Three short tests that look at your color vision from different angles. Together they point to which cone type is affected (red, green or blue) and roughly how strongly.
Start with step 0, the screen check: screen settings and room light can change your score. Even after it, a monitor isn't a calibrated clinical instrument, so treat the result as a well-informed guess and confirm it with an optometrist (an anomaloscope test settles the exact type).
The tests measure very small color differences, so your screen and room matter. A web page can't read or change your display settings, so go through this list once on each device and tick each item as you go. Items that don't apply to your device can be ticked too.
Each strip has five squares, numbered from faintest (1) to clearest (5). On a well-set screen you can see at least squares 3, 4 and 5 in both strips.
Which of these grays looks the most neutral, without any yellow, pink or blue tint?
Light or dark mode doesn't matter: while a test is open the page switches to the same mid gray in both, so your eyes adjust the same way on every run. The theme button is paused until you leave the test.
A ring shaped like the letter C is hidden in the dots. It differs from the background only in color; brightness is scrambled so you can't use it as a clue. Say which way the opening faces. When you can't see a C, say so instead of guessing. Each answer narrows down how faint a color you can see, separately along the red, green and blue confusion lines, so the test adapts to you and stops as soon as it knows enough.
Keys: ← ↑ → ↓ for the opening, Space for "can't see".
About 15–23 trials, 2–3 minutes.
Starting from the reference cap on the left, pick the cap that looks most similar to the last one you placed, and keep going until the tray is empty. Click a placed cap to send it back. People with a color deficiency tend to jump across the color circle along their own confusion line, and the direction of those jumps reveals the type.
Strips A, B and C are the reference recolored the way people with each kind of dichromacy see it. A person with that type sees their own strip as identical to the reference; others look different. Rate each strip honestly, comparing swatch by swatch.
Simulation: Machado, Oliveira & Fernandes (2009), blended linearly for partial strength. Color boost: Fidaner daltonization (a blue-channel variant for tritan). Photos from Wikimedia Commons, CC0 / public domain; see img/CREDITS.md. Tap a photo to see it full screen.
Windows, iPhone, Android and Mac have a built-in color filter for color-blind users that recolors everything on screen so confused colors become easier to tell apart.
| Type | Affected cone | What it's like | Prevalence |
|---|
Prevalence figures are for people of Northern European descent; red-green types are X-linked, so they're about 20 times rarer in women. Tritan types are not sex-linked.
Color blindness can't be cured, but most blindness worldwide can be prevented or treated cheaply. If you'd like to help, these two charities do this work well:
Restores sight with low-cost cataract surgery and trains local eye surgeons in Africa, Asia and the Pacific.
Donate to Fred Hollows ↗Prevents childhood blindness and death with vitamin A supplements. Rated a top charity by GiveWell.
Donate to Helen Keller Intl ↗This site isn't affiliated with either charity and receives nothing from donations. Links go straight to each charity's own donation page, with no tracking codes.