Interface color discriminability should be designed along opponent channels, not isolated hue

What it is

Judging "how easy is it to tell these colors apart" cannot rely on hue-wheel angular distance alone — it needs to look at how far apart the colors read on each of the red-green, blue-yellow, and black-white opponent channels. Two colors that look far apart on a hue wheel, if they happen to sit close together on one particular opponent channel (both near the neutral point on the red-green axis, say, relying mainly on the blue-yellow axis for their separation), can become nearly indistinguishable for a user whose that channel is impaired — a risk invisible from hue-angle distance alone.

This condenses the mechanistic knowledge about opponent coding, afterimages, and forbidden colors covered elsewhere into one directly actionable design criterion: how safe or how salient a color pairing is should be calculated by its separation across the opponent channels, not by hue intuition alone.

Why it happens

Opponent channels work as a basis for color-discriminability decisions because they map onto two real-world constraints. First, the most common color vision deficiency in the population — red-green deficiency — is essentially a substantial weakening or loss of function specifically in the red-green opponent channel; affected users still rely mainly on the blue-yellow and black-white channels to tell colors apart. Second, even with normal trichromatic vision, the red-green and blue-yellow axes are not equal in spatial resolution and detection sensitivity — red-green typically performs better, while blue-yellow is inherently weaker and degrades faster in peripheral vision.

This means the "discriminable distance" between two colors actually needs to be broken down by how much separation each individual channel contributes. If the difference between two colors is concentrated almost entirely on the red-green axis, that difference is essentially absent for a red-green color-vision-deficient user, who can only fall back on whatever residual luminance difference remains. If the difference is spread across the blue-yellow and black-white axes as well, it survives under most color-vision conditions. Judging a palette by isolated hue tends to overestimate the real discriminability of color pairs whose separation rests entirely on the red-green axis.

Where it stops holding

• This mainly applies to scenarios that must guarantee discriminability across different color-vision populations — status indicators, data visualization, map region coding. If a color is purely decorative and carries no information that every user needs to read out, it doesn't need to be constrained this strictly.
• The blue-yellow axis itself weakens further in peripheral vision and under low illumination. If a critical color difference lives only on the blue-yellow axis and also appears in peripheral vision or dim conditions, discriminability problems can still arise — it is not the case that "avoiding the red-green axis" is automatically safe.
• This does not replace the more basic rule of luminance contrast. No matter which opponent axis a color difference is designed along, insufficient luminance contrast will still undermine discriminability; this criterion applies on top of already-adequate luminance contrast, addressing the hue-level discrimination problem beyond it.

Applying it

• When designing a set of status or category colors that need to be told apart, don't just check whether they "look far apart" on a hue wheel — check whether they also differ clearly on the black-white (luminance) channel, since luminance difference is the source of discriminability that survives under every color-vision condition, making it the most reliable one.
• For products with higher discriminability requirements (data visualization palettes, error/warning/success status colors), check candidate palettes with a color-vision-deficiency simulation tool (simulating red-green deficiency), focusing on whether color pairs that look "very different" under normal vision get compressed to near-indistinguishable under the simulated view.
• Avoid betting a critical binary or multi-way status distinction entirely on the red-green axis (the classic "red means error, green means correct" is a textbook risk pattern); layer in at least one redundant channel — shape, icon, or luminance difference.
• Verification: screenshot the final palette under a color-vision-deficiency simulator and check each pair of colors that needs to be distinguished to confirm it remains discriminable in the simulated view; any pair that blurs together needs its luminance adjusted or needs to move to colors on a different opponent axis.

Related

• Same group: A1.05.5 Past the cones, signals are re-coded into red-green, blue-yellow, and black-white opponent channels · A1.05.6 Opponent coding explains why combinations like "reddish green" don't exist · A1.05.7 Color afterimages are caused by adaptive reversal in opponent channels
• Nearby: A1.06 Color vision deficiency · A1.04.1 Discriminability is driven by luminance contrast, not hue difference
• Search terms: opponent color design · color blindness simulation · color-safe palette · data visualization color