Forgot your password?
typodupeerror
Image

Yahoo, Facebook Test "Six Degrees of Separation" 228

Posted by samzenpus
from the small-world dept.
An anonymous reader writes "Yahoo has partnered with Facebook to test the iconic social experiment known as 'six degrees of separation' (everyone is on average approximately six steps away from any other person on Earth). The goal of the Small World Experiment is to determine the social path length between two strangers by tapping into the world's largest social network and its 750 million users, each of whom have an average of 130 friends." Looks like a fun project, but not quite as useful as knowing how close you are to Kevin Bacon.
This discussion has been archived. No new comments can be posted.

Yahoo, Facebook Test "Six Degrees of Separation"

Comments Filter:
  • So does this mean that even though I have explicitly not given my yahoo account to my facebook account, they're about to sync with eachother anyway? Great.
    • I know there's little to no chance that my Yahoo accounts will know about my other accounts. I haven't updated any of my personal information (such as where I live) in either of them. One has a location from 3 moves ago, the other from 4 moves. There's little to no chance of the accounts syncing with each other, much less with anything I might have on a social networking site.

      The key is that all of these social networking sites and email providers only have the information you give them. They can guess

      • by iggymanz (596061)

        you forget about all the facebook web beacon javascript, all the major websites have them. (noscript ftw!)

        • you forget about all the facebook web beacon javascript, all the major websites have them. (noscript ftw!)

          And about everyone he knows who has him in their address book and friends list.

          • by cayenne8 (626475)
            Not to mention, that they can purchase other databases of info....like some states sell drivers license info. You can buy the change of address database from the USPS...stuff like that you can use to 'clean' your data.

            Companies like Acxiom [acxiom.com] do this all the time.

            In fact..likely as not, facebook already uses them to help 'clean' data....lots of companies do like Visa, etc...

            • by Cytotoxic (245301)

              I found out just how accessible information can be several years ago when I worked on a project involving a collections/skiptracing firm. For a quarter you could get conglomerations of various data sources that would pinpoint most targets current location - including address and phone number. One guy they were looking for while I was in the office was tough enough to find they sent it out to one of their private investigators. For $75 he not only found the guy's location (a secluded island only reachable

        • Yep. Even this site uses google-analytics. It's scary how much information is picked up from your surfing habits almost anywhere on the web.

          • by Nursie (632944)

            Noscript, Adblock, Flashblock, Cookie Monster, Better Privacy. Flush browser cache on exit. New email address for every site I join. A few other tricks...

            Makes the web a much nicer place and my habits much harder to correlate between different places.

            I mean except panopticick type stuff browser fingerprinting, but then I'm not sure how much of that data can get back to the site without various forms of scripting.

          • You might want to add this [mvps.org] to your list of privacy-enhancing tools: Should also work with Linux and Mac.

            Example:

            127.0.0.1 www.google-analytics.com #[Google Analytics]
            127.0.0.1 4.afs.googleadservices.com

            Google, alone, has over 100 entries like the above in my HOSTS file. Speeds up surfing. In Firefox, at least, you do take a performance hit when you restart the browser after making any change to this, rather large, file. No noticeable lag at startup after that until you make another change to HOST

      • by rubycodez (864176)
        they don't use the "personal information" fields (too many people same name, same age, same city), that's not how the game of internet marketing and tracking is played. 3rd party javascript on the websites you visit, 3rd party cookies, your browsers cached data compared between successive sites with common 3rd party content, pixel beacons in emails and web pages, friend lists, social networking profiles used as shortcut for registration.
    • by iggymanz (596061)

      yes, and more importantly, they'll have new opportunities for selling marketing data and data mining services.

              win-win all around for the both of them

      Everyone should drop the facebook, and get their own virtual host website for less than $5 a month, and put pictures and have comments / blog there. Or set up a web server at home and use a free dynamic dns, it's good geeky fun. plenty of my-website-in-a-can packages out there.

      • by Lumpy (12016)

        wordpress is easier for non nerds...

      • by Scaba (183684)

        Likewise, everyone should drop their phone company and set up their own CLEC. Or maybe everyone else isn't a socially hostile and ultra paranoid geek and are willing to trade information that's publicly available anyway for a convenient and centralized site for social interaction.

    • It was bound to happen eventually.
  • by foobsr (693224) on Tuesday August 16, 2011 @10:30AM (#37106888) Homepage Journal

    TFS: " each of whom have an average of 130 friends "

    ... where the validity of hitherto common concepts vanishes.

    CC.

    • by alen (225700)

      i'm at about this number. doesn't mean i hang out with everyone but i've known these people at some point in my life. some people i haven't seen for years

      i'm less than 6 degrees from an old country music star and a dentist in florida who's in jail for killing his wife 20 some years ago.

    • Wait, so you're telling me the people that I play Mafia Wars with aren't really friends?!?!!
  • How do they determine real connections vs just having browsed some random strangers' cute/funny/whatever profile and saving it just for that? I'm pretty sure the original theorist would scoff at the notion that's a real social connection. LinkedIn would be much better but even then you'd have to filter out all the recruiters.

    • Oh come on, everyone knows that a "friendship" based on Farmville is just as deep and meaningful as a traditional friendship. None of my "real" friends cared when I got the Golden Tractor...
    • by mfh (56)

      Real science is always scoffed at buy businessmen because it often gets in the way of profit. Profit is only generated by bullshitting people, and science is the act of making people see the facts. No businessman wants you to see the facts or even consider them.

      Think about it. Someone buys something ridiculously cheap and resells it for insane profit. That's crazy! But it happens every day and even SCIENTISTS buy things at five times the standard material price, or more! Except of course, DIY people. These

      • by magarity (164372) on Tuesday August 16, 2011 @11:08AM (#37107336)

        Get help before you end up like Kaczynski.

      • I think it really depends on what you are trying to build... if you want something that fits more precise specs, you are unlikely to even match off the shelf equivalents. Even building a PC at near what some of the OEMs do is difficult, and pretty much impossible in the low-mid range if you include a $100 OEM license. Making a table/couch/cabinet will often cost more in materials alone, let alone time, and investment in equipment needed. Having a hobby, or being an enthusiast doesn't mean it's less expen
    • The article mentions that messages will be sent to the friends. I suppose the theory is that a stranger will just ignore the request while someone with a true connection will participate. It is not just parsing the map of connections among facebookers.
  • Sounds Great! (Score:4, Insightful)

    by jimmerz28 (1928616) on Tuesday August 16, 2011 @10:33AM (#37106906)

    Oh whoops while we were performing this test we accidentally shared a whole bunch of private information with our partners.

    • by alphatel (1450715) *

      Oh whoops while we were performing this test we accidentally shared a whole bunch of private information with our partners.

      ...and transmitted your wifi ssid and geolocation, provided access to your account to a 3rd party without a password, and inadvertently disclosed to law enforcement that you were browsing news articles that favored hacking as a form of right to free speech.

      But it's okay we fixed all that recently.

    • by Syberz (1170343)
      LOL, read the privacy terms, it wasn't accidental.
  • Paul Erdos (Score:5, Insightful)

    by Ecuador (740021) on Tuesday August 16, 2011 @10:33AM (#37106920) Homepage

    The editor should be banished from /. for mentioning the Bacon number and not the Erdos number.

    • by beef3k (551086)
      Yes. Slashdot - a place with noone but maths geeks with no other purpose in life than achieving the highest possible Erdos number. (who's this Kevin Bacon anyway...?)
      • by szo (7842)

        Smallest. Shame on you for this!

        • by vlm (69642)

          Smallest. Shame on you for this!

          Anybody can get a low erdos number. I have a semi-distant acquaintance with a "3" so it would not be terribly difficult for me to score a "4". Perhaps I could help her prove something on a compute cluster, to get credit on one of her papers.

          The real challenge is getting a high number. You must publish or perish. To publish you'll probably have to collaborate, what comes around goes around and if you want to be listed on 10 papers that you didn't do much on them, you've gotta accept ten freeloaders on yo

          • by Ecuador (740021)

            Anybody can get a low erdos number. I have a semi-distant acquaintance with a "3" so it would not be terribly difficult for me to score a "4".

            That's because of your definition of "low". A 4 is low for the number of kittens in cat lady's house, or the magnitude of an earthquake, but it is not low for an Erdos number, with the median being 5 and the average 4.65. So 4 is, as you too realize, very common and thus is far from being considered "low".
            Low Erdos number means 1, 2 or at most 3.

          • The real challenge is getting a high number.

            No, the real challenge is getting a particularly low (or particulary high, but defined) Erdos-Bacon number.

  • I wonder how they get around computational complexity of this kind of thing. I bet it doesn't try to find the smallest number of degrees of sepperation. Or since it is between you and a stranger it just does a random walk and returns who ever is at the 6th position who isn't in either of your fiiends list. http://en.wikipedia.org/wiki/Travelling_salesman_problem [wikipedia.org]
    • by Ragondux (2034126) on Tuesday August 16, 2011 @10:49AM (#37107118)

      It's not a travelling salesman problem, it's a shortest path problem, and as such is much easier. For the distance between two specific people, you'd need the Dijkstra algorithm, and for the distance between any two people, you could use Floyd-Warshall. This one is in O(n^3), where n is the number of users; that's a big number, but it's nowhere near the (supposed) complexity of the TSP.

      • by DavidShor (928926)
        Right. Probably the best way to do it is to randomly draw a couple thousand UID's and run Dijkstra. Then you can do all sorts of neat stuff like geodesic distributions.
      • by Rockoon (1252108)
        ..and it wont actually have a runtime of n^3 steps if the 6-degree theorem is correct.. O(n^3) is the worst case, while it would actually perform like O(n^2) when given high levels of interconnectivity
      • The fact that it is undirected and unweighted makes it even faster -- sub O(n^3) by at least one or more log factors depending on the density of the grpah. The shear size of the graph would make for a fun implementation, but it is entirely doable.

      • Ahh, I remembered the problem wrong. I should read my own links lol. To make it a traveling salemen problem you could do the following. Assuming everyone who is friends with someone on your friends list will accept a friend request. What is the shortest amount of friend requests such that you would become friends with a set of people on Facebook.
        • Even that is not the same because with the traveling salemen your options for the next step is smaller with each move. Where as this your options for friend request would grow with each request. Not the same at all.
    • by grumbel (592662)

      Travelling salesman doesn't seem to be an issue here, as the six-degrees of separation thing is a simple exercise of finding the shortest path between two people, not finding the shortest path crossing N people (with large N). And path finding isn't very complex, just throw some Dijkstra's algorithm [wikipedia.org] at it and be done with it.

      Of course what they are doing in the "experiment" (ad campaign?) isn't actually finding the shortest path, its about letting the users themselves try to find it by sending messages via

  • My friend and I used to play this "connections" game, mostly with movie stars, and always with funny variations like "Connect Tim Burton to Orson Welles, using the movie The Cannonball Run." We almost never needed six degrees to make even the most obscure connections. My favorite example was one we did when Rob Morrow played John Wilkes Booth in a miniseries: "Connect Rob Morrow to John Wilkes Booth." Seems hard no? Not really:

    John Wilkes Booth's brother Junius was married to Agnes Booth, who was in the Pal

  • In real life as soon as you know a person who has met the Queen or the Pope, you then know a lot of people via them. How strong does this connection need to be?
    • I understand your idea, but it mistakes meeting someone with knowing them. Just like Facebook males the mistake of calling people "friends" just because they have something (imaginary or real) in common. While you might think you "know" the queen, just because you've met her there's no reciprocity in the relationship - she does not know you. So the premise falls down.
      • I thought that was about it. The Facebook data doesn't prove a thing if they take into account all those people who have 500+ friends.
        Just to point out I've never met the Queen and have I never "known" her.
        the Pope £uc$ed me once.
  • by lkcl (517947) <lkcl@lkcl.net> on Tuesday August 16, 2011 @10:48AM (#37107116) Homepage

    did facebook and yahoo ask permission of all these users before rifling through their profiles?

  • Using Facebook to test this theory seems kind of dumb. I'm Facebook friends with the Ecumenical Patriarch of Constantinople, so it would appear that there is one degree of separation between me and his All Holiness. And because Ecumenical Patriarch Bartholomew is acquainted with Pope Benedict XVI, then there are only two degrees of separation between me and the Pope -- or between me and any number of world leaders or other important people. But, of course, I have never met the Ecumenical Patriarch, so yo

  • That's exactly the original purpose of Orkut, the (almost defunct) original Google social network site. Orkut Buyukkokten (the Google employee that came with the idea on his 20% paid "free time") was the "Kevin Bacon" and the breadcrumbs of the site used to show the smallest paths between you and Orkut.

    It was a fun experiment in the beginning because not only the path between you and Orkut was drawn, but the path between you and anyone you looked at the profile.

    But then, when the first batch of geeks
    • by 787style (816008)
      There are markets other than America. Orkut is huge in Brazil and India, over 60 million users and one of the 100 most trafficked sites in the world.
  • (potential employer, after reviewing the social "map" of said candidate and seeing they're only 2 "degrees" away from a convicted sex molester...) "Well, we would like to hire you, but it seems that after further review, we don't feel you're quite a solid fit for the position."

    Yeah, tell me that kind of abuse ain't gonna start happening...

  • by cyberchondriac (456626) on Tuesday August 16, 2011 @11:24AM (#37107522) Journal
    I guess I'm 2 steps from Kevin Bacon; I fixed his dad's VCR while working at a shop in Ardmore, PA, back in '91. (it was more interesting however that on the same job, I met Patti LaBelle, and talked to her in her own house - very nice woman)
    • I have a personal Kevin Bacon number of 4. I worked on a student film with a friend who later did costuming on a movie featuring Dan Aykroyd. Aykroyd was in 'Chaplin' with Laurel Whitsett, and Laurel Whitsett was in 'Super' with Kevin Bacon.

    • I had bacon-wrapped hot peppers this weekend, so I got you beat.

    • by Randwulf (997659)
      I'm at 3 steps. A friend of my brother used to babysit Kevin Bacon's children.
    • No separation = same movie.
      One degree was with someone who acted with him, etc.

      I supposed IMDB could be topologized to measure two actors' separation.
      • by mcmonkey (96054)

        No separation = You are Kevin Bacon
        One degree = in a movie with Kevin Bacon

        A guy from my high school drama club was one of the muggers in Crocodile Dundee ("that's not a knife"). If plays are included in addition to movies, that gets me a Bacon number of 6 or 7.

        If you don't restrict connections to performances, Larry David's daughter was my R.A. in college. From there, Larry David was in Seinfeld with the guy who played Neuman, who was in Jurassic Park with Jeff Goldblum, and eventually someone was in som

    • Re: (Score:2, Interesting)

      by Anonymous Coward

      That puts all of us at 3 step max because we visit the same website as you. That's about as valid as Facebook's idea of what Facebook friends are.

  • by jarkus4 (1627895) on Tuesday August 16, 2011 @11:34AM (#37107658)

    Everyone here is bitching about privacy breach, algorithm complexity etc. Actually it has nothing to do with this experiment. From TFA
    "Anyone with a Facebook account can participate to verify if everyone is on average approximately six steps away from any other person on Earth. You’ll be asked to select one of your Facebook friends whom you believe is most likely to know the “target person” that has been assigned to you. A message will then be sent from friend to friend until you get it to the “target person.” The goal is to do this in as few steps as possible. "

    Basically they are just repeating the old mail experiment, but with a new way of passing messages
    - unless you (or one of your friends) participates nothing happens to your privacy
    - no computer algorithm is involved
    - no problem with celebrity profiles linking thousands of people that now nothing about each other

    • I actually signed up for it because I'm like that and it turns out you have to provide the information you want to share and you have to send a message to a Facebook friend you think might get you closer to the end target. You do have to let the Yahoo app have access to your basic information on Facebook to sign up for it. If you want to be a target you have to let the app have more access to your information and provide some additional details. So basically, I doubt this will go anywhere specifically becau
    • by mcmonkey (96054)

      Everyone here is bitching about privacy breach, algorithm complexity etc. Actually it has nothing to do with this experiment. From TFA
      "Anyone with a Facebook account can participate to verify if everyone is on average approximately six steps away from any other person on Earth. You’ll be asked to select one of your Facebook friends whom you believe is most likely to know the “target person” that has been assigned to you. A message will then be sent from friend to friend until you get it to the “target person.” The goal is to do this in as few steps as possible. "

      Basically they are just repeating the old mail experiment, but with a new way of passing messages
      - unless you (or one of your friends) participates nothing happens to your privacy
      - no computer algorithm is involved
      - no problem with celebrity profiles linking thousands of people that now nothing about each other

      You're right about the privacy angle, but not about the algorithm.

      It's still an algorithm, just with people making choices instead of the computer.

  • They tell you whether you have any connections in common up to a certain degree.
  • Problem here is that actors, singers, companies, etc, will act as "hubs" - since they have hundreds of thousands of "friends". So probably it's not 6 degrees but 3 or 4 here.
  • Since Godwin's law will not be contradicted, I have to say that Wikipedia has proved rather successfully that you are never more than 4 clicks away from Hitler.
  • This is how it works.

    Suppose everyone knows roughly 50 other people. If you add together your family (all of them - even distant cousins), the people you know at work, church, through your hobbies, your neighbors, your mailman and so on... the total is around 50.

    Every one of these 50 people knows 50 *other* people, and every one of those 2500 people again knows 50 *other* people, so that the circle expands exponentially in powers of 50.

    Of course, this isn't a complete description since at every stage there

  • I always thought that the *maximum* was six degrees of separation between virtually everyone. Thus, an average would lower than six degrees...and any instances of more than six degrees would be in the minority, or outliers that may simply be because not everyone is accounted for.

    Of course, this also depends on what your criteria for each degree of separation. Do you have to be just facebook friends, or should you actually have had some real physical contact (i.e., being the same room together), or verbal

  • You’ll be asked to select one of your Facebook friends whom you believe is most likely to know the “target person” that has been assigned to you. A message will then be sent from friend to friend until you get it to the “target person.” The goal is to do this in as few steps as possible.

    One of the crucial points of the Small World thing is that you can't predict your indirect acquaintances (or even the regions/groups they are in) more than one or at most two degrees away. We don't have some kind of global routing table in our head. I have dozens of friends in the US. It's quite possible that one of them is related to a former classmate of a relative of a former colleague of someone who knows Obama. Even assuming this is the case, I have no way of guessing who that would be.

    Instead of relyi

  • First off, the entire notion of "six degrees of separation" is based on a study that was later partially discredited. Milgram's study was flawed in a number of ways, later described by Judith Kleinfeld [judithkleinfeld.com] and others.

    Secondly, even assuming that the "six degrees" effect was accurate, using facebook is not the same as the initial experiment at all. Facebook is a radically different method of connecting people.

    It's interesting that Duncan Watts is mentioned in the article but that he isn't allowed to actually exp

A penny saved is a penny to squander. -- Ambrose Bierce

Working...