Why most face-rating apps inflate scores (and where we draw the line)
Hey Reader,
An honest note about face-rating apps. including ours.
A lot of face-rating apps return inflated scores. There are a few mechanical reasons, and it's worth naming them before you take any score (including ours) too seriously.
1. Beauty filters baked into the model
Apps like FaceApp normalize faces toward an "aesthetic ideal" before scoring. The smoother, more symmetric input lifts every score. You're not actually a 9. the model already pre-applied the filter.
2. Photofeeler-style apps measure votes, not faces
Photofeeler returns a score from a small sample of human voters. That's useful for ranking your own photos against each other, but the absolute score moves with sample bias (time of day, who's online). It's a relative tool, not a measurement.
3. Engagement-optimized scoring
Free apps need you to come back. Returning a 7-8 to almost everyone is the path of least resistance. Honest 4s and 5s churn the install.
Where RealSmile is honest. and where we still have limits.
We measure 17 geometric and photo-quality features against published distributions. The score isn't normalized to make you feel good. If your jawline angle is at the 30th percentile, the report says 30th percentile.
But we have limits too: lighting matters more than most users realize, the 17 metrics don't cover every aspect of attraction (personality, voice, context), and the scoring distribution is built on people who chose to scan. not a true population sample. We're transparent about all of that.
No pressure. the email tomorrow is the protocol.