How dating app algorithms decide who you meet

Tinder scored your desirability. Hinge uses Nobel Prize math. Here's how dating app algorithms really decide who you meet — and why Texto doesn't use one at all.

In 2019, Tinder published a strange confession: "Elo is old news." For years, the app had been quietly scoring your desirability like a chess player. That score decided who you saw, and who saw you. The score is gone. The sorting is not. On every major dating app, an algorithm still decides who you meet before you ever get a say.

Here's how it actually works, and why we built Texto without one.

The score you never saw

In 2016, Tinder confirmed to Fast Company that every user had an internal "desirability score" borrowed from competitive chess rankings. When popular people swiped right on you, your score rose. When they swiped left, it fell. High scorers were shown to other high scorers: an invisible class system of attractiveness tiers.

Tinder officially retired Elo in March 2019, saying it now "continuously factors in how members are engaging." But the ranking never left. It just got smarter. Today, the single biggest factor in who appears in your stack is activity: recently online, frequently swiping. The algorithm re-sorts your queue after every single Like or Nope, with changes reflected within a day.

New accounts get a visibility "honeymoon", roughly 48 to 72 hours on Tinder and about a week with a "New Here" badge on Hinge, while the system calibrates your desirability. Then, for most people, visibility collapses. Swipe right on everyone and it gets worse: mass-swiping is read as spammy behavior and pushes your ranking down even further.

Hinge's Nobel Prize math

Hinge takes a different approach with "Most Compatible," one curated pick per day. Behind it sits the Gale-Shapley algorithm, the "stable matching" mathematics originally designed for pairing medical residents with hospitals and organ donors with recipients.

Instead of just guessing who you'll like, it models who is likely to like you back, and pairs people where interest is mutual. As Hinge's director of relationship science Logan Ury put it: "It's all about pairing people who are likely to mutually like one another."

The theory behind it won the 2012 Nobel Prize in Economics for Lloyd Shapley and Alvin Roth. (David Gale, who co-invented the algorithm in 1962, died before the prize was awarded.) It's beautiful math. It's still a machine deciding whose profile deserves to be seen: one match a day, served on the app's schedule, not yours.

The engagement loop

Here's the uncomfortable part. Dating apps don't sell matches. They sell attention. As one analysis put it bluntly: "The apps make money when you use them. They lose a customer when you find a relationship."

So the feeds are tuned like slot machines. Recency boosts keep you checking. The honeymoon keeps you hopeful. The post-honeymoon collapse keeps you swiping to recover what you lost. And when patience runs out, there's a paid exit: Tinder Boost promises up to 10× more views for 30 minutes, Hinge Roses plant you atop someone's screen, Bumble Spotlight pushes you ahead of the queue.

Read that again: you can literally pay to jump the line — because there is a line. Your visibility is a product, rationed by an algorithm and sold back to you by the minute. That's the loop.

Engagement loop diagram of dating appsYou swipeFew matchesYou pay for boostsMore swipingThe algorithm's goal:your time

The kicker: none of it predicts chemistry

All of this ranking, scoring, and matching rests on one assumption: that an algorithm can know who you'll click with. The research says otherwise.

In a 2017 study, psychologist Samantha Joel and colleagues analyzed more than 11,000 speed daters. Machine learning models trained on everyone's stated preferences could not meaningfully predict who would want to see each other again. Not a little bit wrong — useless. Chemistry happens in the room, in the conversation. It doesn't live in a dataset.

So the apps built an enormous machine to solve a problem science says machines can't solve. And they charge you for the privilege of being processed by it.

What if there was nothing between you and everyone else?

Texto takes the opposite bet. No Elo score. No desirability tiers. No honeymoon, no collapse, no queue.

  • No profile ranking. Nobody is scored, sorted, or hidden from you.
  • No boosts. You can't pay to be seen, because nobody can. Texto is 100% free.
  • No engagement algorithm. We don't optimize for time-on-app. We optimize for conversation.

Instead of a curated stack, you get real-time chat rooms. Everyone in the room is visible to everyone else, like walking into a café instead of standing in a sorted queue. You talk to whoever's actually there, right now, and chemistry gets to do what the research says only it can do.

And because no algorithm needs your behavioral history, we don't keep one. Conversations are end-to-end encrypted and ephemeral. They disappear when you're done, leaving no dossier of your romantic life to feed into a model.

The big apps need you single, swiping, and paying to skip a line they created. We'd rather you just meet someone.

No score. No queue. Just people. That's Texto.

Ready to try Texto?

100% free, no ads, no subscription. Private conversations that disappear when you're done.