Is the Hinge Algorithm Broken?

Probably not, and nothing in Hinge's pages says it is. Hinge describes its recommendations as built on deep learning to predict mutual compatibility, and says the system keeps learning from what you do. When the feed looks wrong, the cause is usually what you taught it, the settings you gave it, or the short window it has had to learn you.

Hinge says it rolled out global changes to its core recommendation system in Discover. The updated algorithm is built on deep learning to predict mutual compatibility. The aim is a feed of people who fit your preferences and whom you fit too.

Hinge also says the system uses information you provide or create by using the app: your age, gender, location, preferences, and who you like, skip or match with. That last group is the part you control every day, and it is where most "broken" feeds come from.

If you are new to the app, Hinge says matches can take time because the algorithm is always learning about you and your preferences. A thin first stretch is part of the design.

Three things cause most of the complaints.

You liked too freely. Hinge advises sending likes to profiles you would genuinely match with, instead of liking a larger quantity. The algorithm learns from your activity. A man who taps on nearly everyone teaches it that nearly everyone is his type, and the feed fills with people who do not fit him.

Your preferences are too tight. A narrow age range and a short distance leave the system a small pool, and it shows you the same people again and the same kind of people. Hinge says it nudges daters to be more open-minded about certain filters, like distance, for this reason.

You have not given it much to learn from. A new account, or an account used in short bursts, has little history behind it.

Hinge's own advice is short and practical.

Widen your preferences, and choose Show profiles outside of this preference. Hinge says this lets you see more people and lets more people see you.

Send likes to people you would really meet. Quality gives the system a clearer picture of you than quantity does.

Update your photos and your answers. A page that has been the same for months gives people no new reason to stop on it.

If you want more reach, Hinge says a Boost, or the HingeX feature Get seen by more daters, can show your page to more people.

My coaching on the apps agrees with Hinge on the main point. He tells men that these apps are businesses built to keep users happy and keep women on the platform, and that a man who likes everyone is read as having no standards. In his experience, men who stopped tapping on everyone and liked selectively got better results.

That is a view from coaching, and Hinge itself only says its system learns from your activity. The two line up well enough to act on: be selective, and let your history show a clear preference.

It also explains why a reset of habits works better than a rant. Spend a week liking only the people you would actually want to meet, with a short comment on each, and watch what the feed starts to show you.

Sometimes the feed is simply running out. In a smaller city, a narrow filter and a few weeks of use can exhaust the people who fit. That is not a malfunction. It is a small pool being fully shown to you, and the answer is a wider distance or a wider age range.

1. Are your filters tighter than the people you would honestly consider?
2. Over the last couple of weeks, did you like in bulk instead of selectively?
3. Is your page the same as it was months ago?
4. Have you used the app in short bursts instead of steadily?

If any answer is yes, the feed is doing what it was told. Change that input, give it a week or two, and judge it again.