There is a bug that no crash report will ever surface, that passes every automated test, and that a large fraction of your players hit on their first minute of play: the game encoded something important — which unit is yours, which pipe is safe, whether the last move was right or wrong — using colour and only colour, and a meaningful share of people simply cannot see the distinction. They are not confused because the game is hard. They are confused because the game is speaking a language their eyes do not fully parse. Colour vision deficiency is common, predictable, and almost entirely designable-around, and the studios that handle it well rarely do anything exotic. They just follow one rule with discipline.
The problem, in numbers and in practice
Colour vision deficiency — the everyday term is “colour blindness,” though almost no one sees in pure greyscale — is far more common than most designers assume. Around 1 in 12 men (roughly 8%) and about 1 in 200 women (roughly 0.5%) have some form of it. Put that against a player base of any size and the maths is blunt: in a room of a hundred typical players, several of them are not seeing the palette you designed. It is not a rare edge case you can defer to a “phase two accessibility pass.” It is a standing fraction of everyone who will ever open the game.
The deficiency comes in a few well-understood forms, named for the cone cells involved. Deuteranopia and the milder deuteranomaly affect green perception and are the most common; protanopia and protanomaly affect red. Together these red–green types account for the overwhelming majority of cases, which is exactly why the red/green pairing — so beloved of “good vs bad,” “friendly vs enemy,” and “correct vs wrong” — is the single most dangerous design choice you can make. The much rarer tritanopia affects blue–yellow perception. The practical upshot for a designer is that you cannot assume any two hues are reliably distinguishable, and you must treat red-versus-green in particular as if it barely exists.
Why does colour-only encoding fail so reliably? Because when two colours collapse toward each other in a player’s perception, everything you pinned to that difference collapses with them. The red enemy healthbar and the green friendly one become the same murky bar. The “you must click the red gems” objective becomes a guessing game. The status light that turns from green to red to warn of danger just… stays roughly the same. The information was there on screen the whole time — it was simply carried on a channel the player cannot fully receive. The moment you add a second, non-colour channel, the information survives.
The one rule: colour is a reinforcement, never the message
Every other technique in this post is downstream of a single principle: never encode meaning by colour alone. Whenever a colour is doing a job — distinguishing your team from the enemy, marking a tile as correct, flagging a hazard — it must be joined by at least one channel that does not depend on colour perception. There are five reliable ones, and most good designs use two or three at once:
Shape. Give each state its own silhouette. A circle for “safe,” a triangle for “caution,” an octagon for “stop” — the same logic road signs use, and for the same reason. Shape reads instantly and survives greyscale entirely.
Icon or symbol. A checkmark versus a cross, a skull on the dangerous enemy, a shield on the defensive unit. Icons carry semantic meaning that colour never can on its own, and they scale down to tiny UI far better than a subtle hue shift.
Pattern or texture. When two areas must be told apart — territory on a map, segments of a chart, factions on a board — a hatch, a dot fill, or a stripe distinguishes them independent of colour. This is the classic trick for filled regions where shape and icon do not fit.
Text or number. The bluntest and often the best. A label that literally says “Correct,” “Enemy,” or “+250” removes all ambiguity. It costs screen space, so it is not free, but it is unbeatable for critical states.
Position or motion. Consistent placement (“your resources always on the left”), or a distinguishing animation (the wrong answer shakes, the right one bounces), gives the player a non-colour signal that works even in peripheral vision.
Choosing a palette that survives real eyes
Redundant encoding is the safety net, but you should also pick colours that do more work before the net is needed. Two habits carry most of the benefit. First, avoid pairing red and green to mean opposite things; given the prevalence of red–green deficiency, it is the worst-case combination. Blue–orange or blue–red pairings survive far more eyes. Second, and more fundamentally, differentiate on lightness as well as hue. Two colours that differ only in hue can merge under CVD; two that also differ in lightness — a dark oxblood against a light amber, say — stay distinguishable because lightness is a channel colour deficiency does not erase. A quick sanity check is to imagine the design in greyscale: if two states become the same shade of grey, they were relying on hue alone and will fail for someone.
High contrast helps everyone here, not only colourblind players. Adequate contrast between foreground and background keeps text and marks legible in bright sun, on cheap panels, and for ageing eyes, and it happens to make hue differences easier to resolve too. Contrast is the rising tide that lifts every accessibility boat at once.
Colorblind modes, filters, and testing
Many games ship an explicit colorblind mode — a setting that swaps the palette for a CVD-safe one, or adds symbols to colour-coded elements, or applies a corrective filter tuned for deuteranopia, protanopia, or tritanopia. These are genuinely useful and worth offering when a game leans heavily on colour. But a mode is a supplement, not an absolution: a game that is unplayable with the toggle off has simply moved the problem behind a menu that a first-time player may never find. The healthiest posture is a design that already works for most people without the mode, with the mode as an extra accommodation for those who want a stronger correction.
You do not need colourblind testers on staff to catch most problems — you need a simulator. CVD simulation tools (built into many design apps, browser dev-tools, and standalone utilities) re-render your screen as a deuteranope, protanope, or tritanope would perceive it. Run every important screen through each simulation and look for the states that merge, the objectives that become ambiguous, the healthbars that lose their meaning. It takes minutes and it turns an invisible bug into an obvious one. The greyscale test is the poor-man’s version and still catches a surprising amount: if the screen still makes sense in black and white, you are most of the way there.
Doing it right versus doing it wrong
The contrast is easiest to feel through concrete examples. Wrong: a match-3 puzzle whose only difference between gem types is colour, so a red–green-deficient player cannot tell the red gems from the green ones and cannot form the matches the game is asking for. Right: the same puzzle gives each gem type a distinct shape as well — a red heart, a green clover, a blue diamond — so matching works on silhouette alone. That is not a hypothetical pattern; it is why so many long-lived puzzle games put a symbol inside every piece.
Wrong: a strategy game that marks your units blue and the enemy’s red and stops there. Right: the same game adds a distinct outline, a faction icon, or a base-plate shape to each side, so allegiance reads without colour. Wrong: a “connect the wires” puzzle where the wires are distinguished only by hue. Right: each wire also carries a dash pattern or an end-cap symbol. In every case the fix is the same move — add a redundant channel — and in every case it makes the game clearer for all players, not only colourblind ones, because a second channel is simply more legible than one.
How we think about this at WizusLabs
This is not an abstract concern for us. One of the design intentions we hold across the studio is that colour should reinforce information and never carry it alone — a state that matters should also be signalled by shape, position, a label, an icon, or a number, so a screen reads the same whether or not you can tell two hues apart. We try to be honest about the shape of that commitment rather than stamping a badge on it: it is a principle we design toward and keep auditing as our games grow, not a certified guarantee on every single screen. We lay out that posture — along with how we think about readable text, untimed play, offline support, and sound that is never the only channel — in our accessibility guide, and we would genuinely rather hear where we fall short than pretend we have it perfect.
If there is one thing to take away, it is that designing for colourblind players is not a compliance chore bolted on at the end. It is a small, cheap habit — add a shape, add a label, differ on lightness, run the simulator — applied early, that happens to make the game more legible for everyone who plays it. The rest of the WizusLabs Engineering blog digs into the craft and pipeline behind the games we ship.
Sources
The prevalence figures cited above — approximately 1 in 12 men (about 8%) and approximately 1 in 200 women (about 0.5%) with some form of colour vision deficiency — are long-established and widely reported by clinical and awareness organisations:
- Colour Blind Awareness — “Colour Blindness” (prevalence: ~1 in 12 men, ~1 in 200 women).
- National Eye Institute (NEI), U.S. National Institutes of Health — “Color Blindness” (types: deuteranopia, protanopia, tritanopia).
Beyond these prevalence figures, this is a craft-and-practice piece; it makes no other statistical claims. The guidance on redundant encoding aligns with the widely used principle that colour should not be the only visual means of conveying information (Web Content Accessibility Guidelines, Success Criterion 1.4.1, “Use of Color”).
Keep reading: all posts on the WizusLabs Engineering blog.