Short version: character over qualities, shared direction over shared interests, and a much shorter list than you currently have.
What holds up over decades
Kindness, particularly when unobserved. How they treat a waiter, a parent, someone who can do nothing for them. This is the closest thing to a general-purpose predictor, because most of marriage happens when nobody is watching.
Reliability. Do they do what they said? Over twenty years, this is what safety actually consists of. It is unglamorous and it is the foundation everything else sits on.
The ability to repair. Can they apologise without attaching a defence to it? Can they come back after a difficult conversation? Conflict is guaranteed; repair is the variable.
Emotional steadiness. Not the absence of feeling — the ability to have a hard feeling without the household reorganising around it.
Honesty, including when costly. Notice whether they tell small convenient untruths. People who do it about small things do it about large ones.
Willingness to be wrong. Someone who cannot revise a position is someone you will be managing for a very long time.
How they treat their family. Not whether the relationships are good — families are complicated — but whether they are honest about them and behave decently within them.
What people over-weight
Career and earnings. Stability matters. Prestige does not, and a demanding career has a real cost in presence that people discount at the start.
Height, and its equivalents. The most commonly named trait that people later report mattered least. In the Dating Priorities tool it is the single most frequent answer to “what did you overvalue?”
Shared hobbies. Pleasant. Nearly irrelevant.
Impressiveness. Whether they are impressive to other people is a different question from whether they are good to live with, and the two are only loosely related.
How closely their observance matches yours today. What matters is whether you want the same home in ten years, not whether your current practice is identical. This is covered properly in the article on different levels of observance.
The things that genuinely are non-negotiable
A small set, and they tend to be structural rather than personal:
- Whether they want children, on a timeline you can live with
- Whether the life you each want can be the same life — geography, work, family proximity
- What the home is Jewishly, and whether you both actually want that home
- Whether you are treated well — which is not negotiable at any price
Almost everything else is a preference with a price attached, and the useful work is knowing what price you would pay.
Why your list is too long
Most people hold twenty or more criteria, which is functionally the same as holding none — nobody clears twenty, so in practice you end up deciding on whatever is most salient in the moment. Usually attraction.
Cutting the list is not lowering standards. It is the opposite: it is deciding in advance what you will actually hold the line on, so that you hold it.
The Dating Priorities tool does this by force. Twelve, then seven, then three. Most people find the cut from seven to three genuinely uncomfortable, which is the point — it is the first time they have been made to price the trade-off.
One test worth applying
Imagine you are ill for a long stretch, or you lose your job, or a parent dies. Who do you want in the room?
It sorts a list quickly. Almost nothing on the impressive column survives it, and almost everything in the character column does.