How is the hard mode different from the standard color blind test?

Same plates, same adaptive ladder, same scoring. Hard mode removes the two ungraded warm-up plates, sets the time limit to 3 seconds per plate by default (2 or 5 available), and counts a timeout as a miss instead of granting a second look.

I failed a plate — does that mean I am color blind?

No. Under a short time limit anyone can miss a near-threshold plate, and the engine expects it: a miss opens a best-of-three vote at that level, and the result is drawn from where your threshold settles, not from a single plate. Look at the cone scores and the severity estimate on the result card, and treat them as a self-check.

Is a hard color blind test more accurate than the Ishihara test?

It is more sensitive to mild deficiencies, because it tests at your threshold rather than at fixed contrast, and it covers blue-yellow, which Ishihara plates do not. It is not more accurate about the type or severity — an uncalibrated screen limits both — and it is not accepted for licensing or employment, where a clinical test is required.

Can I practise to get better at it?

You can get faster at the task, and hard mode rewards that a little. You cannot learn the plates, because each one is generated fresh, and you cannot change your cone thresholds by practising. If your score improves on a retake, it is usually because you stopped hesitating.

Hard mode · threshold plates

Hard Color Blind Test

Try more challenging colored-dot plates to explore subtle color differences. Screen conditions affect your result; this is a self-check.

Loading the tool…

Short answer

A hard color blind test is a screening that keeps pushing the colour contrast toward your threshold instead of stopping at the plates most people can read. This one uses the same generated cone-isolation plates and scoring as our standard color blind test, but removes the warm-up plates, cuts the time per plate to three seconds by default, and gives no second look after a timeout — so a near-threshold plate is either read or missed.

A harder presentation does not make the result more accurate: it makes it stricter, and anyone can miss a plate under time pressure. Read the result as a self-check, and for a graded measurement of fine colour discrimination take the Farnsworth-Munsell 100 Hue test.

Plates: Adaptive; Contrast steps down to your threshold, per cone
Time limit: 3 s; Per plate, no re-look; 2 s and 5 s available
Scoring: Same engine; Identical to the standard test — only the presentation is stricter

How hard mode works

1. Straight into diagnostic plates

The standard test opens with two maximum-contrast warm-up plates that teach the task and are not graded. Hard mode skips them: the first plate you see already counts, and the one everyone-sees-it consistency plate is still slipped in to catch guessing.

2. Three seconds, one look

Each plate is shown for three seconds by default (you can pick two or five). In the standard test a timeout grants a free second look with a fresh figure; here a timeout is a miss. Hesitation therefore costs you, which is exactly what makes it hard.

3. Same ladder, same score

Under the surface nothing changes: an adaptive staircase per cone (red, green, blue-yellow) with best-of-three votes at the level you first miss, per-dot brightness noise so only hue gives the answer away, and the same 0–100 cone scores feeding the same type and severity estimate.

What a hard color blind test can and cannot tell you

Does it catch a mild deficiency? Better than a fixed-plate book: the contrast steps down to your threshold on each cone, so a mild weakness that passes fixed plates shows up here.
Is the type and severity more precise than the standard test? No — the scoring is identical and the screen is the limit. Hard mode is stricter about hesitation, not more precise about your cones.
Can a normal-sighted person fail a plate? Yes, under a 3-second clock with no second look. The best-of-three vote absorbs a single miss, but treat a borderline result as a reason to retake, not a verdict.
Want a number instead of a verdict? Take the Farnsworth-Munsell 100 Hue test — hard for everyone, graded continuously, and scored against published age norms.

This is a free screening and self-assessment tool, not a medical diagnosis. Results are affected by your screen, its color settings, and room lighting.

What is the hardest color blind test?

"Hard" means different things: fine discrimination, threshold sensitivity, or clinical precision. The tests below are the ones that actually push, and why.

Farnsworth-Munsell 100 Hue

Why it is hard: 85 caps in tiny hue steps; even people with normal colour vision make errors, and the score is graded rather than pass/fail.
What it measures: Fine hue discrimination around the whole colour circle

This hard color blind test

Why it is hard: Contrast drops to your threshold on each cone, with a 3-second limit and no second look.
What it measures: Screening plus a type and severity estimate

Cone contrast test (Rabin CCT)

Why it is hard: Letters visible to only one cone class at fading contrast; a clinical instrument with a calibrated display.
What it measures: Cone-isolated contrast thresholds, scored per cone

Anomaloscope

Why it is hard: You must match a red-green mixture to a yellow exactly; the gold standard for red-green type and degree.
What it measures: Precise red-green matching range

CAD test

Why it is hard: A coloured target moving through dynamic luminance noise at threshold; used to set aviation limits.
What it measures: Graded red-green and blue-yellow thresholds

Ishihara plates

Why it is hard: Not hard by design: fixed plates at fixed contrast, built to be a quick, reliable screen.
What it measures: Red-green screening

The hardest color vision tests — the Farnsworth-Munsell 100 Hue, this hard color blind test, the Rabin cone contrast test, the anomaloscope, the CAD test and, for contrast, the Ishihara plates — with why each is hard and what it measures
Test	Why it is hard	What it measures
Farnsworth-Munsell 100 Hue	85 caps in tiny hue steps; even people with normal colour vision make errors, and the score is graded rather than pass/fail.	Fine hue discrimination around the whole colour circle
This hard color blind test	Contrast drops to your threshold on each cone, with a 3-second limit and no second look.	Screening plus a type and severity estimate
Cone contrast test (Rabin CCT)	Letters visible to only one cone class at fading contrast; a clinical instrument with a calibrated display.	Cone-isolated contrast thresholds, scored per cone
Anomaloscope	You must match a red-green mixture to a yellow exactly; the gold standard for red-green type and degree.	Precise red-green matching range
CAD test	A coloured target moving through dynamic luminance noise at threshold; used to set aviation limits.	Graded red-green and blue-yellow thresholds
Ishihara plates	Not hard by design: fixed plates at fixed contrast, built to be a quick, reliable screen.	Red-green screening

What makes a color blind test hard?

Three things. The first is contrast: a plate whose figure and background differ only slightly along one confusion axis is hard for everyone and impossible for someone whose cone on that axis is weak. An adaptive test keeps lowering that difference until you miss, so it always ends at a hard plate. The second is noise: when every dot's brightness is jittered, you cannot find the figure by lightness and must use hue alone. The third is time: a strict limit stops you from scanning, squinting or reasoning your way to the answer.

Most "hard color blind test" pages online are simply darker or more saturated plates, which is not the same thing. Saturation makes a plate easier, not harder, and a plate that is hard because it is dim is testing your screen. The plates here are generated on the cone-confusion lines with luminance masking, which is why a genuinely mild deficiency reveals itself here and can pass a fixed-plate book.

Is a harder test more accurate?

Not automatically. Accuracy is about whether the result matches what a clinical instrument would find, and that is limited by your screen, not by the difficulty setting. A harder presentation raises the chance of a false alarm — a person with typical colour vision who hesitates at a threshold plate under a 3-second clock — and does nothing to remove the errors an uncalibrated display adds.

What a hard run is good for is stress-testing a borderline result. If you passed the standard test with a low-ish cone score, hard mode shows whether that score holds up without the second look. If you want a number rather than a verdict, the Farnsworth-Munsell 100 Hue test is the right instrument: it is hard for everyone and grades discrimination continuously, and its scoring is published with age norms.

Is this a color blind test game?

It plays like one — a countdown, fresh plates every time, a score at the end — and there is nothing wrong with taking it that way. The difference from a colour-matching game is that the plates are built on the science of cone confusion lines rather than on hue puzzles that anyone with practice can learn. Because every plate is generated from a new seed, there is nothing to memorise and nothing to look up, which is also why a retake is a fair retake.

If you would rather sort colours than read plates, the Farnsworth D-15 and FM-100 Hue tests are arrangement tasks, which is what most "colour shade test" and "hue test" games are loosely imitating.

Hard color blind test — frequently asked questions

What is the hardest color blind test?: For fine discrimination, the Farnsworth-Munsell 100 Hue test — 85 caps in small hue steps, scored continuously, with errors even among people with normal colour vision. For detecting a mild deficiency, an adaptive threshold test such as this one, or clinically the anomaloscope and the CAD test, which measure at threshold on calibrated equipment.
How is the hard mode different from the standard color blind test?: Same plates, same adaptive ladder, same scoring. Hard mode removes the two ungraded warm-up plates, sets the time limit to 3 seconds per plate by default (2 or 5 available), and counts a timeout as a miss instead of granting a second look.
I failed a plate — does that mean I am color blind?: No. Under a short time limit anyone can miss a near-threshold plate, and the engine expects it: a miss opens a best-of-three vote at that level, and the result is drawn from where your threshold settles, not from a single plate. Look at the cone scores and the severity estimate on the result card, and treat them as a self-check.
Is a hard color blind test more accurate than the Ishihara test?: It is more sensitive to mild deficiencies, because it tests at your threshold rather than at fixed contrast, and it covers blue-yellow, which Ishihara plates do not. It is not more accurate about the type or severity — an uncalibrated screen limits both — and it is not accepted for licensing or employment, where a clinical test is required.
Can I practise to get better at it?: You can get faster at the task, and hard mode rewards that a little. You cannot learn the plates, because each one is generated fresh, and you cannot change your cone thresholds by practising. If your score improves on a retake, it is usually because you stopped hesitating.

Keep reading

Standard color blind test

The same engine with warm-up plates, a 7-second default and a free second look.

Farnsworth-Munsell 100 Hue

85 caps, a graded Total Error Score and a polar plot of what you confuse.

Farnsworth D-15

Sort 15 colours; the pattern classifies the type in a minute.

Tritan test

The blue-yellow axis on its own, the one plate books leave out.