A qualified chart still separates its categories after every color is stripped away in greyscale

What it is

A qualified chart must still read after its colour is stripped: converted to greyscale (or printed in black and white), categories must stay separable, order must stay legible, and critical differences must stay visible. The greyscale test is the strictest shared baseline — it does not simulate one type of colour vision deficiency; it asks directly how much of the chart's separability is being carried by hue alone. Whatever hue is carrying fails simultaneously for black-and-white printing, photocopy circulation, low-gamut projection, and colour-deficient readers.

Why it happens

Greyscale conversion replaces every colour with its lightness. What remains distinguishable after conversion was carried by lightness — the channel that nearly every viewing condition preserves (all colour-vision types, monochrome printing, low-grade screens). What merges after conversion was carried by hue alone, and shatters under exactly those conditions. Greyscale therefore works as a stress test of encoding robustness: a chart that passes has staked its entire separability on the most universal channel and is immune to nearly every population-level degradation.

Where it stops holding

Greyscale readability is a robust sufficient condition, not a one-to-one equivalent of any colour deficiency: deficient readers retain some hue perception, and a few pairs that merge in greyscale remain distinguishable to them (equally light red-green pairs, for instance, separate for some deficiency types); conversely, because their spectral sensitivity curves shift, their lightness perception can deviate from photometric greyscale. So failing greyscale is near-conclusive evidence of unsafety, while passing means "safe under the worst conditions" — it does not guarantee the best experience for every colour-vision type. Greyscale also erodes legend-swatch discrimination, so the legend must be converted along with everything else during testing.

Applying it

• Before delivery, convert the entire chart — legend, annotations, status dots included — to greyscale: categories merging, order reversing, or key points vanishing each send the palette back for rework.
• For categories that cannot be spread across lightness in greyscale, add shape, pattern, or direct labels; do not fall back to "a separate print colourway" — two legends always drift out of sync eventually.
• Put the greyscale check on the release checklist alongside CVD simulation: simulation probes specific populations, greyscale probes every degraded condition, and both must pass.
• Verification: ask a reader to retell the chart's main conclusions from the greyscale version alone; missing or inconsistent conclusions mean information is still bound to hue.

Related

• Same group: U4.04.1 Red-green combinations fail for a substantial share of viewers · U4.04.2 Lightness differences must guarantee separability
• Nearby: U4.05.1 Print, glare, and colour deficiency all strip colour · U4.05.5 Verify alternatives at design time, not after feedback
• Search terms: greyscale test · print safe colour · encoding robustness