Thanks for playing

What should I do differently in a dating app if I want to increase my match rate?

Delete and recreate your dating app account with a brand new Facebook account or phone number. This will reset the algorithm's history with you. Unlike a "boost" or other paid dating app services, this also resets reciprocal history: what users think about you.

How does the algorithm work?

A real algorithm was used in this game to make recommendations for you. It made those recommendations based on your swipes. Learn more about that algorithm here.

What is the state of online dating?

There's lots of evidence that people want racial, ethnic and religious diversity in their romantic lives. Dating apps lag this trend, seemingly because there may be something wrong with the algorithm that powers them. Read more here.

So is this real?

MonsterMatch simulates online dating using the design of the dating app Tinder and a common collaborative filtering algorithm. Your swipes inside the game have a similar impact on what profiles you see as those swipes would on a platform like Tinder.

Is there a diversity problem in dating apps?

Yes, by many measures. That's why there are so many segregated dating apps, like JDate for Jewish users and Amo Latino for Latino users. Something about how the apps are designed prevent them from working for a diverse group of people. The leading hypothesis is that the user experience of swiping on recommended profiles plays the biggest role.

Don't people just like what they like?

If something's not on the menu, you'll never have a chance to try it.

But isn't it the user's fault?

I personally don't think bad match rates are the user's fault.

A lot of people blame the user, dating apps most of all. They release statistics on how their less attractive users receive fewer matches. They never release statistics on the ratio of men to women in their apps. The cherry-picked data in these blogs is meant to lure you into downloading. Apps will only show bad-sounding data if it's consistent with how they think the average person sees the world.

What do dating apps think about MonsterMatch?

I don't know. They will probably show you a survey that says people love their apps, and then add a reset button for your "data". Without being able to see the code, we'll never know what the reset button actually does, especially because it's not clear what to do with data from other people about you. Also, their surveys are bogus, because if you don't like dating apps, you just stop using them and you're never around to answer their surveys!

So what should we do to fix this problem generally?

Dating app makers should open-source the algorithm and data gathering that's used by their apps, and make their datasets available for a trusted third party to analyze. It's unlikely they would do either of these things. Read more here.

What happens when someone has a bad experience on a dating app?

The evidence suggests they use a different app but keep going at online dating. Most online dating users have accounts on multiple apps. So there doesn't seem to be an alternative: dating apps are sort of the place of last resort for relationships.

Why do people use online dating?

The prevailing survey response is people use online dating to "have fun." The second reason is meeting people. Read more here.

Why is a game developer qualified to talk about online dating?

Because if people use online dating to have fun and describe it as a game, there's definitely some real video game psychology going on in how people swipe and why.

Is there a trend that is making people use online dating more?

I agree with the consensus opinion that differing levels of educational attainment between men and women has the biggest impact on the experience of dating in the United States.

Since single women are significantly more likely to have college degrees than single men in their age group, it's less likely they find partners at work, school or daily life. The people you see in those parts of your life depend highly on your educational attainment. Online dating helps you meet people outside of daily life.

What is the specific collaborative filtering algorithm?

The game uses non-negative matrix factorization to figure out how your swipes are related to 5,500 other data points in the dataset. Then, the recommender shows you the profiles with the best predicted ratings first. Additionally, your most recent swipes have the largest impact on your recommendation.