Hacker News new | past | comments | ask | show | jobs | submit login

> There was a very detailed article that used data collected in Tinder.

Survery* OkCupid* and 2011*.

> It showed something like 80% of women fought for 20% of top men and other way around.

It showed that women rated 80% of men below average. Which is pretty understandable, in okcupid and in 2011. That is an insane sample bias, when women look at men in ACTUAL research they tend to rate in a bell curve, and are actually kinder to age than men that tend to rate older women super low.

> This basically meant that large number of men are being left out cold in dating scene.

The data from the same survey showed a lot of those women ended up in relationships outside that 20%. Women rate look as the 4th most important thing in a man.

It seems like you probably need to check first hand sources instead of parroting whatever youtuber you heard this from cause you extrapolated conclusions from data you never checked and got all the conclusions wrong.




I used to work at a dating service pre-Web. I was in the thick of the data, so to speak, and noticed a lot of trends which haven't really changed.

1) Men ask women out a lot; women only rarely ask men, and when they do, they tend to all be the same guys. They really converged.

2) Women rated the importance of their preferences higher than men do, at almost every aspect.

3) When someone says, "Oh, women would never actually judge men based on ..." assume the opposite. Income and height were always in the top five.

4) Despite pushing people to match on commonalities of interest and psych stuff, it was the "chemistry" factor that dictated who got asked out, who accepted, and so on.

5) It was in no way egalitarian and always tended toward "grim and depressing truth about people." All of this has only been accelerated in online dating.

Tempted to see if those old floppies survived and if I could retrieve anything in dBase III or whatever it was. I'd still have to find a document stating what those referred to. Sample size wasn't too bad, a few thousand ... The thing is, I could do all that work, and the people who don't want to believe that people are like that? wouldn't believe it anyway. They'd just look for confounding variables and reasons to dismiss.


> The thing is, I could do all that work, and the people who don't want to believe that people are like that? wouldn't believe it anyway. They'd just look for confounding variables and reasons to dismiss.

That is probably the most frustrating thing about all this discourse. When people see the structure of dating (online and otherwise) presented, they ascribe negative traits to the populations whose behavior is described and assume that everyone else would make the same judgment. Because of that, they assume negative motives and take it on themselves to debunk or censor that data.


> That is an insane sample bias, when women look at men in ACTUAL research they tend to rate in a bell curve, and are actually kinder to age than men that tend to rate older women super low.

I think that the survey was pretty controversial and has since been quoted to death in all sorts of contexts. It would be nice to see this actual research which paints a different picture, so that folks here might actually discuss the findings.

I might have nothing to offer there, but people who work with statistics and whatnot would probably offer a good point or two.

> Women rate look as the 4th most important thing in a man.

I do think that the problem with most surveys is that there will certainly be biases in self-reported answers.

For example, if asked about what I'd prefer in a country, I might claim that I value human rights and liberty really highly, but then turn around and vote for politicians that would promise to lower the taxes instead. I would actually be picking a different set of characteristics/priorities than I had indicated previously - regardless of whether I actually believed my own answers and didn't know, or simply wanted to appear virtuous.

(just a made up example, I'm not sure how one would even measure what people actually value outside of asking them, maybe try to classify people based on various indicators)


> I do think that the problem with most surveys is that there will certainly be biases in self-reported answers.

Yeah and it's very typical with this question especially. You can't ask a woman what she values in a partner, because she will say whatever makes her look like a good person.

That's why so many girls list intelligence as an important thing in a man, and still the most intelligent guys get the least girls.

You need to look at how women actually choose their partners, not ask them how they would hypothetically choose. Look at the men that always have hot girlfriends and what they have in common, that's how you can make an empirical conclusion.


This is actually how you feed your confirmation bias, please be careful as a lot of the language you are using makes me worry about you.

There are many explanations for the phenomenon of dating, but your comment reads like a textbook example of someone looking to confirm beliefs.

Why couldn’t attractiveness simply be a cross product of money and physical activity level? In that case, hot people end up together due to economics and interests.


Isn’t that the whole “system 1” and “system 2” thinking? One being your logical brain and one being your practical brain. Hence why some supermarkets are testing shelving layouts in VR because if you survey people they’ll tell you one thing but in practice they’d actually do something different :)


The OkC study (and the Tinder study mentioned in a sibling comment) are pretty much worthless either way, and people should stop citing them. The only really meaningful way to look at these things are measures of the actual relevant outcomes (likes, matches, conversations, dates), which is hard without access to actual dating app data.

An engineer at Hinge did exactly this for a post on the blog, calculating a Gini coefficient for (IIRC) match rates:

https://hinge.co/hinge-reports/whats-the-biggest-challenge-m...

Naturally, it was taken down, but it did show that heterosexual matches were distributed in a much more lopsided way among men than among women. Though that can't be taken as a justification of "20% of the men get 80% of the women", which is either (obviously) false as stated, also true of the distribution of X among women, or incoherent as a meaningful metric.


> Though that can't be taken as a justification of "20% of the men get 80% of the women", which is either (obviously) false as stated, also true of the distribution of X among women, or incoherent as a meaningful metric.

Not that I agree with the quote, but... why would any of these assertions be true?

I suppose that depends on what "get" means in the quote. I think it's immediately evident that, say, 20% of men aren't married to 80% of women, for example -- so if that's what "get" means, then that doesn't add up. But I see nothing inherently wrong with the notion that 20% of men could be taking 80% of women off the market. Not because these men are dating approximately 4 women each at the same time, but because they could be rapidly playing the field. If these men are engaging in short flings with women, it would seem reasonable that the women might not want to settle on anything less than the 80 percentile of men, knowing that they can briefly hold out until another top 20%er comes along. That effectively takes them off the dating market for the other men.

Again, the stats could disagree with such a hypothesis, but I don't think there's any inherent logical contradiction in the hypothesis itself.

If we're talking about hooking up with people met at the bar, I can maybe see the quote making sense. But if we're talking about meaningful relationships, I actually find it hard to believe that a man is having a hard time finding a partner because "20% of men are stealing all of my opportunities." Though that could be easy for me to say if I'm just that damned attractive, but I don't suspect that's the case :).


Noting that we agree that the actual stats would disagree about the hypothesis.

If I might take the liberty of riffing off your model, let's take it as "looking at the the sexual encounters that occurred over a duration on the scale of months, you can select a certain subset (~80%) of women who will have had sexual relationships exclusively with a certain subset (~20%) of men." I'd put that version down as falling into my first category of obviously false. I.e. sexual relationships don't form a disjoint graph, even excluding the edges from that 20% of men to the nonselected 20% of women. I don't think it's inherently incoherent, just wrong.

We can weaken it to "Over some duration on the scale of months, you can select some relatively small subset of men (~20%) who will have been the partners in a disproportionate number (~80%) of sexual relationships." This is likely and plausible, but it's likely and plausible that you could apply that same procedure to women to get a similarly surprising pair of numbers, because of how social graphs work, i.e. the bulk of a measure is driven by the extremes of the distribution. It's interesting, but it doesn't tell us much about the differences in how heterosexual men and women pair.

I'd read your proposed model as stronger than the second version and perhaps a bit weaker than the first. Maybe modifying the first model by adding something like "Half of those selected 80% of women will also have formed a single edge with the nonselected 80% of men." I think this is a viable way to recover it into something more plausible, but in doing so we've the changed the actual claim to something like "you can select a certain subset (~80%) of women who will have had sexual relationships exclusively with a certain subset (~40%) of men." That's a much weaker and less shocking claim than the original, and I wouldn't be surprised if you could again apply that same procedure to women and get similar numbers out of it.

I think we can make statements about the structure of the sexual relationship graph that point to differences in how men and women form sexual pairs, but those aren't as easily understandable. "The distribution of sexual encounters among men have a Gini coefficient of 0.654, while among women it's 0.562" doesn't really make for a punchy soundbite.


But the Hinge data only shows, a fairly obvious in hindsight, observation. That if we live in a society where men are expected to ask women out, the only men women will ask out are those they really wan't to get.

In other words, if you checked for example superlikes which are capped for men, or Roses on hinge, I am sure the same lopsided percentage shows up.


I don't have the actual Hinge post on hand, but if I recall correctly it measured heterosexual matches, which each involves one woman and one man, and it was the distribution of those matches that was more uneven among men than among women. That wouldn't be affected either way by who initiated.


No, I think they're talking about this study: https://medium.com/@worstonlinedater/tinder-experiments-ii-g...

(Or "study" depending on your standards)




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: