Color Memory Game
Colour vision screening

Color blind test

Ten Ishihara-style plates, about a minute, free and with no signup. Each plate hides a digit in dots that differ from the background in hue alone, and the gap between those two hues narrows as you work through the set. At the end you get a plate-by-plate breakdown and a read on whether your misses followed the pattern a screening test is looking for.

One in twelve men and one in two hundred women has some form of colour vision deficiency, and a large share of them go years without knowing. This will not diagnose you. What it will do is tell you whether the wide-gap plates gave you trouble, which is the signal worth acting on.

Ten plates, about a minute. Each one hides a single digit in dots that differ from the background in hue only, so brightness gives you nothing to work with. The gap between the two hues narrows every second plate. Answer with the first number you see and do not squint the answer into existence.

Set your screen to full brightness, turn off any night-shift or blue light filter, and take it in even lighting. A warm screen filter alone can flip several plates from readable to impossible.

The mechanism

How a colour blind test actually works

Typical human colour vision runs on three cone types, each most sensitive to a different part of the spectrum: short wavelengths (blue), medium (green), and long (red). Your brain does not read those signals separately. It reads the differences between them, and colour is what those differences feel like. When one cone type is shifted or missing, whole ranges of colour that used to produce different signals start producing the same one, and they become impossible to tell apart.

Shinobu Ishihara worked out in 1917 that you could exploit this with dots. Draw a number in dots whose hue sits on the confusion line for a particular deficiency, surround it with dots of a hue on the other side of that line, and randomise the lightness and size of every dot so there is no edge or brightness cue. Someone with typical vision reads a number. Someone whose cones cannot separate those two hues sees an even field of dots.

The plates here are generated rather than reproduced from the printed set, which has one honest advantage and one limitation. The advantage is that the hue gap is a dial rather than a fixed value, so the test can walk from an obvious 110-degree separation down to a 26-degree one and show you where your reading breaks down. The limitation is that a generated plate is not tuned to a specific confusion axis the way a clinical plate is, so it screens broadly rather than telling you which cone is responsible.

The four types

Types of colour vision deficiency

Deuteranomaly and deuteranopia

Around 6 percent of men

The green cone is shifted toward red (deuteranomaly) or missing entirely (deuteranopia). This is by far the most common form. Greens and reds collapse toward a shared muddy yellow-brown, and the classic difficulty is telling a ripe from an unripe tomato, or a green traffic light from an amber one at distance.

Protanomaly and protanopia

Around 2 percent of men

The red cone is shifted or absent. The confusion is similar to the deutan types, but with an extra symptom: red itself looks darker than it should. Brake lights, red text on a dark background, and anything relying on red as a warning signal lose their punch.

Tritanomaly and tritanopia

Fewer than 1 in 10,000

The blue cone is affected. Blue and green become hard to separate, and yellow can drift toward pink or grey. Unlike the red-green types this one is not sex-linked, so it shows up about equally in men and women, and it is more often acquired later in life than inherited.

Monochromacy

Roughly 1 in 33,000

Either only one cone type works or none do. True total colour blindness, where the world is genuinely greyscale, is rare and usually comes with light sensitivity and reduced acuity. Most people who say they are colour blind have one of the red-green types, not this.

The naming is more consistent than it looks. A name ending in -anomaly means the cone is present but its sensitivity is shifted, which produces a partial deficiency. A name ending in -anopia means that cone type is absent altogether. Prot-, deuter-, and trit- simply mean first, second, and third, referring to the long, medium, and short wavelength cones in that order.

Reading your score

What your result does and does not mean

The number out of ten matters less than where the misses fell. Plates one to four use a hue gap wide enough that essentially everyone with typical colour vision reads them without pausing. If you lost two or more of those, that is the pattern worth taking to an optometrist, and it is the only outcome here that justifies anything beyond curiosity.

Losing the last two plates is normal. At a 26-degree hue gap on a screen that has never been profiled, the digit is genuinely faint for everyone. Treat those as the tie-breaker they are rather than as evidence.

Before you read anything into a poor score, rule out the setup. Night shift, true tone, blue light filters, and reading modes all shift the entire hue axis and can cost you several plates on their own. So can a screen at low brightness, an OLED panel at a steep viewing angle, and a warm bulb reflecting off the display. Run it again on a different device in daylight: if the result moves, the screen was the variable.

A real assessment is not expensive and it checks more than colour. If you have never had your colour vision formally tested and you are heading into a career where it matters, aviation, rail, electrical work, and parts of the armed forces all screen for it, get the clinical version rather than relying on a browser tab.

After the test

Where to go next

If the plates went well and you want a harder read on your colour discrimination, the hue sort is the arrangement-style counterpart to a plate test: you order shuffled chips into a smooth gradient, which catches small discrimination errors that plates miss entirely. The multiplayer plate game runs the same plates against friends in a room.

For the science, how many colours the human eye can see covers the cone maths from the other direction, and the tetrachromacy piece looks at the people who may have a fourth cone rather than one too few. There is also a full write-up on whether colour blind glasses work and what the filters are physically doing.

Curious what your colouring suits rather than what your cones do? The seasonal colour analysis quiz is a different kind of colour test entirely.

FAQ

Colour blindness questions

Is this color blind test accurate?+
It is a screening test, which means it is built to flag a possible issue rather than to name one. It runs on your screen, and screen calibration, brightness, and any night-shift filter all change what the plates look like. Missing the wide-gap plates at the start is a meaningful signal. Missing the last pair is not, because those are hard for people with completely typical colour vision.
How does the Ishihara test work?+
Each plate is a disc of coloured dots, and a digit is drawn in dots of a different hue. Both the digit and the background vary randomly in lightness and dot size, so there is no edge, outline, or brightness difference to work from. Hue is the only thing separating the figure from the field, which is what makes it a test of colour perception rather than of pattern recognition.
What are the four types of color blindness?+
Deuteranomaly and deuteranopia affect the green cone and are the most common. Protanomaly and protanopia affect the red cone and also make red look darker. Tritanomaly and tritanopia affect the blue cone and are rare. Monochromacy, where colour is absent altogether, is rarer still. The '-anomaly' names mean the cone is shifted, the '-anopia' names mean it is missing.
Why are men more likely to be colour blind?+
The genes for the red and green cone pigments sit on the X chromosome. Men have one X, so a single faulty copy produces the condition. Women have two, and the working copy on the second X usually compensates. That is why red-green colour blindness affects roughly 1 in 12 men and about 1 in 200 women. Blue-yellow deficiency is not carried this way, so it affects both sexes about equally.
Can this be used as a color blind test for kids?+
It can, with two adjustments. Children who are still learning their digits will fail plates for reasons that have nothing to do with colour, so sit with them and check they can read the same numerals printed plainly. And take the result even less seriously than an adult's: a young child fidgeting through ten plates on a phone is not a screening. If you have a real concern, school vision screening and an optometrist are the right route.
Can colour blindness be cured?+
Inherited colour vision deficiency cannot currently be cured. The cone pigments are set by genetics and there is no treatment that changes them. Acquired colour vision loss is different: when it comes from cataracts, glaucoma, diabetes, multiple sclerosis, or certain medications, treating the cause can sometimes recover some colour discrimination. A sudden change in how you see colour is a reason to see a doctor rather than to retake a quiz.
Do colour blind glasses work?+
They do something, but not what the marketing implies. The filters cut a slice out of the spectrum where the red and green cone responses overlap, which exaggerates the difference between the two signals. For someone with anomalous trichromacy, where the cones are shifted but present, that can make certain colours easier to separate. For dichromats, who are missing a cone type entirely, there is no second signal to separate and the effect is much smaller. They do not restore normal colour vision, and results vary enough between people that trying before buying matters.
What is the difference between this and a real eye test?+
A clinical colour vision assessment uses printed plates under standardised illumination, often alongside an arrangement test like the Farnsworth-Munsell, and is interpreted by a professional who also checks the health of your eye. This is ten generated plates on an uncalibrated screen. Use it as a nudge, not as a result.