Our color methodology
Original color collections, visible scoring rules and an honest account of what a screen can tell you.
Three qualities, twelve starting points
Temperature describes a warm or cool color direction. Depth describes lighter or deeper color. Clarity describes muted or vivid color. Our twelve anchors summarize common seasonal descriptions. Their numeric coordinates are our implementation choices, not a universal professional standard.
How the palettes are generated
Each collection samples CIELCH using the D65 white point. Lightness, chroma and hue ranges differ by season. A deterministic hash chooses samples; colors outside the sRGB gamut are rejected rather than clipped. Each palette contains 24 supporting neutrals, 40 accents and 8 metal-inspired tones. Supporting neutrals and metals use their own ranges. Names are chosen from the CC0 xkcd color survey; our generated hex values are independent of that dataset.
How the quiz scores
Ten questions contribute observations to temperature, depth and clarity. Main fabric comparisons carry weight 3; a repeated clarity comparison carries 1.5; white/ivory, hair depth and contrast carry 1; jewelry 0.75; eye depth 0.5; veins 0.25. Uncertain answers contribute zero. Weighted axis averages are compared with each season using temperature importance 1.4, depth 1.0 and clarity 1.2. The nearest anchor ranks first.
What confidence means
Confidence combines how much of the questionnaire supplied usable clues and how far the first match sits from the runner-up. A small gap, conflicting observations or missing clues can give a low-confidence result. The agreement score measures proximity to our rules. It is not a measured probability of a correct professional analysis.
How the color checker works
The checker finds the nearest palette chip using CIE76: the straight-line distance between two colors in CIELAB D65. Up to 10 is labeled a close palette match, over 10 up to 22 a flexible option, and over 22 a suggestion to try a softer pairing. These thresholds are editorial choices for this tool, not a clinical or personal suitability test. It compares screen colors and always gives three nearby alternatives.
What the photo tool measures
Three selected patches are averaged in CIELAB D65. The displayed photographed bias is b* minus 0.65 times a*. Values above 8 are called warm-leaning, below 2 cool-leaning, and between them neutral-leaning. If the farthest patch differs from the mean by more than 12 CIE76 units, the result is marked inconsistent. These are exploratory rules, not validated biological undertone thresholds. Lighting, camera processing, makeup and reflected colors can dominate the reading. No face detection or season assignment occurs.
Limits and inclusion
Skin darkness, ethnicity, tanning and sunburn do not determine season. Feature observations can be uncertain across all skin depths. Screen and printer profiles change color appearance; hex chips are a reference rather than a textile standard. Celebrity labels remain publicly cited guesses. A physical comparison is more useful than a single photo or label.