There is a specific kind of tired that only shows up around the fourth or fifth time a relationship ends the same way. Not heartbreak exactly. More like recognition. You have had this conversation before, with someone else's face attached to it.
Different city, different job, different name. Same ending.
The easy explanation is bad luck. The more useful one is that the pattern was never in the people. It was in the choosing.
Attraction does most of its work before conscious thought arrives. By the time you decide someone is interesting, a considerable amount of sorting has already happened: how they hold attention, how quickly they fill a silence, whether they lean in or hold back, how they handle being disagreed with.
None of that shows up on the list you would give a friend. Ask most people what they want and you get adjectives: kind, funny, ambitious, loyal. Those are real preferences. They are also almost useless as predictors, because nearly everyone wants them and nearly everyone believes they have them.
The actual filter is running on something narrower and much harder to say out loud. Often it is a temperament rather than a trait. Sometimes it is a familiar emotional weather pattern, one you learned to read early and can navigate without thinking.
This is the part that stings. The nervous system does not rate potential partners on whether they are good for you. It rates them on whether they are legible.
Someone whose moods you can predict feels safer than someone whose moods you cannot, even when the predictable moods are the difficult ones. If you grew up managing a volatile parent, calm can register as flat. If you grew up around emotional distance, someone fully available can feel like too much, too fast.
None of this is destiny and none of it is a diagnosis. It is a bias, and biases can be noticed. But you cannot notice one you have never articulated.
If you want to find your own filter, retrospective questions work better than aspirational ones. Aspirational questions ("what do I want?") return the adjective list. Retrospective questions return data.
Answer those honestly and the shape usually appears within a few minutes. It tends to be more specific and less flattering than the adjective list, which is exactly why it is more useful.
Not much, at first. Naming a pattern does not dissolve it, and trying to date the exact opposite of your type is just the same reflex running backwards.
What changes is the pause. When you can name the filter, you get a beat between the pull and the decision. In that beat you can ask a genuinely different question: is this person good for me, or just easy for me to read?
That question, asked a handful of times, does more than any list of desired traits ever will.