The categorical-colour limit is lower than the category count most charts actually use

What it is

A categorical colour capacity mismatch arises when a chart asks viewers to distinguish more simultaneous categories than its palette reliably supports for the target task and conditions. “Lower than most charts actually use” names a common design-demand mismatch, not a census claim about every published chart, and it supplies no universal colour count. Colour detection, pairwise discrimination, naming an isolated colour, and finding a target category in a multicolour chart are distinct abilities.

Why it happens

Pairs in a categorical palette are not equally distant. Background, mark geometry, neighbouring colours, device gamut, and colour-vision condition change achieved differences. As categories multiply, near neighbours and legend entries usually multiply too; readers must discriminate colours, maintain colour-to-category bindings, and search. Colour-name boundaries sometimes aid memory, but they vary with language, experience, and the sampled colours and are not a fixed capacity law.

Studying it

Factorially vary category count, palette, mark type, and density in the target chart, randomising colour-to-category assignment. Test detection, pair discrimination, identity without a legend, legend matching, and target search separately. Record accuracy, response time, eye movements or legend consultations, and confusion by colour pair. In Tseng and colleagues' multiclass scatterplots, category increases accompanied slower or less accurate performance within their palettes, category range, and mean-judgment task; such bounded results do not establish the prevalence implied by “most charts” or a universal capacity. Include target devices, themes, and relevant colour-vision groups; preserve tasks and success criteria when comparing experiments.

Where it stops holding

Direct labels, stable domain conventions, filtering to a small active subset, or highlighting one object at a time can reduce the identities that colour alone must recover simultaneously. Performance on large swatches does not transfer directly to thin lines or small points, and trained readers do not stand in for first use. If colour also encodes order or state, those meanings may conflict with category identity.

Applying it

• Count categories that must be distinguished by colour alone in the current view, not all categories in the dataset or rows in a legend, and define acceptance for the actual task.
• Test the complete palette on the target background, smallest marks, highest density, and real devices. Report the worst colour pairs rather than only mean colour distance.
• As categories grow, prefer filtering, faceting, direct labels, stable position, or validated redundant shapes over compressing hue spacing.
• Provide category names, structured data, and noncolour state cues. Use colour-vision simulation to flag candidate risks, not as a substitute for real-task validation.

Related

• Same group: U1.09.1 Every visual channel supports a limited number of reliably discriminable levels · U1.09.2 Beyond the limit, added levels bring confusion, not information · U1.09.4 Discriminable-level limits drop as marks shrink · U1.09.5 Over the limit, aggregate categories instead of subdividing the encoding
• Adjacent: U4.01.1 The number of distinguishable categories has a ceiling · U1.05.3 Smaller colour patches need larger hue differences to stay discriminable
• Search terms: categorical colour capacity · palette discriminability · colour identification · visual search