Compatibility Test for Two

Because neither a birth date nor a name is linked to how relationships go: this has been checked on large samples, and no link was found. Other things are linked: plans that cannot be split in half, habits in arguments, how at ease a person feels with closeness. That is what the test asks about.

Your answers are written into the link itself, which you send to your partner yourself, in a messenger or by email. Your partner opens it, answers the same questions, and only then sees the shared result plus a link back for you. If the link is lost, the test has to be taken again: it is not saved to an account.

Yes, except the last question about your mood over the past two weeks: it is not put into the link, and only you see the note about it. Everything else is shown side by side in the result, on purpose: you can’t talk about differences if you can’t see what they are.

Yes, that is what the “Getting to know each other” version is for. It has no questions about a relationship that doesn’t exist yet; more weight goes to life plans and habits in arguments and closeness, which can be known early. To be honest: at this stage the test doesn’t predict a couple’s future, it shows what is worth talking about while it is still easy.

Studies of 23,000 couples in three countries (Dyrenforth and colleagues, 2010) found that personality similarity explains less than 0.5% of the differences in relationship satisfaction. Each person’s own traits matter more: someone whose emotions are steadier usually makes a relationship calmer for both, whoever they are with. So personality, arguments and closeness are scored by each person’s level, and matching counts only in plans and daily life.

After the questions about plans, each person can choose up to three topics where they won’t compromise. A mismatch on such a topic counts double, and a strong mismatch is shown at the top of the result and caps the overall score at 59 out of 100. Topics that can’t be split in half are treated the same way even if nobody marked them: children, marriage, pets when one of you has an allergy, alcohol and smoking for someone who can’t accept them.

It doesn’t predict that, and we don’t claim it does. The best models, built on 11,196 couples, explain up to 45% of the differences in how good people feel together now, and almost nothing about how that will change over time. The test relies on the same signals, but its number is a summary of your answers, not a probability of breaking up.

Yes. The questions don’t depend on gender or roles: even the question about a “traditional family” asks how earning and housework are divided, not who is the husband and who is the wife.

No. A low score means there are many differences, but it doesn’t say which of them you can work out. Start with the “What to talk about” section, which lists questions to begin with. If your arguments involve mockery or insults, or you feel afraid, this is no longer a question of compatibility: see a couples therapist, and if there are threats or violence, contact a support service in your country.