How a chess formula became a dating-app score
Elo was designed to estimate the relative strength of chess players from the outcomes of their games. Before a match, each player has a rating; afterward, the ratings shift according to the result and how surprising it was. Beating a player expected to win earns more than beating a lower-rated one.
Tinder adapted that logic to swipes. Instead of chess victories, the system could treat users’ choices as signals about whom they preferred. A person’s score was therefore shaped not just by how many people swiped right, but by who those people were rated to be. The score was a prediction built from other users’ behavior—not an objective measurement of anyone’s appearance.
Why the score could shape what you saw
A rating system can influence the very outcomes it measures. If the app used scores to help decide which profiles to show, a high rating could bring a user in front of more people. Their responses would then feed into later calculations. Conversely, rejection by highly rated users could count differently from rejection by users with lower ratings, potentially affecting a profile’s future reach.
That creates a feedback loop: past reactions help determine who gets seen next, and those new reactions become more evidence for the system. A private number could thus matter even though users never saw it—and even if it was never intended as a final judgment of their worth.
Tinder later said it had moved away from relying on Elo. The lasting lesson is how easily a ranking tool can turn ordinary choices into an invisible gatekeeper: what looks like a stream of profiles may already reflect a chain of earlier comparisons.
