Color Challenge
Find the one tile that is slightly different.
Ten rounds, each one harder than the last.
Perception varies by display and lighting. This is a color perception game, not a medical eye test.
Round 1 of 10
Difficulty 2/10
Score 0Streak 0 Time 0.0s
How this works
Each round is generated in the browser from a fixed seed, so the same round can be rebuilt exactly and nothing has to be fetched to produce one. The four colours are built in a perceptual colour space rather than by nudging red, green and blue, and the difficulty shown is measured from the real difference between the tiles rather than picked by hand.
A session is ten rounds, and each one is timed. A correct answer is worth 50 points, plus 10 for every level of difficulty, plus up to 100 more for answering quickly — and each correct answer in a row multiplies that total by another tenth, up to 1.5×. Answers past five seconds are still counted, just without the speed bonus: the clock keeps running, it only stops paying.
It is the same engine the rest of PaletteViva runs on. The contrast checker uses it to measure two colours against WCAG, and the color converter to describe one. Nothing you do here is uploaded, and there is no account involved.
The same engine poses two other questions. The color memory game asks you to hold a sequence of colours in mind, and the color reflex game asks you to find one target colour among nine and reports how long it took.
Why two similar colours get harder to tell apart
The smallest difference a person can see between two colours has a name — a just-noticeable difference — and it is a threshold rather than a fixed number. It depends on how large the patches are, how long you look, what is around them, and what kind of screen you are looking at. Two colours closer together than that threshold look like one colour, which is exactly the state this game is trying to put you in.
So the rounds are spaced along that scale. An early round sits around ΔE OK 0.10, something like five times the threshold and findable at a glance; the tenth lands near 0.013, close to the point where two patches stop being separable at all. It never reaches zero: the floor is 0.012, and the two tiles are always at least six steps apart in one channel of eight-bit RGB, so the difference is one a screen can genuinely paint rather than a difference that disappears in rounding.
That is also why the board sits on a fixed grey rather than on the page background, in both themes. A coloured surround shifts how the colour inside it appears — simultaneous contrast — and a game about the smallest visible difference cannot afford a surround that changes with your light or dark setting.
Is this a colour vision test?
No. A result here cannot tell you whether you have a colour vision deficiency, and it is built so that it never pretends to: the tiles differ in lightness as well as hue, so the rounds stay playable for someone who cannot separate two hues, and a screen is not a calibrated instrument whatever the game does with it. The chart an optician uses for that question is a set of printed plates seen under controlled light, and it is the right tool for it.
What the game can tell you is narrower and still worth having: how quickly you picked out a difference of a known size, on this screen, today — and how much that changes when the room is brighter, when you move to another display, or when you play the same level again.
Frequently asked questions
- What does the difficulty number mean?
- It is measured rather than chosen. Once a round’s two colours have been rounded to the eight-bit values a screen can paint, the perceptual distance between them is calculated in OKLab, and the round is labelled with the level whose target that distance matches. Level 1 lands near ΔE OK 0.10 and level 10 near 0.013, so the number on screen is a description of the tiles actually in front of you.
- Is this a color blindness test?
- No, and it cannot be one. The tiles differ in lightness as well as in hue, so the game stays playable without telling two hues apart; a screen is not a calibrated instrument; and nothing here produces a diagnosis. Colour vision deficiency is tested with Ishihara plates and similar charts, by an optician, on paper.
- Why does the same round look different on my phone?
- Because it is a different colour. Two displays differ in gamut, calibration and brightness, so the same hex value is often not the same light coming off the screen — and the room you are in changes how your eyes are adapted when you look. That is why the page says perception game rather than eye test, and why the board always sits on the same fixed grey instead of on the page background.
- How do I score more points?
- Two things, and they pull against each other: answering quickly is worth up to 100 points of the roughly 200 a round is worth, and not missing is worth more, because each consecutive correct answer multiplies the round by another tenth up to 1.5×. A wrong answer costs the multiplier as well as the points, which is why the game rewards a quick look rather than a quick guess.