Zugot™What actually predicts that a relationship will work, and what only looks like it does. Articles on the research Zugot Project is built on — the formula that measures how much you need to match rather than only what you match on, the difference between dimensions where similarity helps and dimensions where it hurts, and what separates a shared hobby that predicts cohesion from one that predicts nothing.
The formulas themselves — what is measured, at what weight, and why that way.
Most matching systems compare positions: you're both a 4 on observance, so you match. The Lutz et al. formula adds a second question — how much do you care about matching on this at all — and that changes the answer entirely.
7 min read
The old argument about whether partners should be alike or opposites rests on a badly framed question. The right answer depends entirely on the dimension: on values and lifestyle, similarity wins; on innate traits, complementary balance does.
8 min read
"We have hobbies in common" sounds like a good sign, but the research shows it depends entirely on which kind. Activity that requires interaction predicts couple cohesion; watching the same series together predicts almost nothing.
7 min read
Age range and height range are the most common filters on dating sites, and the only ones almost nobody checks the output of. Sometimes the price is the best match that was available.
7 min read
Findings from the research literature on what holds a relationship together, and what doesn't.
Most compatibility dimensions ask how alike two people are. Two of them don't: attachment style and emotional stability are measured in absolute terms, because what decides isn't the gap between partners but where each of them stands on their own.
8 min read
Fear of rejection and a need for validation are the two sensitivities that end more matchmaking processes than any values mismatch. Neither is repaired by a similar partner — each defines a specific requirement of the other side.
7 min read
People describe the partner they've built in their head accurately, and the person they'd actually be happy beside for years much less so. How someone behaves in the friend group they already have is observed behaviour, and so it predicts better.
7 min read
How the virtual date is built, how the system learns, and how to fill in your card.
At the end of every date a question is asked about the reason. The answer isn't collected for statistics — it immediately changes the profile of the person who gave it, and only theirs. Four learning mechanisms, and one rule none of them crosses.
8 min read
The virtual date looks odd at first: two coloured circles moving through a gallery, able to hear each other only up close. Every one of those details exists to solve a specific problem the research identified in online dating.
8 min read
A dating site showing hundreds of profiles looks more generous than one showing a single match. The research on choice overload shows the opposite — more options produce fewer decisions, and less satisfaction with the decision that does get made.
7 min read
The closed fields on a card — age, city, level of observance — are easy to fill in and tell you little. The free-text answers carry most of the information, and they're also the part most people fill in worst.
7 min read
Registration is short, and the first date starts in the gallery — not in a list of profiles.