What a Deal Breaker Actually Removes, and What It Deliberately Doesn't
There is a control in your preferences called Deal Breakers. You pick things you do not want, and people who have them stop appearing. That is what it says, and as of recently it is also what it does. Both halves of that sentence are worth unpacking, because the second one is newer than the first, and because the interesting decision inside it is not about the people it removes. It is about the people it leaves alone.
Excluded, Not Deprioritised
The first thing to know is the strength of the setting. A lot of dating software treats a stated dislike as a nudge: the algorithm weights those profiles down and shows them to you later, or less often, or in worse positions. That is not what happens here.
A deal breaker is applied as an exclusion. If a candidate holds a tag you have listed, they are removed from the pool you are being drawn from — not sorted downward, not shown at position forty instead of four. Removed. The reasoning is in a comment beside the code, and it is one line: that is what deal breaker means.
Which makes it worth being sparing
Because it is a hard filter rather than a soft one, every deal breaker you add is a straight subtraction from the number of people you will ever see. That is the point of the control and it is also its cost. There is a cap on how many you can set, and most people will find the useful number is well under it. Two or three things you genuinely cannot live with will shape your experience far more than a long list assembled defensively.
The Decision Worth Reading: People Who Told You Nothing
Here is the question the feature has to answer and nobody thinks to ask. You have said you do not want a particular thing. Somebody's profile does not say whether they have it. They have not claimed it and they have not denied it. Do they get shown to you?
They do. The filter excludes people who hold the tag, and it deliberately keeps everyone who has simply not listed anything. The reasoning, again from the code: a candidate who has listed no tags cannot be shown to hold one, and dropping them would silently shrink the pool for anyone who set a single deal breaker.
Read that twice, because it is a real design choice with a real trade-off and it could honestly have gone the other way. Excluding the silent would make your filter stricter and your results smaller. Including them makes the filter mean what it says — it removes people who have told you something disqualifying, and stays out of the way for people who have told you nothing at all.
What this means in practice
A deal breaker filters claims, not people
It acts on information somebody has volunteered about themselves. Silence is not evidence of absence, and the filter refuses to treat it as evidence of presence either. So a deal breaker narrows who you see without hiding everyone who has left a field blank, which on any dating platform is a large share of everybody.
Why It Is Worth Saying That This Now Works
The Deal Breakers control existed in preferences for months, and it saved faithfully every time somebody used it. What it did not do was get read. The column was written on every save and consulted by nothing; a user's deal breakers were stored perfectly and applied to precisely nothing.
That is fixed, and we would rather tell you than let it pass. If you set deal breakers some time ago and concluded they were not doing much, you were right, and it was not your imagination or a subtle weighting you could not see. Now they are a hard exclusion. It is worth opening the control and looking at what you left there, because whatever it says is now genuinely being applied.
A control that saves and is never read
This is a specific kind of software failure and it is nastier than a button that visibly does nothing. The setting persisted, it showed your choices back to you when you returned, and every signal a user has access to said it was working. The only way to catch it from the outside is to notice results that do not match the setting, which is very hard to do when the thing being filtered is people you would never have seen anyway. We found it by reading the matching code rather than by anyone reporting it.
How to Set Them Well
- Start from what you have actually ended things over. Not what you would prefer to avoid — what has genuinely ended something before. That list is usually short and it is the honest one.
- Leave the merely undesirable to your own judgment. A hard filter is the wrong instrument for a mild preference, because it removes the chance to weigh anything in context.
- Revisit them. A deal breaker set in one phase of your life can quietly go on shrinking your pool long after it stopped being true.
- Remember the silent are still there. If somebody has left everything blank, no filter you set will speak for them — which is another argument for asking rather than assuming. Our guide to talking about something personal while you are still deciding is the other side of that, from the position of the person choosing what to volunteer.
What It Does Not Reach
A deal breaker acts on stated attributes. It does not read messages, it does not infer anything from behaviour, and it is not a safety tool — it cannot know how somebody will act, only what they have said about themselves. If your concern is conduct rather than compatibility, the controls for that are different ones and they are covered in setting boundaries early.
And it sits alongside the rest of your preferences rather than overriding them. Everything else you have set still applies; deal breakers subtract from the result rather than replacing it. If your matches feel thin, the answer is more often the accumulated weight of several settings than any one of them, which is a case for reviewing them together rather than tightening the newest.
The whole feature is about eight lines of code and one comment. The comment is the part we would defend: it says who gets removed, and it says who deliberately does not, and it says why. A filter that could not explain the second half would be a worse filter, however well it worked.