Where the bands come from. The ideal ranges are not invented. They are fitted on SCUT-FBP5500 — 5,500 frontal faces, each rated by 60 volunteers (330,000 ratings, IEEE ICPR 2018) — restricted to the 2,750 male faces. For each measurement the band is the median and interquartile range of the best-rated 20% of faces, and its importance weight is the effect size (Cohen's d) between the best- and worst-rated fifth. The PSL number is a position on a normal curve, which is how the rating communities define it: 4 sits at the midpoint, one point is one standard deviation, and the band cutoffs are theirs (Sub3 below 1.5, LTN 1.5–3.5, MTN 3.5–5, HTN 5–6, Chadlite 6–7, Chad 7+). That is why PSL 5 means ~84th percentile and PSL 6 ~98th, and why the shares beside each band come out as they do.

Where the PSL scale and the data disagree. The canon weights facial dimorphism heavily and the attractiveness data does not: of the twelve features the community treats as decisive, five are not supported. Wider fWHR, a squarer jaw, a low gonial angle and hooded eyes all rate worse with real raters, and canthal tilt — the single most-cited PSL feature — has essentially no correlation at all (r = −0.01). One correction worth stating: the gonial angle, bigonial width and upper eyelid exposure measurements each correlate 0.37–0.46 with camera pitch, and our gonial median is 139° against a literature 120–130°, so those three conflicts are partly photo artefact rather than pure rubric disagreement. The fWHR and canthal-tilt findings stand; the jaw and eye-hooding ones need a pitch-corrected measurement before they count as evidence.

Two PSL numbers, and where each comes from. The harmony algorithm figure is our own measurements mapped onto the PSL scale — geometry only, no community input. The looksmax.org data figure comes from faces the community has actually rated, and is still being built. Both are labelled because the provenance is the point: an unlabelled pair of PSL numbers was the thing that made the old version untrustworthy.

What the harmony number is not. It is a scale mapping, not a validated PSL prediction. We tested it against 24 faces the community tiers by half-PSL — sourced by identity from Wikipedia page images, since an earlier search-based attempt pulled the wrong people — and only 15 had photos clean enough to measure. On those, our ranking correlated with the canon at r = +0.28 and Spearman 0.43, which at n = 15 is indistinguishable from zero. Sean O'Pry (canon 7.5) measured 4.11 and Michael Cera (canon 3.0) measured 3.91. Fitting a curve through that would have manufactured agreement, so none was fitted.

What that is worth. The per-feature band score agrees with the human mean rating at r = 0.43 on held-out faces. The hand-written rubric it replaced scored r = −0.09, i.e. slightly anti-correlated. The joint model shown on the right reaches r = 0.73 as shipped (0.75 in-fit, before dropping race to a population mean). Two measurements are excluded from scoring because they were measured to be invalid: the upper third (the mesh top is the forehead, not the trichion) and symmetry (which behaves as a template-fit residual, not asymmetry).

Method. MediaPipe FaceLandmarker gives 468 mesh vertices plus 10 iris points in 3-D, averaged over five rotation-jittered passes. Landmarks are converted to one isotropic unit, the canonical 468-vertex face model is fitted (Horn quaternion / Kabsch), that rigid transform is inverted, and the ratios are measured in the canonical frontal plane.