There’s no perfect equation for getting laid in the Tinder age
Mathematical matchmaking is more than 50 years old. So why am I still single?
Last July, I joined Grindr and things started off strong. I had a few good screws, a handful of hot dates and an intense summer fling. I saw more action in six months than I’d seen in the past three years. I was winning and a computer was helping me do it.
Then, after an eight-day trip to Las Vegas this past January, all the action dried up.
I’d seen everything that Grindr had to offer and was growing weary of unsolicited dick pics and random old balls. So I turned to Tinder and had a couple failed dates. The first was a disgruntled state worker who wore Tom’s, winced when I told him I had two pitbulls and spent the better part of two hours mansplaining ethics to me. The next was a waifish first-year English teacher and self-professed INFJ who, curiously, didn’t do a lot of reading due to his workload.
I powered through, reminded of months of success, but couldn’t help thinking of my failure. Why had my dates gone so horribly wrong? How did we match in the first place? Why had the cupid in my computer betrayed me? Was it my picture? My profile? Was I too forward? Too passive? Did that dick pic not send?
I obsessively checked my phone, hoping to score and mulling over what I’d done wrong, but the hits weren’t coming like they used to. I swiped left for hours without a single match, and then it occurred to me: I’d been downgraded. Somewhere along the line I’d lost my mojo, and the app knew it. I was now a bottom-of-the-barrel bachelor.
If the computer couldn’t help me, I’d have to help myself. But was I really better suited to finding a mate than my smartphone?
Around the same time, one of my closest friends received an email saying he was “now among the most attractive people on OKCupid” and would subsequently “see more attractive people” in his results.
Here I was, getting hit with the ugly stick on Tinder while in another corner of the online dating universe one of my best friends had just joined an exclusive club of the internet’s most eligible bachelors. Something had gone horribly wrong. So I did what sore losers do and I quit. I deleted Grindr and Tinder and Scruff and swore off online dating altogether. If the computer couldn’t help me, I’d have to help myself. But was I really better suited to finding a mate than my smartphone?
Despite claims that location-based hookup apps like Grindr and Tinder have either disrupted or destroyed dating, computer-assisted matchmaking is nothing new. In his book A Million First Dates, Dan Slater traces the origins of online dating to the university labs of Stanford, Harvard and Iowa State University in the late ’50s to mid-’60s, where engineers used punch-cards to feed questionnaires into massive IBM computers in the hopes of finding like-minded suitors for willing singles. The projects had limited reach but planted the seed for the boom in online dating that started with sites like Match.com, eHarmony and OKCupid in the ’90s and early aughts.
These sites traded on their scientific approach to matchmaking. EHarmony claimed to have used science to “lower the divorce rate,” while OKCupid famously saved itself from extinction by publishing its findings on dating and big data. As the stigma surrounding computer-assisted coupling faded and smartphones went from luxury to first-world necessity, sites like Tinder and Grindr flourished. According to a recent Pew Research study, 15 percent of US adults have logged on to get off, implicitly trusting math to find a mate.
Some companies are more open than others about their secret matchmaking sauce. While the king of all hookup apps is notoriously tight-lipped about its techniques, a recent article in Fast Company revealed that Tinder sorts users with an internal desirability ranking. During an interview with the company’s CEO Sean Rad, Austin Carr was shown his “Elo score,” a nickname apparently cribbed from the chess world. It was the first public admission that such a ranking exists.
It’s clear that we want the algorithm to work and apps like Scruff and Tinder trade on that desire, but evidence to support their efficacy is largely anecdotal.
Unfortunately, that admission is about as much as we know today. There are countless ways in which Tinder could parse our data. It could cull information from our Facebook profiles, Instagram feeds and, of course, our behavior on the app. In November of last year, the company touted big changes to its matching algorithm that would lead “to a significant increase in matches,” but when pressed for details in an interview with TechCrunch, Rad referenced Google’s secrecy over its search algorithm.
Other dating services aren’t nearly as quiet about what makes their matches tick. OKCupid, which built its reputation as a leader in online dating off exhaustive data analysis, has been transparent about its ranking of users based on their supposed attractiveness. This hot-or-not method of pairing perspective dates seems the perfect match for a service as superficial as Tinder, where prospective lovers are presented like trading cards, but it’s certainly not the only way computers are helping us get laid today.
Scruff, a gay hookup app, uses a series of methods and algorithms to suit different user behaviors. The app ranks its users based on how many times other users have “woofed” (the equivalent of a like or fav) a given profile and presents those in a “most-woofed grid.” It also presents an alternative grid of users based solely on proximity.
But it’s the app’s Match Stack function, similar to Tinder’s swiping interface, where algorithms are hardest at work. (Full disclosure: I’ve been on Scruff off and on for the better part of a year, but have yet to make a connection that materialized in a real-world encounter. This could be due in part to the community’s overwhelmingly hirsute focus and my lack of body hair.)
Scruff co-founder Eric Silverberg described the Match Stack as a combination of geo-location and Netflix-style collaborative filtering.
“The simple way to explain it is, if I like Daniel and Daniel likes Chris, it’s going to show me Chris, because, presumably, if we have similar taste in one thing, then our tastes will overlap, potentially in others. ” Silverberg said.
It doesn’t take a team of psychologists to prove that computer-assisted matchmaking, despite decades of work, isn’t a perfect science.
That seemingly simple process is made possible with what Silverberg describes as a “CPU-intensive machine in the Amazon cloud” crunching hundreds of gigabytes of data, including billions of user ratings in order to provide a “stack” of men tailored to fit each user’s explicit and implicit tastes. But, he points out that “those machine recommendations” aren’t everything. The app also peppers in a random assortment of guys in your area to build a more “diverse stack.”
It’s clear that we want the algorithm to work and apps like Scruff and Tinder trade on that desire, but evidence to support their efficacy is largely anecdotal. Yes, there are countless computer-assisted dating success stories, but how much of that success is based on access and volume and how much of it can actually be attributed to fine-tuned mathematical equations?
According to an oft-cited paper published in Psychological Science and the Public Interest, a research team led by Northwestern University professor of social psychology Eli Finkel found that there’s no evidence to prove that algorithms are better than humans at predicting compatibility. The paper’s summary puts it this way:
“Part of the problem is that matching sites build their mathematical algorithms around principles —typically similarity but also complementarity — that are much less important to relationship well-being than has long been assumed. In addition, these sites are in a poor position to know how the two partners will grow and mature over time, what life circumstances they will confront and coping responses they will exhibit in the future and how the dynamics of their interaction will ultimately promote or undermine romantic attraction and long-term relationship well-being.”
And then there’s the ever-important question of chemistry. As a good friend posed it: “How does a computer know who your body wants to fuck?” It’s a very real question. Compatibility goes beyond preferences, appearances and relative attraction. When we meet someone in person, there’s a whole host of biological signals at play that a computer just can’t re-create. It doesn’t take a team of psychologists to prove that computer-assisted matchmaking, despite decades of work, isn’t a perfect science.
That said, in my month living off the online dating grid, I didn’t get laid once. I went on zero dates and aside from a couple of drunken winks across the bar, my flirtations were fleeting. I eventually re-downloaded Tinder, Grindr and Scruff and within days I was back in action. Twenty-four hours in, I was chatting with a handful of men, and making plans for offline encounters.
A few weeks back online and at least one one-night stand later, I got a message from a man I likely never would have met in the real world. He lives three cities over and works nights in law enforcement. With my travel and work schedules being what they are, the chances of us physically being in the same place at the same time are slim.
We’ve been on four dates and my faith in the love algorithm has been partially, if not cautiously, restored. There were, no doubt, multiple equations at work in connecting us in the first place, but something much bigger got us into bed. The truth is, math can only take you so far. The rest is chemistry.