Compatibility
Preference fulfilment is 45% of the score, and never the whole answer
We weight what each of you actually asked for more heavily than any other platform, and we are equally clear about where it stops. A plain reading of the matching model.
James Holloway
Compatibility Research
2 min read
Every introduction we deliver carries a number, and that number is assembled from four components. Preference fulfilment is the largest single one at 45 points out of 100. It is also, by design, incapable of carrying a match on its own.
How the forty-five points are earned
When you built your profile you told us what you are looking for: an age range, the height and body type you prefer, education, occupation, religion and its place in your life, languages, location and distance, whether children are expected. Those answers are not a wish list we file away. They are the largest single input into whether anybody is shown to you at all.
The forty-five points are scored on two things, and both of them run in both directions.
- Reciprocity. It is not enough that they fall inside what you asked for. You must fall inside what they asked for too. A pairing that only works one way scores poorly, because in practice it only disappoints one of you slightly later than the other.
- Distance from the centre. Falling just inside a stated range earns fewer points than sitting comfortably within it. Someone at the very edge of your age band is a match, but an honest score should say that it is a near miss rather than a bullseye.
The scoring is fixed and deterministic. It does not learn from your behaviour, it is not tuned to keep you scrolling, and it produces the same answer for the same two profiles every time you run it.
The other fifty-five
- Values and life goals — 25 points. Marriage timeline, children, how much a shared religion matters, and your five ranked values.
- Lifestyle — 20 points. Habits, work, education, languages, family shape.
- Personality and interests — 10 points. Proximity across four bipolar scales, and what you share.
Why weight it so heavily
Two honest reasons. The first is that our members are looking for marriage, not novelty, and the constraints they name are usually real ones — where they can live, what a family expects, how faith is practised on an ordinary weekday. Quietly overriding those in the name of chemistry wastes everybody’s time.
The second is that it is a framework you can see and argue with. You wrote the inputs. You can change them. You can read the breakdown afterwards and tell us it was wrong. That is more than can be said for most black-box relevance models.
What we will not do is claim it predicts a marriage. A 92 is a strong opening, not a verdict. The conversation is still yours to have.
James Holloway
Compatibility Research
Part of the team that builds, reviews and runs RealVows. We publish how the service works because a matchmaking method nobody can inspect is just a black box with better manners.