Am I pretty or ugly? The question has no answer, and here is why.

When researchers collect attractiveness ratings from panels of people, the results form a single smooth distribution. Most faces cluster in the middle, with fewer at each extreme, and there is no gap anywhere - no point where one group ends and another begins. It looks like the distribution of height, or of shoe size.

A binary needs a boundary. There is nothing in the measurements to put one on, so every tool that returns a pretty-or-ugly verdict has chosen a cutoff and not told you where. Move that cutoff by two points and large numbers of people change category without a single feature changing.

A landmark-based tool can genuinely tell you several things, and this site reports all of them:

• How closely your paired features mirror each other across the centre line of your face, as a percentage deviation.
• How evenly your face divides into vertical thirds and horizontal fifths.
• How your height-to-width ratio compares to the classical 1.618 target.
• Which of six face shapes your proportions sit closest to, and by how much.

Those are real facts about the geometry in one photograph. They are the same kind of fact as the distance between your pupils. None of them is a category you belong to.

The single most reliable finding in this field is that composite faces - many photographs digitally averaged together - are rated as more attractive than nearly all of the individual faces used to build them. Averaging removes asymmetry and pulls every proportion toward the population mean, and the result is rated pleasant.

It is also rated forgettable. Composites do not become anyone's favourite face, and the features that make a particular face compelling to a particular person - a strong nose, a wide mouth, an unusual eye shape, a gap in the teeth - are the exact deviations that averaging removes and that a proportion test scores down.

So a geometric score is measuring closeness to an average. Reaction is measuring something else entirely, and the two only loosely overlap. A tool that collapses them into a verdict is not making a mistake in the arithmetic. It is making a mistake about what the arithmetic is for.

It happens often enough to be worth writing down. Some things that are true and easy to lose sight of:

• Rating tools are usually built to produce shareable results, not accurate ones. A number tuned for engagement is not evidence.
• The same face measures several points apart across two photographs taken minutes apart. If the result moves that much, it was never measuring you.
• Preoccupation with a perceived flaw in your appearance, to the point where it affects your day, is a recognised and treatable thing. Speaking to a GP or a therapist is a reasonable response to it, and taking another test is not.

Am I pretty or ugly - is there a test that can tell me?

No. Tests can measure facial geometry, and this site runs one of them. What no test can do is sort faces into pretty and ugly, because that boundary does not exist in the data. Attractiveness ratings from real people form a smooth, wide spread with no gap in the middle - there is no point on the scale where one category stops and another begins. Any tool that returns a pretty-or-ugly verdict has invented the line it is drawing.

Why does this site refuse to give a verdict?

Because it would be a lie with consequences. The tool measures symmetry and proportion in one photograph. Turning that into 'ugly' would require knowledge it does not have, and would hand a stranger's arithmetic authority over how someone feels about their own face. The score is reported as a measurement, with bands that describe geometry rather than judging it.

What do am I pretty or ugly tests actually do?

Most detect facial landmarks, compute a few ratios, compare them against target values, and then map the result onto a verdict with a threshold the developer chose. The measurement part is real. The threshold is arbitrary, undisclosed, and usually set to produce a spread of results that keeps people sharing screenshots.

Does facial symmetry decide whether someone is attractive?

It contributes and it does not decide. Across large samples there is a real but modest relationship between measured symmetry and average attractiveness ratings. On any individual face it predicts very little. Almost every human face is asymmetric to some degree, including every face you have ever found beautiful.

I already think I am ugly. Will this test help?

Probably not, and it is worth being direct about that. If you have already reached that conclusion, a number will either seem to confirm it or seem wrong, and neither outcome changes anything. Persistent distress about your appearance is a well-documented thing that responds to talking to someone, and does not respond to more measurements. The most useful thing this page can do is say so plainly.

Is it bad to search this?

It is extremely common, which is not the same as harmless. A great many people type this question, most of them late at night, most of them after looking at photos of themselves. The searching is not the problem. The tools that answer it with a confident verdict are.

• Am I pretty? The photo test
• The 1-10 pretty scale, and what it measures
• Pretty privilege: what the research actually found