Real love demands real stakes

Half of the people I’ve talked to about my hopes to build a new dating site believe the entire premise is doomed. They say that dating is fundamentally a lemon market. In other words, “sellers” have more information about the true nature of their “product” than the buyers do. They can exploit this by misrepresenting the true nature of their goods. Someone might exaggerate their financial status, career achievements, physical appearance, or relationship history in their dating profiles (or in messages) in order to appear more attractive than they really are. This misrepresentation erodes trust in the whole market. How can you tell if someone is actually good for you, and thus worth your time?

Getting people to reveal themselves honestly is a hard problem. Some people propose building authentication systems. Have someone actually fulfill a background check, upload their diplomas, volunteer the phone numbers of employers in order to verify employment. Maybe this could work. It would require work to build it in a way that scales. And I suspect that mostly straight men would be the relevant subjects here.

But such authentication elides all of the deeper correlates of compatibility. Do you really care if someone graduated from Harvard or not if they don’t have compatible values in the areas that you actually care about?

Plus, there’s a much cheaper way to force people to reveal their true nature. Particularly in matters of values. Ask scissor questions.

Scott Alexander coined the term scissor statement, which Charles Arthur roughly describes as a statement that divides “and does so absolutely. You can’t be on the fence about it.”

These are statements that are divisive and polarizing. They reveal schism. Here are some examples:

• Trans women are women.

• Men cause 100% of unwanted pregnancies.

• Assimilation promotes unity and national identity.

• Guns keep us safe.

• Cancel culture holds people accountable.

You can probably immediately imagine the argument someone might make for why each of these statements is true or false yet also have a deep innate sense of your views on them.

What if such statements were blatantly turned into questions? It’s almost impossible to imagine someone answering such a question in a way that disagrees with their values. This is a big component of what Offline’s question set will be like.

It’s not that Offline wants to promote polarity in the culture, generally. It’s that Offline wants to help like-minded, value-aligned individuals meet. In order to that, we need to find where people actually start to disagree with one another. If you’re a bit disturbed by this premise, remember you can also rank the importance of any given question when looking for matches. (You can also optionally write a little blurb about why you made the choice you did or if there’s any additional context or nuance you want to offer!)

You see, solving for compatibility is about revealing where people are actually different from each other, and from the mean, rather than where they are the same. Most dating apps encourage too much sameness. When a pick up line is successful, it goes viral and you’ll see canned messages like “looking for the Pam to my Jim” on many profiles. People optimize for mass appeals because they want any matches, at all, since there is no way of truly filtering for the right people for you.

We’ll of course still include questions about other topics, like relationship and sexual preferences, lifestyle and habits. But we are hopeful that including a sizable portion of scissor questions, in general, will promote a culture of honesty across all questions. If you already have to make real choices, why not just be honest everywhere else, too? Especially if you are rewarded for it by discovering compatible people.

Comparing your answers to that of a prospective date also gives you something real to relate to and, in the cases where you disagree, to actually talk about.

In other words, find a real match. Have a real conversation. And hopefully find a real relationship.