The 15 Face Scan Tests, Explained: What Each Score Means
All fifteen tests on this site, in plain language: what you upload, what comes back, and what the number is really based on. No mystique required.
What this page is
Every test on this site runs on the same pipeline: detect a face, place landmarks, run a formula. What differs between tests is mostly the formula — which measurements it reads and how it weights them.
This page lists all fifteen tests and, for each one, what it actually reads. The details come from the code that runs the scans, not from the marketing copy. Nothing here is a secret; it is just rarely spelled out.
Two general points before the list, because they apply to every row:
Every score comes from geometry. The inputs are distances and ratios: the width of your mouth against the width of your face, the height of your forehead, the angle of your jaw. Coordinates and blendshapes. No test consults anything about your life.
The same few measurements keep reappearing. Nose width, jaw width, face width-to-height ratio, lip thickness, and a handful of expression blendshapes. Notice how often they recur. Fifteen tests, but a much smaller set of inputs — which tells you the tests are variations on a theme rather than fifteen independent instruments.
The fifteen tests
| Test | Score range | What it actually reads |
|---|---|---|
| Attractiveness ✨ | 0–100 | Facial symmetry, eye size relative to face width, the proportions of the facial thirds, smile intensity, jawline width |
| Rich Destiny 💰 | 0–100 | Nose width against face width, forehead height measured from the top of the bounding box, lower-lip thickness |
| Body Count 💔 | 5–50 | Mouth width divided by face width, jaw openness, and two smile blendshapes |
| Cheating Tendency 😠 | 0–100 | Asymmetry between the eyes, asymmetry along the jaw, and how tightly the mouth is pressed closed |
| Libido Index 🔥 | 0–100 | Distance from nose to chin, jaw openness, lower-lip thickness |
| Real Size 🍑 | 0–10 | Ratio of lower-face width to upper-face width, jaw openness, mouth width |
| Love Brain 💘 | 0–100 | Eye roundness, lip fullness, resting smile, distance between brow and eye |
| Virgin or Not? 🙈 | 0–100 | A blend of eye-region measure, mouth tightness, mouth puckering, overall symmetry, jaw tightness, and smile restraint |
| Boy or Girl 👶 | 0–100 | Face width-to-height ratio, jaw width ratio, forehead fullness, nose-bridge height |
| Celebrity Match 🌟 | 0–100 | Eye spacing ratio, nose width ratio, jaw width ratio |
| Height Predict 📏 | 0–100 | Face width-to-height ratio, nose-bridge height, jaw width, lower-face length |
| Future Looks ⏳ | 0–100 | Symmetry, jawline score, facial thirds, a youthfulness estimate |
| Peach Blossom Luck 🌸 | 0–100 | Smile blendshapes, smile-eye curve, lip fullness, and cheekbone width against face width |
| Night Owl Index 🌙 | 0–100 | Eyelid openness, under-eye area, and how far the mouth corners turn down |
| Noble Luck 👑 | 0–100 | Forehead height ratio, brow gap, and nose-bridge length |
Reading the table honestly
A few things fall out of that list once you look at it as a whole.
The score ranges differ, and none of them are calibrated. Some tests return 0–100, one returns 0–10, one returns 5–50. Those endpoints were chosen by whoever wrote the formula. A 70 on one test and a 7 on another are not saying anything comparable — they are different scales with different anchors.
The inputs are small in number and reused constantly. Look at how often nose width, jaw width, and lip thickness appear. It is entirely possible for two tests with very different names to be reading nearly the same measurements with different weights and different labels. That is not a bug; it is what a small formula library looks like.
The labels are the authored part. The measurement stage is real. The step where "jaw width ratio" becomes "Cheating Tendency" is a naming decision, made once, in code. That single step is where the entertainment value and all of the plausibility live.
Some of the result text contains fixed numbers. The messages attached to score bands include statistics — a percentage of people with your face type, a size for a reference database. Those numbers are literal strings written into the code. They are not computed from your photo and not drawn from any dataset. Treat any percentage you see in a result message as part of the script. A share of users, a database size, a confidence figure — these are written by hand, not computed from your photo, and a number you cannot check is worth reading as advertising rather than as a measurement.
Why this matters for how you use them
Knowing the mechanics does not ruin the tests. It changes what you get from them.
- Comparing two tests is not comparing two things. If Body Count and Real Size both partly read mouth width, agreement between them is not corroboration.
- A high score is not a high score across the board. Different ceilings mean the numbers are not interchangeable, so there is no such thing as a "good" total.
- Consistency is guaranteed by the formula, not by accuracy. Same pose, same lighting, same inputs, same output — every time.
- The interesting part is the label, not the number. Seeing which human trait a jaw measurement has been assigned to is the actual content of the experience, and it is the part written by a person.
The honest summary
Fifteen tests, one pipeline, a small shared library of facial measurements, and fifteen authored stories about what those measurements supposedly mean. The computer vision does its job well. The interpretation is a design choice dressed as a discovery.
That is not a reason to avoid the tests. It is a reason to enjoy them for what they are: a fast way to get a group of friends arguing about their faces. Take the number as a prompt, keep the label at arm's length, and the whole thing stays fun.
Try the free face scan tests
15 entertainment-only tests. Upload one selfie, get a score, take it exactly as seriously as it deserves.
Open the tests