METHOD / VERSION 1 / 31 OCTOBER 2026
Numbers need a definition.
BAUS Score describes how a styling idea fits your selected preferences and setting. The golden-ratio observation is a separate mathematical comparison of two photo ratios. Neither is a measurement of your attractiveness.
BAUS Score: preference fit
First, the planner removes ideas that conflict with your explicit settings: for example, a new piercing is never prescribed, facial-hair ideas require your opt-in, and a top hat requires an evening white-tie brief. It then ranks eligible ideas with five transparent components.
| Term | Definition, each from 0 to 1 |
|---|---|
| G · goals | Match between your chosen goals and an idea's editorial tags. |
| A · accessories | 1 − |idea intensity − selected intensity| / 100. |
| D · dress code | 1 for the selected level; 0.75 for an eligible, less expressive level. |
| S · setting | 1 after the idea passes the setting filter. |
| P · presentation | 1 for your selected presentation or an all-styles idea; 0.5 for another presentation. This is a soft preference, not a restriction. |
When no goals are selected, G is 0.75. Otherwise it uses matching goal tags, with a half-point low-maintenance bonus for low-effort ideas, divided by the smaller of two or your goal count, capped at 1. For a look containing several selected ideas, the overall BAUS Score is their arithmetic mean. It does not model interactions between garments.
The weights and tags are product choices, not coefficients fitted to an attractiveness study. A score of 80 means a strong match under this rule set; it does not mean 80% attractive, an 80th percentile, or an 80% chance of success.
The golden ratio: mathematical closeness
The constant φ solves φ² = φ + 1. Its positive value is:
For an accepted front-facing capture, we compare two fixed projected ratios: mesh face height divided by midface outline width, and mouth-corner span divided by nasal-base mesh span. These are model landmarks in one photograph, not measurements of physical bone or tissue.
Closeness = 100 × exp(−D)
The log distance treats equal multiplicative deviations symmetrically: φ × 1.2 and φ ÷ 1.2 have equal closeness. Both ratios must be available; missing data never becomes a zero score. The choice of these two ratios and this mapping is an explicit product convention.
Worked example
If both ratios equal φ, closeness is 100. If one equals 1.2φ and the other φ/1.2, closeness is about 83.3. That result describes arithmetic similarity to a constant. It establishes no preference, diagnosis, ideal face or need to change anything.
What the research supports
Rossetti and colleagues' primary study does not establish a universal relationship between golden proportions and judged facial attractiveness. Its limited sample and rating design cannot validate this product. We therefore keep φ out of styling rankings and do not recommend correcting a face toward it.
Ward and colleagues' perspective analysis explains why camera distance affects projected nasal proportions. Lighting, expression, pose and model error also change observations. Our capture thresholds are operational checks, not calibrated confidence intervals.
200 observations across seven photographs
The guided capture protocol defines 200 view-specific slots: four frontal captures × 41, plus three profile captures × 12. Frontal views repeat 33 mesh measurements, three derived ratios and five photo fields. Profile views repeat seven experimental model projections and five photo fields. The interface shows the number actually measured, with an explanation for each missing result.
These are correlated observations across expressions, not 200 independent anatomical traits. The expression names are supplied by the person taking the photographs; the engine does not verify a smile or frown. Profile geometry has separate pose checks and uses a bridge-to-chin reference span, with no iris-distance scale. Its accuracy on real profile captures is unvalidated.
Style, profession and personal choice
Industry gives a starting accessory range. Change the dress code or setting to match your actual workplace. The band is an editorial suggestion, not an employer rule. Safety and respirator requirements take precedence over style; OSHA's respirator standard requires that facial hair not interfere with a tight seal.
Men's, women's and all-styles labels are browsing choices. We never infer gender from a photograph. A moustache or frame-shape trial can change visual emphasis, but we do not claim that a wider nose requires either, or predict an attractiveness improvement.
Seasonal references include Vogue's September 2026 accessories review and GQ's September 2026 menswear forecast. These are editorial observations, not scientific evidence of an appearance effect.
Try the guided capture · Build your style plan · Measurement methods