FaceRating.ai Data Study
Potential vs Actual: How Much Can a Face Rating Really Improve?
Based on 17,032 real analyses · January – August 2026 ·
Key findings
- • The average face rating is 5.59/10 and the average potential score is 6.9/10 — a realistic gap of 1.31 points (median 1.3).
- • 78% of everyone falls between 1.0 and 2.0 points of headroom. The gap is remarkably consistent (standard deviation just 0.42).
- • Only 4.4% have a gap of 2 points or more. Just 15 people out of 17,032 have a gap of 3+, and nobody has a gap of 4.
- • Headroom shrinks as your score rises — from 2.02 points below a 4.0 down to 0.32 points at 9.0+.
- • Of the 4,307 people scoring under 5, not one has a potential score of 8 or higher (the highest is 7.5).
Every analysis on FaceRating.ai produces two numbers: your current score, and a potential score — the model's estimate of where the same face could land by improving only the things a person actually controls (skin, grooming, hair, styling), with no surgery assumed. That second number is the one the looksmaxxing internet argues about, usually without data. So we measured the gap between the two across 17,032 real analyses. The answer is narrower, and more consistent, than almost anyone claims.
The gap is about 1.3 points — for almost everyone
The striking thing is not the average, it is how tightly everyone clusters around it. Four out of five people sit between 1.0 and 2.0 points of headroom. Fewer than 3% have almost no room at all, and only 4.4% have 2 points or more.
| Improvement gap | Share | Distribution |
|---|---|---|
| Under 0.5 | 2.8% | |
| 0.5 – 1.0 | 14.4% | |
| 1.0 – 1.5 | 36.7% | |
| 1.5 – 2.0 | 41.7% | |
| 2.0 + | 4.4% |
Share of 17,032 analyses by size of the current → potential gap. Percentiles: 10th = 0.7, 25th = 1, 75th = 1.7, 90th = 1.7.
Lower scores carry more headroom
Headroom is not evenly distributed. It broadly shrinks as the current score rises: people below a 4.0 average 2.02 points of estimated gain, while those already at 9.0+ average just 0.32. That makes intuitive sense — someone already near the top of the scale has little controllable ground left to make up. One honest wrinkle: the trend is not perfectly smooth. The 7.0–7.9 band averages 0.73 points, slightly more than the 6.0–6.9 band's 0.87, so the decline is a general pattern rather than a clean line.
| Current score | Share | Avg potential | Avg gain | Headroom |
|---|---|---|---|---|
| Below 4.0 | 0.5% | 5.17 | +2.02 | |
| 4.0 – 4.9 | 24.8% | 6.47 | +1.75 | |
| 5.0 – 5.9 | 49.6% | 6.66 | +1.36 | |
| 6.0 – 6.9 | 17% | 7.53 | +0.87 | |
| 7.0 – 7.9 | 4.8% | 8.33 | +0.73 | |
| 8.0 – 8.9 | 3.1% | 8.84 | +0.39 | |
| 9.0 + | 0.2% | 9.51 | +0.32 |
Bands with fewer than 100 analyses are merged into "Below 4.0" so no figure rests on a handful of scans.
Nobody goes from a 5 to a 9
This is the clearest result in the study, and the one worth quoting back at anyone selling a transformation. Among the 4,307 people currently scoring below a 5, the number whose potential score reaches 8 or above is zero. Not a small percentage — none. The highest potential score anywhere in that group is 7.5. Across the whole dataset, only 15 people out of 17,032 have a gap of 3 points or more, and not one person has a gap of 4.
Current and potential scores correlate at r = 0.94 — almost perfectly. In plain terms, your potential score is essentially your current score plus about a point and a quarter. The model is not identifying hidden transformations; it is estimating how much polish is available on the face that is already there.
What this means in practice
A 1.31-point gain is real and worth having — it is the difference between a 5.5 and a 6.8, which is a visible change in how a face photographs. But it is a change of degree, not of category. If you want the practical route to closing your own gap, our data-backed guide to improving your face rating works through the feature-level scores, which ones are genuinely improvable, and what the evidence says actually helps. To read your current number in context, see the full score distribution.
Frequently asked questions
What is a potential score on a face rating?
A potential score is the model's estimate of where the same face could realistically land by improving only controllable factors — skin, grooming, hair, styling — with no surgery assumed. Across 17,032 analyses the average potential score is 6.9/10 against an average current score of 5.59/10.
How much can you realistically improve your face rating?
About 1.3 points. Across 17,032 real analyses the average gap between current and potential score is 1.31 points (median 1.3). 78% of people fall between 1.0 and 2.0 points of headroom, only 4.4% have a gap of 2 or more, and no one in the dataset has a gap of 4 points.
Can you go from a 5 to a 9 face rating?
Not according to this data. Of the 4,307 people currently scoring below 5, not a single one has a potential score of 8 or above — the highest potential in that group is 7.5. Claims that a routine can move you from a 5 to a 9 are not supported by the model's own estimates.
Do lower scorers have more room to improve?
Yes, broadly. Headroom shrinks as the current score rises: people below 4.0 average 2.02 points of potential gain, those at 5.0-5.9 average 1.36, and those at 9.0+ average just 0.32. The trend is not perfectly smooth — the 7.0-7.9 band averages 0.73, slightly above the 6.0-6.9 band's 0.87.
Methodology
This study aggregates 17,032 completed face-rating analyses on FaceRating.ai (January – August 2026). Each analysis records an overall score (0–10) and a potential score (0–10) estimated by the same model in the same pass. The gap is potential minus current, computed per analysis. Bands with fewer than 100 analyses are merged. The correlation is Pearson's r between current and potential scores. All figures are aggregate and anonymized — no individual results, images, or personal data are included. Two limits worth stating: the potential score is a model estimate rather than a measured outcome, and this is a self-selected sample of people who chose to scan their face, not a random sample of the population.