A New Formula Could Turn Online Dating On Its Head

Traditionally, online dating websites match users based on things like shared preferences, interests, and goals. There's no doubt that all those things are important elements of a successful relationship, but there's one hugely important factor that dating algorithms don't take into account: physical attraction.

Sure, online daters can list vague specifications about things like height and body type preferences, and of course most of us are guilty of spending more time on photos than profile essays, but matchmaking formulas give very little weight to attraction. Until now, that is. A new potential matchmaking formula created by computer scientists at the University of Iowa puts mutual physical attraction front and center by only matching you with partners who are likely to find you attractive.

The formula works by analyzing your message history in order to make date recommendations based on shared tastes and the kinds of people you found attractive in the past. The scientists who designed the algorithm described it using a fictional online dater named Mike: “It is Mike’s taste that affects whom he approaches through initial contacts, and his attractiveness that determines whether he can get responses.”

In other words, Mike – or any of us – can initiate conversations with anyone he wants, but he won't get very far if he continually reaches out to people who don't find him attractive. If the people you contact never reply, all the recommendations in the world will never be of any use. The most effective online dating algorithm would be able to suggest matches based on the likelihood that they will respond to your messages.

“Considering the match of both taste and attractiveness, our model tries to improve dating partner recommendations by boosting a user’s chance of getting responses,” the researchers write.

To test their theory, they used anonymized data on 47,000 users from a real dating website over the course of a period of nearly 200 days. Their program analyzed the replies each user received and used them to evaluate the user's attractiveness or unattractiveness based on the assumption that people who receive more replies are more attractive. When compared with other methods of matchmaking – like pairing people based on shared interests and other variables they had in common – the results showed that the beauty-based method was significantly more effective.

“If a user approaches a partner recommended by [our engine], he/she will have a better chance of getting responses,” the researchers say. So far the attractiveness sorting formula is just a concept, but it's probably only a matter of time before a dating service seeking to be the next best thing turns it into a reality.