wedrifid comments on Open Thread, September, 2010-- part 2 - Less Wrong

3 Post author: NancyLebovitz 17 September 2010 01:44AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (858)

You are viewing a single comment's thread. Show more comments above.

Comment author: Relsqui 25 September 2010 12:03:50AM 3 points [-]

While OKTrends does contain a lot of interesting, well-explained, and often entertaining statistics, I would be very cautious about mentioning it in the same breath as anything to do with how to attract people. The folks at OKC can describe the kind of photos which get a user lots of messages--basically, pointing the camera down your cleavage if you're female, cropping your head out in favor of your abs if you're male--but a fat lot of good that does you if you have a small chest or a belly. (Or if you want to get messages other than "UR HOT WANNA GO OUT 2NITE.") Similarly, knowing that "vegetarian" is a first-message keyword which disproportionately leads to conversation is not all that helpful if you aren't one.

Relatedly, their data is collected from a large and surprisingly mainstream userbase; if LWers are as atypical as we say we are, very few of those users are going to be similar to whomever you're trying to attract. Maybe you know a geeky, intelligent woman who wouldn't roll her eyes at the headless ab shot, but I don't. If you don't believe that the target audience of OKC is very different from LW's readerbase, read the comments on any OKTrends post. They're not all inane, but ...

(I suppose I'm making the assumption there that LWers looking for dates are more interested in someone smart and with common interests than someone who relies on having tits to get attention. If I'm wrong, feel free to disregard.)

By the way, just in case you don't feel insecure enough yet, OKC claims to be quietly segregating its users by hotness. I've seen it theorized that that was a publicity stunt or a sneaky way to pull back inactive users, which seems quite plausible to me, but doesn't make the stunt any less scummy.

Don't get me wrong; I like and use OKC. Just remember that, no matter how clever and statistically sound their algorithms are, most of their data still comes from people who think that what checkout stand magazines say about people, dating, and sex is actually the gospel truth. The site's judgments are based on that standard. So don't take them too seriously.

Comment author: wedrifid 25 September 2010 12:25:44AM 4 points [-]

By the way, just in case you don't feel insecure enough yet, OKC claims to be quietly segregating its users by hotness. I've seen it theorized that that was a publicity stunt or a sneaky way to pull back inactive users, which seems quite plausible to me, but doesn't make the stunt any less scummy.

That seems to be a desirable outcome and one that I expect would be the natural outcome from applying statistical measures to interaction patterns. I expect and prefer OKC to provide matches that are most likely to lead to rewarding interactions. These do tend to be more likely between people of approximately equal hotness.

Comment author: Relsqui 25 September 2010 12:33:25AM 0 points [-]

That's conventional wisdom, certainly, but I'd be interested in seeing an actual study.

Comment author: mattnewport 25 September 2010 12:59:31AM 2 points [-]

Who You Find Attractive is Based on How Hot You Are. I couldn't find a published study with a couple of minutes but I recall seeing experimental evidence for this before.

Comment author: Relsqui 25 September 2010 01:11:05AM 1 point [-]

Thanks, that was interesting. One of his conclusions has a funny side effect: if you care a lot about how hot people are, you're probably pretty attractive. ;)

Comment author: wedrifid 25 September 2010 12:58:49AM 0 points [-]

Have you tried google? I've seen several but do not have the links on file.