywhere else.

If I was DARPA I'd build a map showing all Twitter/Facebook/Google user locations. Then I'd put many of the balloons where there were no dots. Think DARPA doesn't have that map already compiled somewhere?

Just putting the balloons where Twitter and 4chan will find them would make the contest pointless. I assume DARPA is looking for hard things to solve, not completely trivial stuff.
posted by y6y6y6 at 3:58 PM on November 2, 2009 [1 favorite]


rural locations will be poorly ocvered by Fark...

Yeah, I dunno about that. There are a lot of people on Fark, sometimes because they're stuck somewhere with nothing to do. MeFi, sure. If the balloons were hidden in grad schools across the country, or in 10 locations across Williamsburg, that's a different story.
posted by GuyZero at 3:59 PM on November 2, 2009


Admiral Haddock: "Is anyone else shocked--SHOCKED--at how ridonkulously bad DARPA's logo is? It's like they're doing advanced research into suck."

I don't know. I think it has a nice OCP vibe to it.
posted by brundlefly at 4:01 PM on November 2, 2009


Actually.......... Can we get a map showing all MeFi/Facebook users? Shouldn't someone with too much time on their hands already made that map? Overlay it with a highway map and you have a map showing - nearly exactly - "places in the US where it would be very hard for 4chan to know about a 8 foot red balloon, even if we offered them $40,000".
posted by y6y6y6 at 4:03 PM on November 2, 2009


Can we get a map showing all MeFi/Facebook users?

I'm not seeing how to pull the information on members who logged in at least once in the last 3 mos. and who give a lat/lon in their profile. Put in some random variable for the name and it should be fairly simple to map using google's API. I was thinking of going further and taking away all areas not within 4 miles of an interstate, excluding metro areas. It should cut down a lot right there.

In any case I doubt we have more than a couple of thousand members that fit such criteria.
posted by geoff. at 4:10 PM on November 2, 2009


I'm not seeing how to pull the information on members who logged in at least once in the last 3 mos. and who give a lat/lon in their profile.

I should say I cannot how to see to put this together from waxy's stats, it might be fairly simple if the right data is there.
posted by geoff. at 4:11 PM on November 2, 2009


Seems to me that if this contest is way for DARPA to look at the way social media can be used, they'd plant the balloons in places where lots of people use social media. So, probably not rural Montana.
posted by Quietgal at 4:13 PM on November 2, 2009


Selflink: I've written extensively about this in a post called How to Win the DARPA Network Challenge, and I've also started a wiki collating resources and links about the challenge (it is not a place for a group to co-ordinate though) - plus here's a Twitter list for accounts related to the challenge.

Trying to win this with a secure, closed group is going to be very tricky; it is far too easy for someone to infiltrate an online group and feed in false information. It's much smarter to assume you'll be receiving false positives and develop a robust system that can assign reliabilities to each report.

Cash award to correct reporters will not work. On my blog post, the first comment was from someone offering "at least $1000 for verified reports". Only three comments later was an offer for $3000 for verified reports. It's an arms race no-one wins.

Money is not going to be a primary motivation here; charity, or perhaps kudos, is. I can imagine a large group of people getting together to donate the prize to charity - in fact, many of the groups are already promising to do so. Another possibility is a large company or organisation (e.g. Google) deciding to throw a bit of cash behind an effort in order to get some good publicity.

I see a lot of people casually talking about monitoring Twitter and Facebook to win. It is very sad to see this, because it will be an enormous waste of time; just wait until December 5th, and I guarantee you there will be thousands if not hundreds of thousands of false reports on every channel you can think of.
posted by adrianhon at 4:20 PM on November 2, 2009 [6 favorites]


"Balloons will be in readily accessible locations, visible from nearby roadways and accompanied by DARPA representatives."

So, do you walk up to the DARPA rep and ask for ID?
posted by longsleeves at 4:20 PM on November 2, 2009 [1 favorite]


Cash award to correct reporters will not work. It's an arms race no-one wins.

I think the smart way to do a cash award is to make each first correct report a lottery ticket for the full $40,000. Obviously this would only work well if the organizers were also searchers, like in the Post Office case.
posted by jedicus at 4:28 PM on November 2, 2009


Yes, the lottery idea is a nice one, particularly when combined with someone like the Post Office. The challenge there would be finding someone within the organisation with enough time and power to co-ordinate the workers there and publicise their group, which could be tricky given institutional roadblocks and bureaucracy.
posted by adrianhon at 4:34 PM on November 2, 2009


Is this something I'd need to be in the World Wide Web USA to play?

How quaintly local this game is.

/obvious
posted by dash_slot- at 4:34 PM on November 2, 2009 [1 favorite]


Interesting. I'm trying to think through the feasability of mustering a mefi-only effort, and these are the things jumping out to me:

- raw size of source group (i.e. the mefi userbase) is not as important as distribution. 1000 people in the US and 900 of them in NYC (FIGURE NOT TO SCALE) is not much more help than 101 people in the US, 1 of them in NYC.

- adrianhon's point about infiltration of a secure closed group is a reasonable one, but the question of enforcing security seems like the key thing there. If the mefi team consists only of people who were already members of the site as of, say, a month ago, we reduce the likelihood of someone directly joining with infiltration in mind. (Secondary weaknesses abound, but it's a good start.)

- A group commitment to send the lump sum to a charitable third party mitigates the pay-for-play issues. A trusted figurehead who can't afford to fuck everyone over obviates the solo-betrayal issue. I'd be comfortable being that guy whose mefi life would be ruined by said ethical lapse, as far as that goes.

- The plausibility of making it work, all else aside, comes down to the sheer volume of coverage we can muster. So distribution + availability + mobility spells out raw sighting resources; some math should tell us how much road we can cover per hour, how much road needs covering, and most importantly how the covering of that road can be most efficiently (e.g. evenly) accomplished with the resources we have available.

I suspect we could put together a pretty solidly designed and managed effort. I'd be impressed if we managed a full eight sightings. I'd be shocked if we were actually competitive for a win. I think kicking ass and not winning is a totally acceptable idea.
posted by cortex at 4:39 PM on November 2, 2009 [7 favorites]


Dash: If you're interested in doing something like this in the UK, then get in touch - I've been planning similar sort of thing for some time, but British (i.e. really cheap and set in a pub).
posted by adrianhon at 4:39 PM on November 2, 2009


Cortex: Mefi is certainly as well-placed to tackle this as any other site, I think (others would include the Straight Dope, Ars Technica, Fark, Something Awful, etc); it has a good mix of very smart and motivated people with a range of skills.

Restricting entry to people who signed up just before the competition was announced would absolutely help a lot, although I think we're all aware that there are plenty of spoilers amongst us ;) Still, it shouldn't be too difficult to work out a system to check and corroborate reports. The issue will be whether keeping things restricting outweighs the benefits of being more open (as, say, 10loons appear to be).
posted by adrianhon at 4:44 PM on November 2, 2009


Trying to win this with a secure, closed group is going to be very tricky; it is far too easy for someone to infiltrate an online group and feed in false information.

I think this is right; I also think it's a factor (maybe the only factor?) where we have a competitive advantage. We have a community that is remarkably diverse, for a web-based community, in terms of talent and geographical location. We have some built in controls to keep out the riff-raff (no one who joins after this was posted should get any access). We don't have any red tape, top down institutional bureaucracy that keeps us from getting off the ground in time. We're used to organizing together for little financial gain. We have loyalty to each other-we trust each other the way the SA Goons and 4chan fags(is that right? forgive me, I am not one of them) don't, for example.

I really, REALLY believe that we're as well-positioned as any group out there to do this for little capital. If Google decides to jump in and spend a bunch of money to do it, good for them I guess, but to me that's outside the spirit of the thing. We just have to be better than everyone else, and I think we can be.

Uh, that's my speech.

On preview, yay cortex!
posted by Kwine at 4:45 PM on November 2, 2009


If you can submit ten and only ten locations, you're right-the complexity becomes totally unwieldy. If you can submit arbitrarily more than that, we're ok.

At 12:01 AM on December 5, someone should submit a file containing a list of every valid lat/long combination in the continental US.
posted by ook at 4:48 PM on November 2, 2009 [3 favorites]


I expect that you will only be able to submit ten locations at a time. I doubt they'll only let you submit once, but they'll probably rate-limit submissions to perhaps one every few minutes (that's how I'd design it if I was making this as part of an ARG)
posted by adrianhon at 4:51 PM on November 2, 2009


At 12:01 AM on December 5, someone should submit a file containing a list of every valid lat/long combination in the continental US.

Yeah, I was just thinking of this, and so did you, and that means 5,000 other people did too. Whatever measures are in place, they'll have to disqualify that approach on pain of having a really shitty contest. On the other hand, they can't possibly limit you to ten and only ten guesses...can they? We'll see what they say...
posted by Kwine at 4:53 PM on November 2, 2009


Well, Adrian, if only I had the resources and the, er, global vision that an advanced research projects agency had, then - yes!

[quits carping from the cheap seats]

Good luck, my american friends!
posted by dash_slot- at 4:54 PM on November 2, 2009


How quaintly local this game is.

As much as I love harping on the US-centrism on MeFi, this is actually a research project sponsored by the US Government just like the one where they have robot cars drive around the desert. It's US-centric by virtue of its sponsor who is doing this to actively work against the rest of the world.
posted by GuyZero at 5:03 PM on November 2, 2009


Yeah, now that I get around to reading the Register link I see they suggested the same thing.

Easy fix: winner is whoever gets the greatest number of correct locations from the shortest list of locations. Curious that they didn't clarify this in the rules to begin with, though, it's such an obvious trick...
posted by ook at 5:03 PM on November 2, 2009


Yeah, that Register article sure had some great hacks in it. If, by "hack", you mean "suggestion".

Hack. Hack hack hack hack hack. Hack.
posted by lalas at 5:05 PM on November 2, 2009


lalas: You would love Merlin Mann's Surprising House Hacks!
posted by amuseDetachment at 5:07 PM on November 2, 2009 [1 favorite]


This contest will take place on 12/5 so I'd suggest a sidebarring is in order. If we decide a bit of monkey-wrenching is in order I'll be in Seattle and have been successfully impersonating an adult for some time now.
posted by vapidave at 5:08 PM on November 2, 2009


Coordinates must be entered with an error of less than one arc-minute to be accepted

I wonder if the lack of a coordinate system is deliberate?

Anyone with GIS knowledge, fast internet connection and programming skills could probably find all 10 location working alone pretty easily in a day or so. The only tricky part would be that you have a bunch of people who might or might not know how to use a GPS and you have no way of viewing the metadata. If you get info from some place where everyone routinely uses their state plane system instead of WGS84, you'd never know to convert it.

If I were DARPA I would locate the balloons right on the edge of the arcs too, just to be a dick.
posted by fshgrl at 5:18 PM on November 2, 2009 [1 favorite]


Set them free at the break of dawn
'Til one by one, they were gone
Back at base, bugs in the software
Flash the message, "Something's out there"
posted by tellurian at 5:24 PM on November 2, 2009 [2 favorites]


Sounds like a good project to draw on amateur radio and storm spotters.
posted by acro at 5:54 PM on November 2, 2009


And stormchasers! We could see if that nice Heene family would like to help.
posted by cortex at 6:06 PM on November 2, 2009 [2 favorites]


Prisoner remake viral!
posted by cowbellemoo at 6:34 PM on November 2, 2009 [2 favorites]


39.782878, -105.080109
47.264596, -122.517128
38.479511, -122.731276
36.157133, -96.095352
30.274689, -82.123919
43.090821, -78.992386
39.607901, -77.707157
42.251740, -71.055794

You're welcome.
posted by odinsdream at 6:50 PM on November 2, 2009 [5 favorites]


So is it safe to say that these will all be placed in Large Metropolitan Areas™ with maybe one token balloon in flyover country?
posted by crapmatic at 8:16 PM on November 2, 2009


DARPA would already have a list, on some server somewhere, of the locations where the balloons are going to go, wouldn't it?
posted by flabdablet at 8:33 PM on November 2, 2009


amuseDetachment, yes. That was hacktastic.
posted by lalas at 8:48 PM on November 2, 2009


Assuming this is going to require Actually Going Outside, then the numbers we can count on are probably at most equivalent to the 10th anniversary meet-up.

What kind of numbers and coverage were we talking about there? We had a big chunk of the continental US covered right?

We at least have a pretty sizable contingent of people in the DC/Nova area, and I'd be shocked if there wasn't at least one balloon there. Are we going to divy it up? Should we do a kind of travelling salesman thing where we figure out what path's people should drive on the day of to cover the major roads?

Also, I agree that yes, we do this for charity, but let's set aside 10g for the people that find the 10 balloons. The potential for $1000 is probably enough for a lot of people to get out of bed in the morning, especially paired with a fun game, and if people don't want to take it, they can throw it into the kitty with the rest of the money.

How accurate is the iphone GPS? Close enough for the contest?
posted by empath at 9:46 PM on November 2, 2009


Also, do what do people think about espionage?

I'd think the number one seed would be Goonswarm if they went for it. They have numbers and they're frighteningly well organized and ruthless. I mean, they have their own intelligence agency. How many message boards have that?
posted by empath at 9:51 PM on November 2, 2009


Is this a viral ad for a reissue of the movie?
posted by Cranberry at 10:54 PM on November 2, 2009 [1 favorite]


MeTa thread for planning purposes.
posted by empath at 11:14 PM on November 2, 2009


DARPA would already have a list, on some server somewhere, of the locations where the balloons are going to go, wouldn't it?

Speaking from experience, you're not getting it that way.
posted by scalefree at 11:18 PM on November 2, 2009


Does anyone find it odd that the rules state the balloons will be on display from 10AM ET to 4PM local time? If read literally that means 7AM to 4PM in the pacific time zone.

Quick back of the envelope calculation: according to this there are about 4 million highway miles in the US. Assuming a full 6 hours of driving at 60 MPH, that means you would in theory need about 11,000 participants to conduct a full brute-force sweep. I'm sure the actual number would be a lot higher because you can't achieve perfect efficiency of coverage. Eleven thousand dedicated volunteers is nothing on the internet scale, however it is quite large on the scale of a single site -- unless you're talking about facebook. But facebook suffers a different problem in that their massive resources will be splintered into countless factions.

Several people have commented on the idea of whether there would be balloons placed in remote areas. If that were the case it would bias the competition towards the teams that are focusing on shoe leather (i.e. going out and looking) and away from those that plan to infiltrate and/or google. I tend to think that the focus of the contest is more on the social side than the actual physical act of searching side, which is why my guess is that they will all be hiding in plain sight of many people, i.e. in major metro areas near busy thoroughfares. The real challenge then is not so much organizing teams to go out searching but rather to simply get word out to as many people as possible to know what to do if they see a red balloon during their normal Saturday activities. If this theory is correct then it would be better to have a team of 5000 people who don't actually modify their Saturday behavior in any specific way but who are on the lookout and will report to their team if they see something, compared to a team of 1000 dedicated hunters who plan to spend the whole six hour window doing nothing but driving around looking.
posted by Rhomboid at 11:26 PM on November 2, 2009


I'm actually imagining a combination of people going out looking and a team of people at home, browsing the web for tips and who have contact info for lots of mefites around the country so they can send them out to verify.

Maybe a 50/50 online/real-world split.
posted by empath at 11:31 PM on November 2, 2009


I'm just thinking...is there a way to involve the citi-data forums in this effort? These are committed communities of people from all across the USA, both rural and urban, far and wide, that maintain a lively and constant presence on what's happening in their communities. Not sure how or if to involve this resource, but it's worth putting on the table for consideration.
posted by zagyzebra at 12:29 AM on November 3, 2009


I wonder how they'll ultimately use this capability?

Namely... leveraging disparate user-generated social networks and WOM to geo-locate multiple small independently-moving targets on earth?

How do you say: "Anybody seen Bin laden?" in Urdu?
posted by LakesideOrion at 7:47 AM on November 3, 2009


LakesideOrion: If I were a conspiracy theorist, which I'm not, I would say it would be a great way to benchmark the most efficient user-generated networks and diagnose how they operate at the media level in order to create an efficient method of disrupting these networks in the event of large-scale operations that would not be possible to cover up without disrupting them. The force, power and scope of the internet as a social connectivity tool must be terrifying to people ostensibly in charge of controlling a populace. It sure would be nice to know how it works at its most ruthlessly efficient.
posted by odinsdream at 8:05 AM on November 3, 2009 [5 favorites]


Oh come on sheeple! Don't you realise that there are no red balloons? They're going to be looking out for reports of red balloons so they know where the trouble makers are.

*adjusts peak on bacofoil cap*
posted by twine42 at 8:23 AM on November 3, 2009


The rules say each balloon will be "accompanied by DARPA representatives," so best to plan on faking those as well. I'm sure metafilter can supply sufficient numbers of middle-aged men with glasses and facial hair.

Heh. Or young geeks. I think DARPA folks would be a mix of lots of types.

Not much I can do as one of a million mefites in the Boston area, but I'm so up for this.
posted by clauclauclaudia at 10:40 AM on November 3, 2009


So is it safe to say that these will all be placed in Large Metropolitan Areas™ with maybe one token balloon in flyover country?

DARPA doesn't generally do "easy" challenges. "[R]eadily accessible locations and visible from nearby roadways" covers a lot of territory. "Readily accessible" could very easily mean a small farm road in the middle of rural Wyoming. There's no reason to believe that it means "readily accessible to internet readers in big cities on the coasts who don't own cars." In fact, if the challenge is worth anything, it will force those people to expand their reach significantly in order to participate.

I wouldn't expect a balloon to be placed on a 4-wheel-drive-only road in the mountains somewhere. But I would expect several to placed on seldom-traveled two-lane paved roads in the mountains somewhere. A balloon could be in the desert 50 miles in the distance from I-15 north of Barstow and visible only by someone with a very keen eye who is looking for it and not only fit the definition, but be a really obvious one, comparatively (since that's the general vicinity of other DARPA challenges).

Several people have commented on the idea of whether there would be balloons placed in remote areas. If that were the case it would bias the competition towards the teams that are focusing on shoe leather (i.e. going out and looking) and away from those that plan to infiltrate and/or google. I tend to think that the focus of the contest is more on the social side than the actual physical act of searching side, which is why my guess is that they will all be hiding in plain sight of many people, i.e. in major metro areas near busy thoroughfares.

I agree that the focus seems to be on the social side. But that does not mean "major metro areas near busy thoroughfares." In fact, I think it would make a lot more sense to force the competition to include social and internet involvement of people in remote areas. They have the internet in Ketchum, West Yellowstone, Munising, and Gallup, too. Forcing teams to stretch their use of available social networks is something that I would expect DARPA to do.

Frankly, I would be shocked if more than 2 or 3 of the balloons were in metropolitan areas - big or small. I would also be shocked if any of them were visible from a busy thoroughfare. If the idea is to celebrate the anniversary of the internet, it seems to me that demonstrating the incredible reach of the internet to remote areas would be an obvious part of that celebration.

Does anyone find it odd that the rules state the balloons will be on display from 10AM ET to 4PM local time? If read literally that means 7AM to 4PM in the pacific time zone.

To me, it seems to say 10AM to 4PM local time to the balloon itself. So the balloon could be in the front yard of a farm house on the outskirts of Firth, Idaho and the DARPA folks only have to stay with it during the day and can then travel to their lodgings for the night. The timeframe suggests to me that they are definitely going to put the balloons in the middle of nowhere.

Quick back of the envelope calculation: according to this there are about 4 million highway miles in the US. Assuming a full 6 hours of driving at 60 MPH, that means you would in theory need about 11,000 participants to conduct a full brute-force sweep.

A brute force sweep of the highways. And there's no reason to believe that any of the balloons will be visible from any highway, is there? I mean, the announcement says "roadway," not "highway." They could be at the end of suburban cul-de-sacs. The sheer vastness of the United States "roadways," combined with the surprising reach of the internet into remote areas make the challenge much more interesting than if it was just a question of demonstrating that people in big cities can communicate with each other.
posted by The World Famous at 11:24 AM on November 3, 2009


Everyone's a superhero--everyone's a Captain Kirk, with orders to identify, clarify, and classify.
posted by not_on_display at 11:58 AM on November 3, 2009 [1 favorite]


LakesideOrion: "I wonder how they'll ultimately use this capability?"

Pure speculation, but my guess is that the purpose of the competition is to measure/observe how information propagates via social networks, and also how people might assemble various bits and pieces of information from diverse sources into something "actionable" (list of the correct locations and only the correct locations).

It seems like it has greater theoretical than immediately practical value, and would also explain the rather low prize payout compared to challenges with more immediate applications (e.g. self-driving cars).

I thought of odinsdream's scenario (learning how the networks function so they could be disrupted) as well. I don't think that's really tinfoil-hat category at all; knowing how to disrupt or feed bad information into a possible intelligence-gathering mechanism seems fairly logical.
posted by Kadin2048 at 12:55 PM on November 3, 2009 [1 favorite]


This is way cool, thank you for posting it.
posted by LobsterMitten at 1:03 PM on November 3, 2009


I just bought stock in all the big red balloon manufacturing companies.

I win.
posted by Bonzai at 1:08 PM on November 3, 2009


Well, if I spot a big red balloon in my part of rural Texas, I'll certainly let Mefi know.
posted by threeturtles at 1:50 PM on November 3, 2009


If they meant 10AM in local time why didn't they just say that instead of "10AM ET"? They went out of their way to specify the start and end times differently. It just doesn't make any sense.

And there's no reason to believe that any of the balloons will be visible from any highway, is there?

The reason is that they want it to have some small chance of this being doable. If you take their suggestion literally to mean that it could be visible from anywhere that there's any kind of road then the chances of a team being able to get all ten goes down to vanishingly close to nil, at least with a paltry 40k budget. As already mentioned their previous challenges had prizes on the order of a million dollars and I seem to recall that the winning teams spent something like 1.5x - 2x the prize amount (in other words, they wanted to win for the acclaim and the money was not a factor.)

Actually, that 4 million miles figure might be a bit high. I just checked the CIA world factbook and they list 2,615,870 miles paved and 1,401,791 miles unpaved (4,017,661 roadway miles total.)
posted by Rhomboid at 2:17 PM on November 3, 2009


« Older Two new biographies examine the life and legacy of...  |  An insurance industry CEO expl... Newer »


This thread has been archived and is closed to new comments