This is What I Learned About Humans! (interesting stats on Human Brain Cloud)

kids1.gifMy just-for-fun side project Human Brain Cloud, (online game that gives you a word, you type in the first thing that comes to mind, and it builds a big network of connected words etc) has been up for almost two weeks now, and I’ve noticed some interesting trends. Quick perspective: nearly 800,000 associations have been submitted, connecting a little over 100,000 unique words (or phrases). So given that, here are some interesting stats/trends/etc:

  • The Top 10 Most Submitted Words List. Here are the top 10 most submitted words EVER as of right now:

    Go, sex, go! Woo! If this experiment had an scientific credibility, I’d say humans were more horny than narcissistic or greedy.

  • Phrase Completion. I thought this would be a game just for associating words, but it looks like there is some phrase completion happening too. Some are funny. Some are heartbreaking:
    • “I have…”
      • “…no idea” (3 people)
      • “…a penis” (2 people)
      • “…crabs” (1 person)
      • “…failed” (1 person)
    • “I am a…”
      • “…person” (3 people)
      • “…badger” (2 people)
      • “…miscreant” (1 person)
      • “…hero” (1 person)
  • Numbers. An entire sub-cloud of NUMBERS has popped up. Try this: Start from the number 1 and see how high you can count. I got up to around 30 before the sequence broke down. I guess smallish numbers (<30) are generally seen as counting numbers, where larger numbers are seen more as quantities or have other meaning attached. ie. 1337 -> sp33k, 1984 -> Orwell, etc
  • Marketing Messages in the Public Consciousness, leggo my brain eggo! It is scary how certain words appear to trigger loyal marketing message recitation / product related association or whatever:
    • lego -> my eggo
    • tiger -> tony
    • have it -> your way
    • snap -> crackle pop
    • subway -> eat fresh
    • this is -> SPARTA!
    • (there were a bunch more I’m not remembering – anyone see any good ones?)

    But we sure as hell aren’t gonna be happy about it!

    • “Marketing Campaign…”
      • “…money” (4 people)
      • “…evil” (3 people)
      • “…annoying (2 people)
  • Women Totally PWNd Men. This is the first game I’ve made that I think actually appeals to both women and men. If we look at this game as a battle of the sexes, women have totally dominated in terms of both quantity and especially quality of submitted words. And ok – I don’t actually KNOW who’s a boy and who’s a girl, but browsing some of the names on the leaderboards, it is possible to make some good guesses. ie. Congrats to wandergrrl for being the first person to break 1000 words submitted in a single session (and having some of the highest quality connections too)
  • Racism. Type in just about any racial slur you can think of, and you’re sure to find it. Even the plural spellings. I’ve gotten only one complaint – from someone who found a naughty word connected to his religion of choice – and I’m glad he alerted me to it, because it made me solidify my completely hands-off approach in letting the cloud grow and prune itself.
    People do submit garbage – no question. Luckily, each word has “legitimacy points” attached to it. When a word is flagged by users, it’s legitimacy goes down. As people submit and make connections to and from a word, it’s legitimacy goes up. Racism, sexual insensitivity, etc do tend to go away on their own. I guess I’ll be highbrow here and concede that “just like in the real life Human Brain Cloud of life”, little pockets do keep bubbling up. It reprezents society!

Anyway, on that note, I’ll admit that the number one thing that surprised me right off the bat with this experiment is that people are, in general, overwhelmingly funny, friendly, articulate, and willing to play along. I don’t have a lot of restrictions in place in the game. It would have been very easy to turn Human Brain Cloud into a giant dumping grounds for spam and profanity (and of course there is some), but out of sometimes up to 10 words being submitted per second by players, almost ALL of them are high quality words and connections, firmly in the spirit of the game. So thanks, this experiment has been absolutely worth it, and my cold black heart has thawed just a little. :)

79 Responses to “This is What I Learned About Humans! (interesting stats on Human Brain Cloud)”

  1. Manuel Says:

    This is too funny. And highly addictive
    i see you can change your name at the bottom, but can you use that name again and again? like a kind of login?
    You could give active members with a login rising credibility, according to how strong their words and connections are

  2. The Human Brain Cloud - word association for the masses | BlogSchmog Says:

    […] July, Gabler published some early statistics when the network was about one-fifth its current size. The Human Brain Cloud has the potential to […]

  3. » Have a look inside your brain Says:

    […] Multiplayer Word Association Game (MMWAG?). Macht Laune – probiert es aus. Hilft uns das weiter? Logo! […]

  4. Tim Jones Says:

    This game is doing dangerous things to my neurons! But, speaking as someone born in England who now lives in New Zealand, it would be interesting to know how the network differs according to the country of origin of the players. There are associations that come readily to my mind that might not occur to an American, and vice versa. I’d love to see how the networks created by different nationalities differed, if that was feasible. In any case, this is one of the best and most elegant web applications I’ve ever seen – well done!

  5. William Says:

    I’m a little surprised everyone is referring to this as an original, novel idea. The word association game has been on the web and IM for at least 5 years. I personally wrote an AIM/MSN version over 4 years ago after seeing other poorly implemented versions.

    To the creator: there should be no need to flag wrong words or misspellings. The brain should be able to do that on its own, through the link strength mechanism you are describing. Junk works, spam, misspelling should all be filtered to the bottom and appear rarely.

    Also, I have to say the brain isn’t very smart. From ‘fruit’, it came up with ‘imbibed’, from ‘drunk’, it came up with ‘430kmh’, from ‘scientist, I get ‘cabbage patch’. Am I missing something? Neither of those make any sense. I can see that it is showing common matches on the right side, but is it actually trying to make smart replies or not, because it clearly isn’t.

    Good try, the interface is simple and nice, but the game could use some “brains”.

  6. Kyle Gabler Says:

    Thanks guys, several people have requested I add support for other languages, so that will be the next big thing when I get some time.

    William – Cool, how did the aim/msn version work out? Yeah, link strength alone is absolutely enough to filter the words. The flagging buttons simply help junk words move away faster, not to mention they act as a bit of an apology to players who might be offended when a naughty word pops up and they want revenge. Revenge in the form of a spam button. It works well for everyone. :) Hmm.. not sure where you are getting those connection results ie. for “fruit” – were you using the “view” page? Try this

  7. William Says:

    Kyle, the aim/msn version worked great, but 90% of the time, I was in a battle with flooders and spammers. Spam was easy to deal with using token and link strength to weed out the junk. But flooders were difficult because, especially with the AIM version, you had to deal with message limits. Anyway, a web version is definitely the way to go. And now I understand the usefulness of your spam button. Revenge can be sweet. :)

    The view page works great. I get words definitely associated with ‘fruit’, like ‘mango’, ‘kiwi’, etc. But with game play, I type in ‘fruit’ and get ‘figure skater’. Again and I get ‘eight nine ten’. I don’t know what’s up with that.

    BTW, if you’re interested, I have a Lost (the tv show) themed version of my word association game at – it isn’t as clean and organized as yours, but it works pretty well. Give it a try if you can. Thanks.

  8. Kyle Gabler Says:

    William – oh yeah, that’s because in the “play” mode, the words you receive are completely unrelated! :) If they were related, play sessions would follow a chain of consciousness, and I wanted to avoid that. My reasoning was that the network would grow wider and more reliably if input words were randomly selected.

    Interesting Lost associator – I hope they are paying you!

  9. William Says:

    Kyle – aha, well that explains what’s going on. I can understand where you’re coming from with wanting a very broad dataset, but isn’t the point of the word association game to demonstrate a chain of consciousness between the user and the game? Obviously, most of the players here are having a good time with the game as is, but for me, if I’m going to be teaching the bot, I’d like to see him demonstrate some intelligence in return. And when I see random tokens come up, it goes against what I’d expect. Humbly, I don’t see the fun in the game unless it is trying to respond in a smart way.

    One of the most interesting things I got out of the Lost association game and my earlier wide-open versions was watching them learn from an empty dataset. I gave it a dozen seed words and then let it learn. It was stupid at first, but depending on the number of players and the minimum token and link strength I required, it learned very fast. And to watch it reach the “tipping point” where it started acting smart was pretty exciting. Anyway, I’m rambling now. :)

    Did you write the Flash link viewer yourself or is it open source? I’m interested in implementing something similar for my games, but I’m a humble programmer and don’t know flash.

  10. shino Says:

    i got the new phrase

    ‘i like pie’


  11. reece Says:

    I assocated “zag” with “zig” and “scoot” with “around the corner” and they’re ALL in the song “bop to the top!”

  12. lon Says:

    From what I’ve read about this experiment/game, it really sounds like people had a blast playing it. I just wish I found out about this mini-game sooner. =(

  13. Jerry Lee Lewiston Says:

    “The Internet is an elite organization. Most of the population of the world has never even made a phone call.”

    Something to keep in mind about the “human nature” question and the top ten. This isn’t the human race you’re sampling from, but narrower group with a strong background. The narcissism is disappointing as ever, but not surprising.

  14. Peter Says:

    Hey Ive been following you for a while, and I love your style, both artistic and creative, Im a fellow EGPer and I was wondering if 2dboy is acepting interns? Now about the game, I had alot of fun playing with a couple of my friends over skype, and Id love to see multilingule support

  15. KILLroy13 Says:

    LOLest Game ever.. =)

  16. Crank Says:

    I made a new game,

    1) think of a word
    2) go to view cloud
    3) try and find it using word association.

    Excellent site, great fun.

  17. Ladyluck523 Says:

    I’ve always enjoyed words and language. This is a fun way to see how you ‘tick’. I’ll be spreading the uh…word :)

  18. Le zapping Google du 21 février 2008 | Qualis Opinions Says:

    […] un calamar géant ! Il y a des statistiques d’utilisation de cette sympathique machine sur le blog de l’auteur mais aussi en cliquant sur l’onglet […]

  19. Niru Says:

    Ooh, this sounds great – I await the game’s reappearance with bated breath. :D

    Until then I’ll mope around some.

  20. Sid Says:

    Strangely addictive – Damm you

  21. headlight Says:

    I love the work you do! I love the Zoners who get this far too!

  22. Nostril Soup Says:

    Love the game… must have typed thousands by now! It would be great if you could give us full access to the data somehow though… so we can search for the number of times a words or a specific connection (not just yours!) has been submitted.

    Also would be GREAT (and would require a LOT of work on your part) if you take IP origin into account, and colour the connection lines to show which country has made the strongest connection between the two words. The british, given the word “chips” would probably associate “fish”. Americans (i’m guessing) would favour “dip”. And.. y’know… it’d be nice to have that reflected.

    What else do I want? Ah… yes – the ability to see users who have the highest number of matches phrases with your own, and their nationality. And pic and phone number. Not that I’m a lonely wordophile or anything.

    Also – an additional mode of gameplay, so you can either play “random” (as it is now) or “connected”… so the word you suggest affects the next word your given, and moves along the chain (semi-randomly). If you enter a new word or connection, the chain is broken and you get given a new word.

    Sorry to bother you with my demands, but when I rule the world I’m going to order you to do it anyway, so you may as well get it over with now and avoid the twenty lashes when you’re not programming fast enough.

  23. Webstruxure » Blog Archive » Addicted to Words? Then Beware! Says:

    […] app was written by Kyle Gabler, and his blog has some interesting musings on what he’s learned from the game. Least surprising fact: “sex” is the word that’s been submitted most often. Most interesting […]

  24. Wetterdew Says:

    I want to try it, please…

  25. Ian Flournoy Says:

    Hey I think this is an excellent experiment ….. One thing though can I see the PHP source? :D

  26. Paul Says:

    The site is gone. Safari can’t load it. :(

  27. Jules Says:

    Where did it go? Is it gone forever? I just came across a reference to it and it sounds fascinating!

  28. moonvalleymama Says:

    I miss you. Where are you??? Have you totally forgotten about us cloudy brains??

  29. Josefa Buckle Says:

    Dans la recherche de sites liés à l’hébergement web et plus spécifiquement la comparaison plan d’hébergement web Linux, votre site a été soulevée. :)