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

AFAIK from past discussions on here, the Hinge model is to make it harder with people with a low Elo on the app, and to provide them with more fine-grained options to pay to boost their chances ("roses" etc.)

If you have a higher than average Elo, it will seem like having a free lunch. But that's because it's fueled by someone else's desperation at being caught in low Elo feedback loop early.

You then have high Elos who cycle through the app quickly and don't pay, and the rest who cycle more slowly and are more likely to pay during their time there.




I assume "Elo" is their hotness credit score system? I doubt I had a high one - kinda nerdy looking mid 40's dude, not particularly tall or athletic-looking, not particularly wealthy, no constant stream of matches, etc.

I found the people in the "rose list" to just look like Instagram profiles with fairly vanilla profiles and standard TV/socmed good looks, whereas I usually fit better with other nerdy weirdos of average/decent looks. It's oversimplification, but that's about as good as I can get as a generalization.

I still spent some weekends taking care of things at home and plenty of others with dates over the course of several months. I met a few lovely people who I'd totally say hi and chat with if I ran into them out and about today. I went out a few times with one or two that eventually fizzled. And I went out with one that I wouldn't have met otherwise (lives 45min away) and am still involved with a year later.

I never paid for it (Hinge) and just treated it as a way to "surface" potentially interesting people to say hi to, chat with, and potentially date. For all the flaws, it's a more effective way to do this than counting solely on meeting people in your physical sphere.


How do they make it harder for people with low Elo? Surely it being harder for low Elos is the natural way of things? They are just cashing in on the fact that low Elos (which are disproportionately male, of course) will get desperate and more likely to try the "premium" options.


Attractive people will have options, so the app has to show them other attractive people to retain them. That's why there is a containment system where if your Elo drops low enough you get shown other low-Elo accounts. You also need attractive people to retain the low-Elo accounts (Elo being a relative measure to the userbase, by definition) so you can seed them with attractive people once in a while but generally keep them confronted to low-Elos. Ideally, you want to create an impression of potential at first, which will get the user to spend in the face of loss aversion as the Elo bucketing sets in.

That said, this is not the only thing happening, and the platform apparently also does the opposite, or does batches (exclusively low-Elo followed by exclusively high-Elo for a while). At the end of the day, it's an ML blackbox.


One of the main criticisms I see is that low-Elo people also get show mainly high-Elo people. So there's much less chance for two low-Elo people matching with each other, the algorithm works against it.

Since you have a limited number of "likes" per day, you have to actively "say no" to people that's attractive to you.


Yeah, that sounds about right. I am actually single again (ha...), so the quality of the match is not gonna be as deep as a compatability based algo like OKC - Hinge profiles are relatively shallow, and hard to get right. It's much better for getting laid, not finding the love of your life.

I don't plan on returning to any apps, because the whole experience feels dehumanizing, and I'm looking for something enduring now.


Isn't that exactly how Tinder also works, or at least used to work? I remember Tinder being heavily criticised for using Elo ranking.


We'd have to know the algorithm but it's all about the details of how that Elo is implemented, which traits are selected for in which apps, the monetization options etc.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

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

Search: