Are you clickier than a 4th grader?


  • Trolleybus Mechanic

     My wife forwarded me this dumbshit experiment that's polluting her school's newsletter. An experiment that "can tell you" how old you are based on how you click things on screen.

    Both her and I, with vastly different computer usage styles, were told we are 10 years old. Given the utter retarditude of the premise and execution, this means either that the researchers are right at this very moment spending grant money on some ground-up monkey spleens that they'll snort out of a hooker/co-ed (cause there's a difference, right?) eyeball-- or there's some fantastic double-blind switcheroo meta-experiment going on for which this jizz-curtain is just a veil.

    The link: http://kgajos.eecs.harvard.edu/ag/

    The evidence for the "brilliant mastermind" case: It's hosted on Harvard servers and-- umm-- that's all I got

    The evidence against: 

    1) Click on the link. Do you get a timeout? I did. Several times. Should we allocate additional server resources for a world-wide study? NAH!

    2) Despite using jquery 1.9, there is this brilliant warning right on the front page: "This test will not work on mobile devices (including iPads and other tablets). You need to use a regular computer or a laptop. This test will also not work with Internet Explorer 8 or older."

    Bathe in that for a moment. Given that all the test really does is track a path from one clickpoint to another, they can't get it to work in IE8.  I could almost understand the iPad if they're tracking mousetrails-- but what's so special about IE8 that doesn't provide them with the details they need?  "Hey, we could put a bit more effort into this-- or we can eliminate a large chunk of our data."  Not to generalize, but I would suspect that people who voluntarily use Firefox and Chrome are, on average, of a different technical and computer-savvy level that people who use the default browser. Eliminating those people won't bias your stats in any way, will they?

    3) Do the experiment. Get your age. Also get a comment box to send them some feed back.  Also get a hidden field with your age value in it.  Should we save the age in a session variable of some sorts between the data analysis page, and the finalize page? WHAT'S A SESSION? Dump it in a hidden field. I'm sure no one will fuck with it for the lulz.

    4) Given the technical competence that went into this experiment, I have to wonder how many people's age are "); DROP DATABASE;--"?


  • Considered Harmful

    It says I'm 31. I feel insulted.



  • @Lorne Kates said:

    1) Click on the link. Do you get a timeout? I did. Several times. Should we allocate additional server resources for a world-wide study? NAH!

    It worked for me
    @Lorne Kates said:
    3) Do the experiment. Get your age.

    It says I'm 33, how quaint
    @Lorne Kates said:
    I have to wonder how many people's age are "); DROP DATABASE;--"?

    Dammit, I'm doing it again ;)



  • I happen to know one of the researchers involved in this study (LabInTheWild). A couple of things to keep in mind: First, the occasional bad guess doesn't meant the algorithm doesn't work. Of course you can't guess someone's age with 100% accuracy just from their mouse movements, but if you can get a fairly good estimate for a significant proportion of people tested, that's useful in itself.

    Second, projects like this are almost always coded by a Masters or even Bachelor student, and code quality for the non-core parts of the system is usually not a priority. Using a hidden field may not be good practice, but as long as it works, who cares? The proportion of people who'll know how to manipulate it is probably insignificant (as you noticed, the game is so popular it has taken down the lab's server multiple times). I haven't looked at the code, but I'd have thought the information they're collecting is not just the estimated age, but the data going into those calculations; in that case, it doesn't really matter if people manipulate it. If it's a security hole that can affect their results, then yes, it should of course be plugged.

    Finally, the team says: "To guess your age, we look at over 60 factors. Speed definitely matters, but it is only one thing that our software looks at. Other factors include the duration of tiny pauses people make while moving the mouse pointer, the timing of the peak velocity, the shape of the movement, and many others" So it's not just a matter of tracking a path from one clickpoint to another, and there may be legitimate reasons why they couldn't support IE8 (without disproportional effort). Also, the head of the lab, Krzysztof Gajos, is an expert in how different factors affect ergonomic/mechanical performance of various computer tasks, so I'm sure he has considered how the selected and self-selected group of test-takers will or will not bias the results.



  • @Lorne Kates said:

    2) Despite using jquery 1.9, there is this brilliant warning right on the front page: "This test will not work on mobile devices (including iPads and other tablets). You need to use a regular computer or a laptop. This test will also not work with Internet Explorer 8 or older."

    Dunno about the IE < 8 business, but the touchscreen thing makes utter sense: they are testing your hand-eye motions with regard to shifting the cursor across the screen, so even if a tablet would run the js, it wouldn't measuring the same thing at all (because there's no distance to traverse, you're just stabbing at the screen).



  • It seems like this thing doesn't work on Chrome, either.



  • @Cad Delworth said:

    It seems like this thing doesn't work on Chrome, either.

    Works fine in chrome and was pretty damn close for me (34 vs 33).



  • @Lorne Kates said:

    1) Click on the link. Do you get a timeout? I did. Several times. Should we allocate additional server resources for a world-wide study? NAH!

    I didn't have any problems with timeouts.
    @Lorne Kates said:
    2) Despite using jquery 1.9, there is this brilliant warning right on the front page: "This test will not work on mobile devices (including iPads and other tablets). You need to use a regular computer or a laptop. This test will also not work with Internet Explorer 8 or older."

    I have a feeling the first part (no mobile devices) is intentional. It's a very different skill to touch a circle on a tablet than to click on one on a screen using a mouse.
    @Lorne Kates said:
    3) Do the experiment. Get your age.

    It says I'm 30 years old.

    I'm not even 20 yet.



  •  That was fun. According to the test I'm 29.  The testers also don't seem to know about trackballs.  I don't know if that made the test harder or easier for me (I'm going to say harder--although I scored under my actual age, LOL!).  :)  Hey, works in FF.



  • Once I got past the million page resets it worked fine. Told me I'm 30. Actually I just turned 25.



  • That was kinda fun. Interesting results though: It said I'm 32, but I'm 23. Considering I play a lot of FPS with mouse+keyboard, I don't think that's quite right...



  • It's ironical that I only had connection resets when posting the form, after stating I hadn't experienced any technical difficulties.

    They don't post-get-redirect. I resent the form a bunch of times then gave up.

    After a while I thought "hey, I'll give it another try, and I'll leave a note saying I've done the test before". That time the test stopped at the "You have completed the test. Thank you!" end screen without showing me the survey or any results. I guess I'll never know what age they think I am.

     

    Slightly but not completely off topic, have you seen http://donottouch.org/ ? Including the web console?



  • @lethalronin27 said:

    That was kinda fun. Interesting results though: It said I'm 32, but I'm 23. Considering I play a lot of FPS with mouse+keyboard, I don't think that's quite right...
    Said I was 32 also.  Off by more than twice what it was for you.



  • It gave me 31, which is correct.

    I am going to assume it now returns a random result from {29,30, 31, 32}

    @lethalronin27 said:

    That was kinda fun. Interesting results though: It said I'm 32, but I'm 23. Considering I play a lot of FPS with mouse+keyboard, I don't think that's quite right...
     

    I don't think gaming is precisely the same as clicking targets.

    @Ben L. said:

    It says I'm 30 years old.

    I'm not even 20 yet.


    It means you're already old and past your prime.



  • @Zecc said:

    ironical
    wut


  • Considered Harmful

    [quote user="Grammar Girl"]

    "Ironical" is a standard word—it is an alternative adjective form of "irony"—but it means the same thing as "ironic."

    "Ironical" is the more old-fashioned form of the word, and "ironic" is the more common form today. You can choose to use either one, but as Highway Star's question shows, people are more likely to look at you funny if you use "ironical."

    "Ironical" is more common in Britain than in America, so that may be why it shows up in the work of English writer Ian Fleming, the creator of James Bond.

    [/quote]


  • "Ironical" is also untouched by Alanis Morrihoweveritsspelled.


  • Considered Harmful

    How... ironical.



  • @lethalronin27 said:

    That was kinda fun. Interesting results though: It said I'm 32, but I'm 23. Considering I play a lot of FPS with mouse+keyboard, I don't think that's quite right...

    The average gamer is 32 these days so it thinks you're average!



  • @Zecc said:

    "Ironical" is also untouched by Alanis Morrihoweveritsspelled.
    Ironical is a brand-name of a low-calorie dietary supplement for people with anemia.



  •  It guessed me as being 31. I'm in my late 40s.



  • @lethalronin27 said:

    That was kinda fun. Interesting results though: It said I'm 32, but I'm 23. Considering I play a lot of FPS with mouse+keyboard, I don't think that's quite right...

    Yeah, your wrist is just shitty now. See that dark area there... all yeuch and hardened... they get like that, they're not good any more.



  • @dhromed said:

    I am going to assume it now returns a random result from {29,30, 31, 32}

    So it's not so much of a "how old are you based on this test" as it is "how old are you based on statistics about people who take this test, disregarding the actual results".



  • Make that {29,30,31,32,33}.



  • I have a beef with their "make your window as large as possible" instruction. I think they vastly underestimate how big my monitors are.

    It read about 5 years too old. I attribute this mostly to me not really giving a shit and not making any effort to complete it as quickly as possible. Or possibly the fact that my monitor is fucking huge.

    Loaded and posted form fine.

    It's broken if you've changed your Windows DPI, but that's Chrome's fault (probably) not the test's. (Yes, Chrome still doesn't correctly handle Windows DPI set to anything except default. Wankers.)



  • @blakeyrat said:

    I have a beef with their "make your window as large as possible" instruction. I think they vastly underestimate how big my monitors are.

    It read about 5 years too old. I attribute this mostly to me not really giving a shit and not making any effort to complete it as quickly as possible. Or possibly the fact that my monitor is fucking huge.

    Loaded and posted form fine.

    It's broken if you've changed your Windows DPI, but that's Chrome's fault (probably) not the test's. (Yes, Chrome still doesn't correctly handle Windows DPI set to anything except default. Wankers.)

    And what exactly do you think Chrome should do when you change the DPI? Change the size of a pixel?

    Is this text the same size as this text?



  • @blakeyrat said:

    I think they vastly underestimate how big my monitors are

    When they asked me about medical conditions I put down "Fapism". I see you suffer from the same thing.



  •  It guessed that I was 44; I'm actually 43. I didn't take any of those quick breaks, but perhaps I should have mentioned that I do have a medical condition that affects this test, because my retinas are pretty much shot, peripheral vision is not very good and often I lose track of the mouse pointer, although not during this test.

    Pretty interesting concept, though.



  • @dhromed said:

    I am going to assume it now returns a random result from {29,30, 31, 32}

    Might not be far from the truth. I got 30 as a result. Which, for the record, is off by two years. Not bad.

    Anyone else not seeing a submit button? After a few minutes, I did a simple 'inspect element' on the text field that said it was waiting for results, found a submit button with a display=none, enabled it, submitted the form and got my results. Seemed odd. There was also a random quotation mark at the end of that text. Shoddy coding right there.



  • @FragFrog said:

    Anyone else not seeing a submit button? ... There was also a random quotation mark at the end of that text.

    Yeah, I saw the same thing. I dealt with it by pounding on Back and Reload until it came good. Maybe that's how I lost 18 years, though there was some kind of test ID included in every URL beyond the first that was long enough to be plausibly unique for me.



  •  Looking at the responses in this thread, this might just be a giant ploy to get the real ages of the people on this board.


  • Discourse touched me in a no-no place

    @dhromed said:

    It gave me 31, which is correct.

    I am going to assume it now returns a random result from {29,30, 31, 32}

    Don't think so. Gave me 37, which was only 4 years out.


  • Discourse touched me in a no-no place

    @FragFrog said:

    Anyone else not seeing a submit button?
    I just left the page open for ~5 minutes, and when I went back the button eventually appeared.



  • @Ben L. said:

    So it's not so much of a "how old are you based on this test" as it is "how old are you based on statistics about people who take this test, disregarding the actual results".
     

    So far, it's really poorly calibrated.



  • I tried it a while ago (days? a week?). It alleged I was 42, which is, of course an excellent choice of number. In what passes for reality around here, I am cough somewhat older.



  • @Lorne Kates said:

    1) Click on the link. Do you get a timeout? I did. Several times. Should we allocate additional server resources for a world-wide study? NAH!

    I got "Error 1: Can't log the fuck in."



  • @da Doctah said:

    @lethalronin27 said:

    That was kinda fun. Interesting results though: It said I'm 32, but I'm 23. Considering I play a lot of FPS with mouse+keyboard, I don't think that's quite right...
    Said I was 32 also.  Off by more than twice what it was for you.


    Hmm, the difference is 18 or more so you are less than 14 years old? Is that your actual picture?



  • @Lorne Kates said:

    Given the utter retarditude of the premise and execution, this means either that the researchers are right at this very moment spending grant money on some ground-up monkey spleens that they'll snort out of a hooker/co-ed (cause there's a difference, right?) eyeball-- or there's some fantastic double-blind switcheroo meta-experiment going on for which this jizz-curtain is just a veil.

    Now, why can't your front page articles contain such eloquent prose?


  • Trolleybus Mechanic

    @Sutherlands said:

    @Zecc said:
    ironical
    wut
     

    Ironical is a dollar-store knockoff of [url="http://en.wikipedia.org/wiki/Bionicle"]a more popular toy line[/url]. Contrary to expectations, the knock-off is a much better toy.



  • Gee, I thought it was interesting.

    It guessed that I am 38, but I'm 58, so it also put me in a good mood. :)



  • Took several minutes for the site to initially load. Took several more minutes for the test to actually begin. Made it through pretty quickly, filled out the info at the end, tried to get to my results and after a few minutes I got "Connection reset by peer". Doesn't seem like a very resource-intensive website so I don't know wtf was up with the painfully slow performance.

    I sincerely hope this is a meta-experiment like Lorne mentioned, because otherwise it sounds like the worst way to estimate someone's age.



  • It said I was 42... not even close.

    Also it doesn't work in IE < 9 because it uses a canvas to draw those pretty circles.



  • @serguey123 said:

    @da Doctah said:

    @lethalronin27 said:

    That was kinda fun. Interesting results though: It said I'm 32, but I'm 23. Considering I play a lot of FPS with mouse+keyboard, I don't think that's quite right...
    Said I was 32 also.  Off by more than twice what it was for you.

    Hmm, the difference is 18 or more so you are less than 14 years old? Is that your actual picture?
    Other way, whippersnapper.

    Perhaps it's very accurate but my clicking style hasn't changed since 1991?

     



  • 30 for a 35er not bad... when i'm doing this in a HD 23'' monitor!!!



  • It thought I was 15 years younger than I actually am, and I'm in the middle of acclimating to not using mouse acceleration (just an experiment to see if I can mouse more accurately without it), so... yay me?



  • @Lorne Kates said:

    Ironical is a dollar-store knockoff of a more popular toy line.
     

    I once tried to show someone what I considered a very funny dollar-store knockoff of a popular brand of cat food.

    "Look at this", I said.  "Some poor foreign schlub didn't understand that 'Nine Lives' is a common cultural reference involving cats, so they decided to raise the stakes and market '10 Lives' instead.  Sort of like that 61-star made-in-China US flag that turned up a few summers ago.  And what's worse, they made the label black, not realizing the morbid implications for your kitty."

    I grabbed one of the cans to show it closer up.  Amazingly, there was no '1' before the zero.  "Zero lives?!  That's even worse!  Can you believe someone would actually make it onto the shelves with such a poorly named product?"

    She looked at the can, then at me, then back at the can, then back at me.  "I don't get it.  Something wrong with this can of OLIVES?"



  • @da Doctah said:

    @Lorne Kates said:

    Ironical is a dollar-store knockoff of a more popular toy line.
     

    I once tried to show someone what I considered a very funny dollar-store knockoff of a popular brand of cat food.

    "Look at this", I said.  "Some poor foreign schlub didn't understand that 'Nine Lives' is a common cultural reference involving cats, so they decided to raise the stakes and market '10 Lives' instead.  Sort of like that 61-star made-in-China US flag that turned up a few summers ago.  And what's worse, they made the label black, not realizing the morbid implications for your kitty."

    I grabbed one of the cans to show it closer up.  Amazingly, there was no '1' before the zero.  "Zero lives?!  That's even worse!  Can you believe someone would actually make it onto the shelves with such a poorly named product?"

    She looked at the can, then at me, then back at the can, then back at me.  "I don't get it.  Something wrong with this can of OLIVES?"


    What the hell is a Iolives?


  • Trolleybus Mechanic

    @Ben L. said:

    What the hell is a Iolives?
     

    An abstract olive interface.



  • @Ben L. said:

    What the hell is a Iolives?
    Something visible only to a reader whose choice of font doesn't distinguish 1, l and I.  If you people would quit being so stubborn and just switch over to Comic Sans, we wouldn't have these legibility problems.

    (ObDigression: someone asked if my avatar was a picture of me.  It's not.  I harvested it from the Web many years ago to illustrate what I visualize when I hear the expression "out of the mouths of babes".)



  • @da Doctah said:

    (ObDigression: someone asked if my avatar was a picture of me.  It's not.  I harvested it from the Web many years ago to illustrate what I visualize when I hear the expression "out of the mouths of babes".)

    By way of contrast, these are my real shoes and trousers. I also have it on good authority that Ben L. actually does look like a cat faced cupcake.


Log in to reply