Looksmaxxing Statistics
What 387,170 AI face scans reveal about real faces: how rare Chads actually are, the average face rating, the weakest male facial feature, and the selfie-vs-side-profile gap. Primary data, published nowhere else.
Where Faces Actually Land
Overall attractiveness scores from a 30,000-scan random sample of full app analyses. The curve is brutally narrow: one in three scans lands between 65 and 70, and fewer than 1% clear 90.
Overall attractiveness score distribution
% of scans per 5-point bucket · n = 30,000 · 0–100 scale
Jawline definition
the weakest metric · wide spread, heavy low end
Facial symmetry
the strongest metric · tight and high
Key Findings
- Only 2.9% of Mogger Test scans rate Chad. On the PSL Score test, 4.9% reach Chad or above.
- Half of everyone who takes an AI face test lands in the Normie tier (50.1%).
- The average face rating is 65 out of 100. The middle 80% of people fall between 55 and 75.
- Jawline definition is the weakest average metric at 54/100, roughly 13 points behind facial symmetry.
- Neck posture averages just 54.5/100, making it the most overlooked looksmaxxing lever in the data.
- Side profiles score 5.4 points below front-facing photos on average (59.9 vs 65.3).
- The average measured glow-up ceiling is 84.7/100, about 19 points above the average current score.
- Skin quality is the most variable metric: the middle 80% of people spread across 40 points (45 to 85).
How Rare Is a Chad?
Tier results from every completed scan on the two most-used free tools. The pyramid the PSL community always claimed turns out to be real: the top tiers are single-digit percentages.
PSL Score test
n = 223,886 scans (all genders)
Bar length = % of all 223,886 scans (0–100% scale)
Mogger Test
n = 71,694 scans
Bar length = % of all 71,694 scans (0–100% scale)
The Mogger Test maps to four tiers only, which makes its Chad bar stricter: 2.9% versus the PSL test's 4.9% Chad or above. Stacy and Gigachad exist only on the PSL scale.
The Average Face, By Metric
Every structural metric on the same 0–100 axis, sorted best to worst. The band is where the middle 80% of people land; the dot is the mean. Jawline sits at the bottom with the widest, lowest spread.
Band = middle 80% of people (p10–p90) · dot = mean · hover a row for detail
View as table
| Metric | Mean | Median | p10 | p90 |
|---|---|---|---|---|
| Facial symmetry | 67.2 | 70 | 60 | 75 |
| Skin quality | 66.3 | 70 | 45 | 85 |
| Nasal angle (side) | 66.1 | 70 | 45 | 80 |
| Overall attractiveness | 65.3 | 65 | 55 | 75 |
| Eye area | 63.5 | 62 | 45 | 80 |
| Side profile (overall) | 59.9 | 62 | 40 | 75 |
| Neck posture (side) | 54.5 | 55 | 30 | 80 |
| Jawline definition | 54.0 | 55 | 40 | 70 |
The side-profile gap
Side profiles average 59.9 versus 65.3 for front-facing photos. Chin projection, nasal angle, and neck posture only show from the side, and they score worse than anything a selfie measures. 98.2% of app scans include a side photo.
The 19-point ceiling
Where a potential score was computed (n = 4,211), the average ceiling was 84.7 against a 65.3 average current score: roughly 19 points of improvement attributed to changeable factors like body composition, skin, hair, and posture rather than bone structure.
Methodology
Sources. Two datasets feed this page. First, 295,626 completed scans on the free web tools at moggedupapp.com (Mogger Test, PSL Score, and related analyzers) since June 23, 2026, by 92,115 unique visitors, with bot traffic excluded. Second, 91,544 full face scans inside the Mogged iOS and Android app, drawn from 328,035 registered accounts.
How scans work. Each scan analyzes a user-submitted front photo (and in the app, a side photo, present in 98.2% of scans) with an AI vision model that scores structural features from 0 to 100 under a fixed rubric: jawline definition, facial symmetry, skin quality, eye area, and for side photos, chin and nasal angle and neck posture. Web tools map the result to PSL-style tiers; the app produces the full per-metric report. The full scoring rubric, evidence standards, and what we deliberately don't claim are documented on our methodology page.
Sampling. Tier percentages use every completed web scan. Metric averages and distributions use a 30,000-scan random sample of app scans (roughly a third of the corpus), which puts sampling error near ±0.6 points at a 95% confidence level. Counts are per scan, not per person: users can and do scan more than once.
Limitations. This is a self-selected population of people curious enough about looksmaxxing to take an AI face test, which skews young and male; it is not a representative sample of the general population. Scores are AI estimates from photos, sensitive to lighting and angle, and are not clinical or anthropometric measurements. Tiers like Chad and Normie are internet-culture constructs applied consistently, not scientific categories.
Privacy. Everything published here is an anonymized aggregate. No individual photos, scores, or identities are included, and web tool photos are processed for analysis, not stored.
Citing This Data
Journalists, creators, and researchers are welcome to use these statistics with attribution. Cite as "Mogged AI face scan data (2026)" and link to this page. Numbers are refreshed periodically as the scan corpus grows, so linking beats screenshotting. For data questions or custom cuts, reach out via the contact address in the site footer.
Last updated 2026-09-04 · first published 2026-09-04
See Where You Land
Every stat on this page started with someone taking the free 60-second test. Find your tier, then see the full breakdown in the app.