Attractiveness scale 1–10: what each number actually means

The ten-point scale is the one everyone carries around, the PSL scale is the one the forums built, and BecomeTen's 0 to 100 is the one the rating uses. They look like three systems. They are one shape drawn at three resolutions: a middle where most faces sit, a lower range that is short in practice because few adult faces are rated there in good faith, and an upper range that thins out fast. Any honest scale has to look like that because that is how a population of faces is distributed. So the question "what is a 7 out of 10?" has a real answer, and it is not the answer the internet gives. Read across the table and 7 lands in the Above average tier, at about PSL 5 to 6, on a scale where most people are a 5. A 7 given by a stranger online usually means a 5 said kindly.

Two things to notice. The bands are not equal: a single point of the ten-point scale covers six to fifteen points of the 0 to 100, and the Normie tier absorbs both 5 and 6. And the top compresses: 8, 9 and 10 together cover less of the population than 5 alone. The tier names are BecomeTen's own, chosen to be neutral, and the looksmaxxing tiers entry explains each; the PSL column follows the mapping on the PSL score page, where 50 is about PSL 4, 62 about 5 and 75 about 6.

• A 7 out of 10 is clearly above average: 65 to 77 on BecomeTen, the Above average tier, about PSL 5 to 6. On an honest scale it is uncommon. The 7/10 face card the internet hands out is usually a 5 or a 6 said politely.
• A 6.5 out of 10 sits on the line between the top of Normie and Above average, about 63 to 66 on BecomeTen and about PSL 5. It is a good-looking, ordinary face with one or two areas holding it back, and those are most often skin, hair or body composition rather than bone, which is why 6.5 is the number where the gap to potential is usually widest.
• A 6 out of 10 is the upper end of average: 58 to 64, still Normie, PSL 4.5 to 5. Nothing is wrong; the face is a little above the crowd.

Half points exist because a single point of the ten-point scale is wide: the gap between a low 6 and a high 6 is as big as the gap between a high 6 and a low 7. The 0 to 100 score is there to say which side of the line a face is on.

Because a scale on which the average person is a 7 has three points left for everyone above average and six points for everyone below, and it stops telling you anything. The midpoint has to be the middle, and on BecomeTen that is deliberate: an average, healthy, groomed adult lands at 50 to 60, inside Normie. The reasoning, and what it does to the rest of the ladder, is in why average is fifty. Average is also not a failing grade in the research. A 1994 study in the Journal of Comparative Psychology found that faces closer to the population's average configuration are rated more attractive, independently of symmetry, which also correlates. A 2011 review in Philosophical Transactions of the Royal Society B summarises the evidence for averageness, symmetry, skin condition and adiposity as the signals with the strongest support. Ordinary is a well-studied, well-liked configuration, not the bottom of a scale.

Ask a stranger to rate a face and you get a social act, not a measurement. Nobody wants to tell someone they are a 4; 6 and 7 are the numbers that feel honest without feeling cruel, so the whole population migrates upward and the scale loses its middle. Apps have the same incentive with a commercial reason on top: a flattering number is shared, a strict one is not, so a rater drifts toward 7 and 8 for ordinary faces. The forums went the other way and built a scale strict to the point of cruelty, with a ladder of labels for the lower rungs. The strictness is useful, the labels are not, and BecomeTen keeps the first and refuses the second: the PSL number is shown as an equivalent, the labels are described on the PSL scale entry and repeated nowhere. Why two raters give the same face different PSL numbers, and what the photo adds, is in why PSL ratings vary.

1. Find the row, not the point. A 58 and a 62 are the same face on a different day; both are Normie, upper end, about a 6. Read the band, and treat a difference within two points between two scans as noise.
2. Subtract the inflation before comparing. If a friend or an app gave you a 7 and BecomeTen gives you a 56, they are probably describing the same face. The 7 was kind; the 56 is calibrated.
3. Check the capture before the face. A phone held close at chest height, overhead light or a smile each move the number by a few points. Every scan is graded for light, angle and sharpness, and a poor capture is flagged, so the score is read with that caveat.
4. Ask where the points went. The tier is a label; the area split is information. The gap between score and potential is the part of the number you can do something about, and for most faces it sits in skin, hair or body composition.
5. Keep the scale fixed. The tiers do not move with other users and there is no leaderboard, so a 65 next year means what it means today. That is what makes a second scan worth taking.

The score is an estimate of how one photo reads against a fixed rubric, not a measurement of a face and not a verdict on a person. The attractiveness test page sets out what a single photo can and cannot support, and the AI face rating is where the number comes from.