Tap the pad to start, then click or tap it or press any key on the green circle, and keep still on the red square.
green circle
red square
amber diamond
blue triangle
purple cross
grey circle
grey square
grey diamond
grey triangle
The cards: red ink sits in a square frame, green ink in a circle and blue ink in a triangle, so the ink can be told without its colour.
This test switches shapes and lamps on and off, never more than three times a second and with no fades. Skip it if flashing lights trouble you.
More options
The standard run. Ten signals after random waits of 1.5 to 4 seconds: seven green circles and three red squares, shuffled afresh every time. Press on green and keep still on red. The result is your average on the green circles in milliseconds, plus your false alarms and misses.
–This run
–Your best
–Last run
10 signals after a random wait of 1.5 to 4 seconds: 7 to go on and 3 to hold on, shuffled every run. Go on the green circle; keep still on the red square. A signal stays for 1.5 seconds. A press on a hold is a false alarm, and a go signal left unanswered is a miss. The result is the average of your hits in milliseconds, with the false alarms and misses beside it. A press before the signal is a false start and the try is repeated.
Green go, four colours holdTwelve signals: eight green circles and four holds, one each of the red square, amber diamond, blue triangle and purple cross. W. E. Hick's paper of 1952 found that reaction time lengthens with the number of possible signals when each signal has its own response. Here the response never changes and only the holds multiply, so the law makes no promise. Play it after the standard round and see whether your average moves.
Green go, red holdThe standard run. Ten signals after random waits of 1.5 to 4 seconds: seven green circles and three red squares, shuffled afresh every time. Press on green and keep still on red. The result is your average on the green circles in milliseconds, plus your false alarms and misses.
Mostly go, rare holdTwenty signals, and 17 of them are green circles. Pressing turns into a rhythm, and the three holds arrive when the rhythm is strongest. Set your false alarm count here beside the one from the standard run, where three signals in ten were holds.
Mostly hold, rare goTwenty signals and only five are green circles. The other 15 are red squares and amber diamonds that you sit through without moving. No single signal is hard. The danger is a miss: a green that comes and goes while your hand has settled. The average is taken over your hits alone, so five at most.
Show all 25 rounds and notes (21 more)
Change the go colour 3
Blue triangle goGo on the blue triangle, hold on the green circle and the red square. Ten signals, seven of them go. Green has meant go for so long, on the road and on this site's other pads, that keeping still for it takes a run or two to learn. Its best is kept apart from the standard run's.
New go colour every runWhen a run starts the page draws one of the five coloured shapes as the go signal and names it on the pad, for example: Go on the amber diamond. The other four are holds. Twelve signals, eight of them go. Read the rule before the first wait ends, because the next run will have a different one.
Red go, green holdThe standard run with the meanings swapped: press on the red square, keep still on the green circle. Same ten signals, same seven to three split, same waits. Any gap between this average and your standard one comes from the swap alone.
Harder rules 7
Colour words, go on green inkThe pad shows the word RED, GREEN or BLUE in red, green or blue ink, inside a frame of the ink colour's shape. Go when the ink is green, whatever the word says. J. R. Stroop showed in 1935 that a clashing colour word slows people naming the ink and draws errors. He timed spoken answers and this round times a press, so treat the gap between matching and clashing cards as your own finding.
Colour words, go on the word GREENThe same cards with the rule turned round: go when the word reads GREEN, whatever ink it is printed in. Studies of the Stroop effect find that a clashing ink hardly disturbs word reading, because a fluent reader takes in a word's meaning automatically. When both word rounds are done the page sets the two averages side by side.
Grey shapes onlyEvery signal is the same mid grey, so colour tells you nothing: go on the circle, hold on the square, the diamond and the triangle. Play it if you cannot tell this page's colours apart, or to find out whether a shape takes you longer to recognise than a colour does.
Pale colours, same shapesThe standard run with every colour washed out towards the pad's background, though never below a contrast of 3 to 1. Henri Piéron's formulae of 1920 describe reaction time falling as a signal gets stronger, down to a floor, so a faint signal costs time. Find out how much on your own screen.
Quick fireFifteen signals with waits of only 0.8 to 1.5 seconds, and each signal leaves the pad after 1,000 ms instead of 1,500. Ten are green circles, five are red squares or amber diamonds. A faster pace leaves less time to check the colour before your finger moves, so watch what it does to your false alarms as well as to your average.
Stacked lampsThree lamps in a column, in the order of a vertical traffic signal: red square at the top, amber diamond in the middle, green circle at the bottom. One lights at a time, and you go on the bottom one. Position is a third cue beside colour and shape, and it is the one that drivers who confuse red and green learn to use.
Two go colours, two hold coloursGo on the green circle or the blue triangle, hold on the red square or the amber diamond. Twelve signals, eight of them go. You carry two go rules at once, and the result lists an average for each go colour as well as the overall figure, so you can see which rule is the weaker.
Warm up and compare 3
Colour warm-up, six signalsFour green circles and two red squares, and nothing is saved. Use it to learn the pad: where the shape appears, how long a hold lasts, what a false alarm looks like. A cold first press costs nothing here.
Green every time, no decisionSeven green circles and no holds, on the same pad with the same waits. This is a simple reaction: nothing to recognise, only something to notice. Franciscus Donders subtracted one kind of reaction time from another to estimate how long a mental step takes, and the page does that sum for you: standard average minus this one.
Twenty signals, early against lateFourteen green circles and six red squares in one long run. The result prints the average of your first five hits beside the average of your last five. Read the false alarm count with it: if the last five hits are quicker and you also pressed on red, that is haste, not sharpening.
Colours, shapes and scoring 8
Amber diamondPeople with red and green colour blindness commonly mix up red, green, orange and brown with one another, and yellow with bright green. Amber sits in the middle of that set, which is why it never appears on this page without its diamond.
Blue triangle and purple crossBlue and purple are another pair that red and green colour blindness runs together: to a protan, violet, lavender and purple look like shades of blue. The rarer tritan type has trouble separating bluish hues from greenish ones instead. A triangle and a cross cannot be mistaken for each other or for the circle.
Colour is never the only cueThe web accessibility guidelines say colour must not be the only visual means of prompting a response (WCAG success criterion 1.4.1, Level A), and suggest shape or text beside it. Every signal on this page carries a shape, the word rounds carry a framed shape for the ink, and Stacked lamps adds position.
Green circleThe default go signal. The pairing is borrowed from traffic signals in the eastern provinces of Canada, where lights are sometimes told apart by shape as well as colour: a circle for green, a diamond for yellow and a square for red. On this pad the shape is the whole signal, never a small badge in a corner.
Hit, miss, false alarm, correct holdThe four things that can happen on a signal, in the terms of signal detection theory. A go signal and a press: a hit, the only outcome that is timed. A go signal and no press inside the window: a miss. A hold signal and a press: a false alarm. A hold signal and no press: a correct hold, which the textbooks call a correct rejection.
How common red and green colour blindness isUp to 1 man in 12 and 1 woman in 200, among people of Northern European descent. The commonest single type is deuteranomaly, which affects the cones tuned to medium wavelengths. That is a lot of visitors, and it is the reason no round here asks anyone to separate red from green by hue alone.
Red squareThe default hold signal. To a person with the protan type of colour blindness, reds are dimmed, sometimes so far that they pass for black or dark grey and a red traffic light seems to be switched off. So every shape here has an edge in the pad's text colour, and a red that looks dim still reads as a square.
Why speed and false alarms are scored togetherAnyone can lower an average by pressing at everything. Detection theory calls that a liberal bias: more hits, paid for in false alarms. So the page keeps two bests for every round: your lowest average, and your lowest average from a run with no false alarms and no misses. Chase the second one.
How this list was compiled: Rounds and notes written for this page. Every round changes one thing about the standard run (the go colour, the number or share of hold signals, the cue that carries the rule, the pace, or the strength of the colours), so two results can be compared fairly. The arithmetic was computed on 2026-09-19: the go and hold counts of every round (tries minus holds), the shares (7 of 10 is 70%, 17 of 20 is 85%, 5 of 20 is 25%), and the flash ceiling (the shortest wait on the page is 800 ms, so signals can appear at most 1.25 times a second, under the limit of three). Facts checked on 2026-09-19. Donders' ordering of simple, recognition and choice reactions, his subtraction method and the later failure of its additive assumption, go and no-go tasks, reaction time lengthening with the number of possible signals, and Henri Piéron's formulae of 1920, in which reaction time falls as stimulus intensity rises, down to a floor: https://en.wikipedia.org/wiki/Mental_chronometry. Hick's law, decision time growing with the logarithm of the number of equally probable choices, from W. E. Hick's paper of 1952, whose experiment paired each of ten lamps with a key of its own: https://en.wikipedia.org/wiki/Hick%27s_law. Colours confused in red and green colour blindness and in the tritan type, reds dimmed for protans to the point where a red traffic light may appear extinguished, deutan types tied to the medium-wavelength cones, prevalence among people of Northern European descent (up to 1 male in 12 and 1 female in 200), deuteranomaly as the commonest single type, memorising the position of lights as the main coping mechanism, the red, amber, green order from top to bottom, and the shaped signals of the eastern provinces of Canada (red a square, yellow a diamond, green a circle): https://en.wikipedia.org/wiki/Color_blindness. Hit, miss, false alarm and correct rejection, and liberal and conservative response bias: https://en.wikipedia.org/wiki/Detection_theory. The Stroop effect, first published in English by John Ridley Stroop in 1935 (clashing colour words slow the naming of the ink and draw errors, and the interference disappears when the task is to read the word): https://en.wikipedia.org/wiki/Stroop_effect. Colour must not be the only visual means of prompting a response (WCAG 2.2 success criterion 1.4.1, Level A): https://www.w3.org/WAI/WCAG22/Understanding/use-of-color.html. No more than three flashes in any one second (success criterion 2.3.1, Level A): https://www.w3.org/WAI/WCAG22/Understanding/three-flashes-or-below-threshold.html. Nothing on this page tests colour vision. Licence: own work
How it works
This color reaction test shows a coloured shape after a random wait of 1.5 to 4 seconds. Click the pad, tap it or press any key on the green circle. Keep still on the red square. Ten signals give your average on the greens in milliseconds, plus your false alarms and misses. Every colour has its own shape, so colour-blind players can play.
Read the full guide
Hits, false alarms and misses
Pick Colour warm-up, six signals in the Round menu and tap the pad. Nothing is saved, and the shapes follow by themselves. Then play Green go, red hold: seven green circles and three red squares in a fresh order. A press on green is a timed hit. A press on red is a false alarm, never timed. A green left for 1,500 ms is a miss. Then try Blue triangle go or Stacked lamps.
What the decision costs
In 1868 Franciscus Donders reported that a simple reaction is quicker than one where you must first recognise the signal. Green every time, no decision gives your simple figure on the same pad, and the page subtracts it from your standard average. Treat that gap as rough: later experiments found that mental stages do not add up so neatly.
Judge a run by both numbers. Pressing at everything lowers the average and piles up false alarms. Chase the clean best: no false alarms, no misses.
Learn the five shapes first. Then a dim screen or a night-mode tint cannot cost you a signal.
Trying this with friends? Compare each person's own two rounds, never one person's time with another's.
Questions and answers
Can I take this test if I am colour blind?
Yes. Every colour is tied to one shape: green circle, red square, amber diamond, blue triangle, purple cross. Red and green colour blindness affects up to 1 man in 12 of Northern European descent, and to some a red light can look switched off. Here a dim red is still a square. Grey shapes only drops colour altogether.
Why is my time slower here than on the reaction time test?
You must recognise the signal before you may press. The reaction time test asks only that you notice a change. Hick's law covers a related case, a separate response for every signal: decision time grows with the number of choices, though doubling them does not double it.
Can this test tell me whether I am colour blind?
No. The shapes let you score well without telling red from green, so no run says anything about your colour vision. A slow average can come from a cold start, a dim screen or a sticky mouse button. If colours give you trouble, see an optometrist.