"Face card" means roughly this: your face as the thing you present, the way a card is what you hand over at a till. Saying someone's "face card never declines" means their face holds up in any photo, any lighting, any day. Saying a face card "declined" means the opposite, usually about one bad picture rather than about a person.
It is a good piece of slang because it captures something real, which is that appearance is inconsistent. The same face photographs differently across a week. What the metaphor gets wrong is the shape of the thing it describes. A card either goes through or it does not. Faces do not sort that way, and the data says so fairly bluntly.
The metaphor is binary, the distribution is not
Across 17,032 analyses run between January and August 2026, the mean score was 5.59 and the median was 5.3. Half the dataset sits between 5.0 and 5.9. That is a very narrow strip of a ten-point scale holding roughly one person in two.
| Score band | People | Share |
|---|---|---|
| Below 4.0 | 77 | 0.5% |
| 4.0 to 4.9 | 4,230 | 24.8% |
| 5.0 to 5.9 | 8,450 | 49.6% |
| 6.0 to 6.9 | 2,891 | 17.0% |
| 7.0 to 7.9 | 824 | 4.8% |
| 8.0 to 8.9 | 524 | 3.1% |
| 9.0 and above | 36 | 0.2% |
Nearly three quarters of the dataset, 74.7%, scores 5 or above. A quarter, 25.1%, reaches 6. Only 8.1% reach 7 and 3.3% reach 8. Just 36 people out of 17,032 scored 9 or higher.
So if "never declines" means what people usually mean by it, something like unambiguously striking in every photograph, it describes a top-3% group at most. And if "declined" is supposed to mean the bottom, that group barely exists either: 0.5% of people, 77 out of 17,032, scored below 4.0. The metaphor has two large categories and the data has one large category with thin tails on both sides.
What the crowding actually means
The 5.0 to 5.9 band holds 8,450 people. Inside that band, the difference between the person at 5.1 and the person at 5.8 is seven tenths of a point, and thousands of people are stacked between them. Move up one band and 6.0 to 6.9 holds 2,891 people in another single point of range.
The practical consequence is that "face card" talk is describing tenths of a point as if they were categories. Two people who read as noticeably different in a group photo are frequently separated by less than half a point of measured geometry, and that half point can be produced by lighting, sleep, camera height or expression rather than by structure.
This is also why the metaphor holds up better as a description of days than of people. A face card that "declines" on a Tuesday and not on a Thursday is not a face changing by a full band. It is the same face measured under different conditions, which is exactly what the slang was reaching for before it got turned into a ranking.
Where the number comes from
A face card rating is not a vibe check with a number attached. It scores twelve features separately - eyes, eyebrows, nose, lips, chin, cheeks, forehead, jawline, skin, bone structure, symmetry and overall facial harmony - and combines them.
Symmetry does most of the work. It correlates with the final score at 0.95, and the gap between groups is large: people scoring 8 or above overall average 8.54 on symmetry, while people under 5 average 4.74. Golden-ratio compliance, which gets far more attention in looksmaxxing content, averages 77.28% across the dataset and correlates with the overall score at only 0.67.
That matters for reading your own result. A mid-range score is usually mid-range across most of the twelve, not one catastrophic feature dragging down eleven good ones. The feature breakdown is where the useful information is, and the headline number is a summary of it.
How to read your own face card score
Against the table, not against 10. A 5.3 is the median. A 5.9 is above half the dataset. A 6.5 is comfortably in the top quarter. A 7 is roughly 1 person in 12, and an 8 is 1 in 30.
Two limits are worth stating. First, this is one model scoring geometry and presentation in a single photograph. It has nothing to say about expression in motion, voice, style or how someone comes across in a room, and those carry enormous weight in actual attraction. Second, the people who run a face rating tool are not a random sample of the population, so "top 3%" describes the people who showed up, not humanity.
And the humane reading of the table is the accurate one, not a softened one. Half of everyone lands between 5.0 and 5.9. A score in the fives is not a declined face card. It is the most common result in the dataset by a wide margin, which is another way of saying it is what normal looks like when it is measured. If you want the same number reported with your percentile against this distribution, the attractiveness test does that, and the full band-by-band figures are in the average face rating study.