LOOKSMETER · CALIBRATION & METHODOLOGY
What is the average PSL score? The bell curve & calibration study
Most people taking their first calibrated PSL test are startled when an everyday good photo scores 4.1. Here is the statistical math behind the 0.25-to-8.0 bell curve, how PSL decimals map to population percentiles, how a PSL score maps to an everyday 1-10 rating and how camera distance shifts measured landmarks. Test a 5-foot daylight photo on our calibrated free PSL test →
The PSL bell curve: score-to-percentile & z-score table
Morphometric traits in human biology - stature, interpupillary distance, bizygomatic width and composite craniofacial harmony - follow an approximately normal (Gaussian) distribution. Because individual ratios are only partially correlated, scoring in the upper tail on six ratios at once becomes exponentially rarer.
To prevent top-end bunching, the 0.25-to-8.0 PSL rubric maps roughly one standard deviation (~0.85 to 1.0 PSL points) per step away from a 4.00 mean:
| PSL Score | Tier | Std Dev (z) | Population Percentile | Everyday 1-10 Rating |
|---|---|---|---|---|
| 2.00 | Low Tier Normie (LTN) | -2.0 SD | ~2nd-4th percentile | 5.5 / 10 |
| 3.00 | LTN / MTN Boundary | -1.0 SD | ~16th percentile | 6.5 / 10 |
| 4.00 | Mid Tier Normie (Mean) | 0.0 SD | 50th percentile (Average) | 7.4 / 10 |
| 4.50 | MTN / HTN Boundary | +0.65 SD | ~74th percentile | 7.9 / 10 |
| 5.00 | Solid HTN | +1.0 SD | ~84th-85th percentile | 8.4 / 10 |
| 5.50 | Chadlite / Stacylite Entry | +1.65 SD | ~95th percentile (Top 5%) | 8.9 / 10 |
| 6.00 | Chad / Stacy Entry | +2.33 SD | ~99th percentile (Top 1%) | 9.4 / 10 |
| 7.00+ | True Adam / Eve Ceiling | +3.3+ SD | >99.95th percentile | 9.9 / 10 |
How a PSL score maps to an everyday 1-10 rating
A PSL score and an everyday 1-10 rating use different centre points. An everyday 1-10 rating tends to place an average face near 7 while the PSL scale fixes the average at 4.00. The table above gives the conversion. Three parts of LooksMeter's calibration keep the PSL number steady:
1. A fixed 4.0 median
Scores pass through an explicit calibration curve locked to the 4.0 median, so a 4.3 (High MTN, ~62nd percentile) reflects real statistical standing.
2. Geometric measurement first
The on-device 3D mesh extracts craniofacial landmarks and corrects for head roll before measuring ratios and bilateral landmark symmetry.
3. Deterministic scoring
Deterministic image hashing means the same photo returns the same score.
How LooksMeter's two-layer calibration pipeline works
LooksMeter separates geometric measurement from qualitative visual synthesis so scores stay consistent, repeatable and verifiable:
First, our on-device 3D mesh extracts craniofacial landmarks directly in your browser, correcting for head roll before measuring interpupillary Eye Separation Ratio (ESR), Facial Width-to-Height Ratio (fWHR), Midface Ratio, vertical Facial Thirds, bilateral landmark symmetry, nasal-to-mouth width and profile angles.
Second, consistent scoring rubrics and identity anchoring combine those measured corridors with surface cues (skin homogeneity, brow contrast, hairline framing) through an explicit calibration curve locked to the 4.0 median.
Optical variance: how camera distance shifts PSL scores
Even a deterministic algorithm can only measure the 2D projection you feed it. Because perspective distortion depends strictly on camera-to-subject distance (not megapixel count), moving a phone from 12 inches to 5 feet changes measured facial ratios dramatically:
| Capture Setup | Distance | Nasal Width Bias | Jaw Width Read | Typical PSL Shift |
|---|---|---|---|---|
| Close Phone Selfie (24mm) | 12 in / 30 cm | +26% to +30% wider | Tapered / narrowed | -0.4 to -0.9 PSL penalty |
| Extended Arm Selfie (26mm) | 26 in / 65 cm | +10% to +14% wider | Slightly narrowed | -0.15 to -0.3 PSL |
| Rear Timer / Mirror Distance (50mm eq.) | 5 ft / 1.5 m | ~0% (True human eye view) | Accurate bigonial width | 0.0 (Calibrated baseline) |
| Studio Telephoto Portrait (85-135mm) | 8-10 ft / 2.5-3 m | -3% to -5% compressed | Maximized jaw & cheek plane | +0.15 to +0.35 PSL boost |
Frequently asked
What is the true average PSL score?
On a calibrated 0.25-to-8.0 PSL scale, the true 50th-percentile population median sits at 4.00 (the middle of Mid Tier Normie, 3.00-4.50). Roughly 55% of unedited everyday portraits land between 3.00 and 4.50 while roughly 80% of all faces fall between 2.75 and 5.00.
What percentile is a 5.0 or 5.5 PSL score?
A 5.00 PSL score (solid High Tier Normie) places a portrait near the 84th-85th percentile (top ~15% of the population, roughly +1.0 standard deviation above the 4.0 mean). A 5.50 PSL score marks the entry to Chadlite / Stacylite around the 95th percentile (top 5%) while 6.00 marks roughly the 99th percentile (top 1%).
Why can a ChatGPT rating differ from a PSL test score?
General-purpose chat assistants are built for conversation and take a different approach. LooksMeter anchors roll-corrected 3D facial geometry to a bell curve centered at 4.0 with deterministic image hashing so a 4.3 (High MTN, ~62nd percentile) reflects real statistical standing.
Where can I inspect LooksMeter's live score distribution?
We publish our live, aggregated scoring distribution on looksmeter.com/stats so anyone can verify that real-world scans center near the MTN/HTN bell curve.
Email me the cheat-sheet
Get the cheat-sheet + a first-time member perk
Get our facial harmony & styling reference sheet emailed to you.
Live data, PSL tiers & measurement tools
Live Platform Score Distribution
Real-time histogram and tier shares across LooksMeter scans
Free Calibrated PSL Score Test
Deterministic 0.25-8.0 score, percentile and 4-category bars
The PSL Scale (0.25-8.0)
Complete 7-tier guide and 1-10 conversion chart
MTN (3.0-4.5): The 50th Percentile
Why ~55% of human faces cluster around the 4.0 mean
Craniofacial Literature Review
What Farkas, Perrett & Rhodes studies show about facial harmony
Halos & Failos Breakdown
8 facial regions ranked by perceptual weight and fixability
Keep reading
The TrueRateMe scale, explained (1-9.5 to PSL)
Complete 1-9.5 Gaussian TrueRateMe chart for men and women, exact percentiles, how it converts to 0.25-8.0 PSL and a private free test.
Am I attractive?
How an AI attractiveness test reads your face and how to make sense of the result.
Face rating, explained
What an AI face rating measures, what the score means and how to read the breakdown.
This page is general information about facial aesthetics terminology, not medical or cosmetic advice. Scores from our tools are calibrated reads of one photo that describe presentation, never your worth.