Hue shifts alone are not enough—add shape or pattern redundancy

What it is

A common colour-blind mode swaps one confusable colour pair for a distinguishable one (red-green becomes blue-orange). This only fixes that specific pair for that specific deficiency type; change the palette or the deficiency and the problem returns. Redundant non-colour encoding is the more fundamental fix: every colour-coded meaning simultaneously carries a second, non-colour channel—shape (triangle for enemy, square for ally), pattern (stripes, dots), luminance difference, or a text label. Colour remains (it is the most efficient channel for colour-typical players) but is no longer the only carrier.

Why it happens

Redundancy works by changing how the information exists: colour is a single-dimensional signal (one pixel holds one colour value) and any deficiency type can lose resolution on that dimension; shape is structured spatial signal unaffected by colour vision. With both encodings present, information redundancy rises—a loss in any single perceptual channel no longer destroys the information. This is the core logic of the general accessibility principle "never use colour as the only carrier of meaning." Hue-swap schemes assume there exists a palette distinguishable to all deficiency types simultaneously—an assumption that nearly holds for one type but yields a very limited set across all types, and cannot accommodate individual variation.

Where it stops holding

Redundant encoding carries an information-density cost. Shapes and patterns consume area—small icons (status-effect badges) may become unrecognisable with added patterns, requiring a trade-off between pattern complexity and icon size. Not every visual element can take redundant encoding: photorealistic character models cannot wear printed patterns, so the substitutes are outline strokes, overhead icons, or effect markers. Redundancy should also not be hidden behind a colour-blind toggle—well-designed shape encoding helps colour-typical players too (in fast scenes shapes are caught faster than colours), and gating it behind the mode throws away a general readability improvement.

Applying it

• Assign every colour semantic a fixed non-colour companion—shape, pattern, luminance tier, or text—and write the mapping into the design spec so all subsystems share it.
• For elements that must read at speed and distance (friend-or-foe markers, target indicators), use outlines, overhead icons, or edge strokes rather than surface material colour.
• Verification: desaturate the frame entirely (strip all hue information) and play the core loop. Any information that disappears under greyscale is an element lacking redundant encoding.

Related

• Same group: W8.03.1 Colour-blind modes must cover every colour-coded element · W8.03.3 Different colour-deficiency types need different adjustments · W8.03.4 Colour-only friend-or-foe identification creates real fairness problems
• Nearby: J2.03 Colour accessibility · F4.03 Visual salience · W5.06 Indicators and markers
• Search terms: redundant coding · shape coding · greyscale test · non-colour cues