This year’s Nobel Prize in Economics went to Alvin Roth of Harvard and Lloyd Shapley of UCLA and was awarded for “the theory of stable allocations and the practice of market design.” Their respective bodies of work focus on how to best match up pairs of individual people (or things) based on their preferences and desires.
What does this have to do with online dating, you ask?
What is matching and why does it matter?
Matching is the process of allocating resources by pairing together different factors, variables, or agents in the best or most efficient way. Matching can be used for lots of things: matching organ donations, doctors with hospitals, college admissions, marriage, even speed dating.
To determine how to match most efficiently, Shapley developed a series of algorithms – complex formulas – to sort and organize people’s preferences relative to others’ in order to create matches among large numbers of people, big data sets, etc. Roth then put them to the test in the real world. The Gale-Shapley deferred acceptance algorithm finds ‘stable matchings,’ which is what happens when all the matches in a big set of options are made such that there’s no other way everyone could be made better off.
Game theory says this is the optimal way to use dating sites
Such an algorithm is clearly useful in healthcare and education settings, but it can also apply to what some call the most important investment we make in our lifetimes: picking a life partner (or significant other). This year’s Nobel Prize winners focused mainly on marriage – but can their work apply to dating too?
In short: yes. In an over $4 billion industry such as online dating, getting your algorithm right matters. A lot.
An academic paper by several other economists, Matching and Sorting in Online Dating, explains why these algorithms can be so powerful and effective. The research applies the Gale-Schapley algorithm to a data set from an anonymous online dating service, to find that the website did a good job creating matches that are “approximately efficient.”
Case Study: OkCupid
OkCupid, a free dating site with 7 million users that was acquired last year by Match.com, claims that, “our matching system is the best in business, and you can easily train it to find the best matches for you.” Is this true? Do they and other online dating websites use algorithms as solid as this year’s Nobel Prize winners when they recommend good matches?
OkCupid lays out a rough approximation of its algorithm for public use. Keep in mind that no two dating service companies use the exact same algorithm – OkCupid currently even has a patent pending on its own. But some universal lessons can be drawn from the math of romance.
In a hypothetical case where 10 women and 10 men must be matched up based on their preferences, the Gale-Shapley algorithm teaches us this gem:
“The specific setup of the algorithm turned out to have important distributional consequences; it matters a great deal whether the right to propose is given to the women – as in our example – or to the men. If the women propose, the outcome is better for them than if the men propose, because some women wind up with men they like better, and no woman is worse off than if the men had been given the right to propose. Indeed, the resulting matching is better for the women than any other stable matching.”
This might be analogous to the advice OkCupid administrators give women new to OkCupid: the more messages you send, the more successful and satisfied you’ll be at finding appropriate matches. It sounds like being proactive is the way to go, whether you’re talking a proposal or just sending a first message.
Will online dating algorithms work for you? We think one thing is for certain: your free time is better spent on dates than on the Internet hunting down algorithms.