Anyone can call their own test honest. What makes it checkable is answering the hard questions in one place, in public, with sources attached. That's this page. If an answer here stops being true, that's a defect — tell us.

Question 1

WHO MADE IT?

One person: Vince Gonzalez, an independent researcher in Punta Gorda, Florida, United States.

Not a university department. Not a hospital. Not a company that sells eyewear. There is no team behind this, no institutional review board, and no outside funding. That cuts both ways and you should weigh it: there is no commercial incentive bending the results, and equally there is no institution whose reputation is staked on them.

Question 2

WHAT EXACTLY IS BEING MEASURED?

Each test states its own measure on its own page, rather than collapsing everything into one vague score.

  • Color testsCIEDE2000 difference between colors before and after a Machado, Oliveira & Fernandes (2009) simulation of color-vision deficiency.
  • AcuityThe smallest optotype correctly identified, at a viewing distance you set.
  • Stereo depthThe smallest horizontal disparity, in screen pixels, at which a random-dot shape was located.
  • Flicker fusionThe frequency, in Hz, at which flicker stops being visible to you on your display.

Where a quantity can't be honestly derived, the test reports the raw measure and stops. The stereo test is the clearest case: it refuses to convert pixels into arc seconds, because doing that needs your pixel pitch and viewing distance, and a web page knows neither.

Question 3

WHAT DOES IT ASSUME ABOUT YOUR SCREEN?

It assumes an sRGB display — and nothing else, because a browser cannot measure the things that actually matter here.

It does not know your display's real gamut, gamma, brightness or white point. It does not know your ambient lighting, your viewing distance, or your pixel pitch. Every one of those changes the result. Screen saturation is bounded by the sRGB gamut, so any test that depends on very faint or very saturated color is limited by the display itself, not by the method.

Where the error lives Display calibration is the dominant error term here, and it applies to every test on the site. No browser can remove it — a clinic controls illumination, viewing distance and a physical target; a web page controls none of the three. That is exactly why every test reports its raw measure and names its assumptions instead of manufacturing a precision it hasn't earned.

Question 4

IS THE METHOD PUBLISHED?

Yes, open access, with DOIs you can check independently of anything said on this site:

The color work rests on Machado, Oliveira & Fernandes (2009) for deficiency simulation and CIEDE2000 for color difference — standard published methods, not a private in-house formula you're asked to take on trust. Read the research →

Question 5

CAN YOU READ THE CODE?

Yes. The site and the engine are MIT licensed and public, and the plate generator is deposited alongside the paper.

The same engine is also published as installable packages on PyPI and npm, so you don't have to trust our numbers — you can run the math yourself and compare. If your output disagrees with ours, that's a bug worth reporting. The packages & API →

Question 6

HAS IT BEEN CLINICALLY VALIDATED?

Not clinically validated — and here is what has been The underlying methods are published and standard. This browser implementation, running on your uncalibrated screen, has not been through an independent clinical validation study against calibrated instruments. Said plainly, because you should be able to weigh it.

What is established instead is every layer underneath: the method is published with a DOI, the colorimetry is standard rather than proprietary, the code is open and installable so you can reproduce the numbers, and each test states its own measure. That is a stack you can check line by line — which is more than most vision tests offer.

It still means no result here equals a Titmus, TNO or Randot stereo test, a physical Ishihara plate book under standard illumination, or a clinic acuity measurement. Anyone who tells you an online test is "as accurate as" those is selling something.

Independent review that has happened, stated precisely: the extension passed review at four separate stores — Mozilla's Firefox Add-ons, the Chrome Web Store, Microsoft Edge Add-ons and the VS Code Marketplace. Those processes vet security, privacy, permission scope and readable source; Mozilla's is the strictest of them. They are real checks on the software, and none of them is a review of the color science — every link is in the footer if you want to confirm them yourself.

Question 7

WHAT CAN IT NOT MEASURE?

Eye health, in every sense. Specifically, it cannot:

  • PressureMeasure intraocular pressure or screen for glaucoma.
  • StructureExamine the retina or optic nerve, or detect cataract or macular disease.
  • RefractionProduce a spectacle or contact-lens prescription.
  • DiagnosisClinically diagnose color-vision deficiency, or its exact type and severity.
  • StereoacuityGive a stereo threshold in arc seconds.

Everything here is a screening and educational aid. A flag on this site is a reason to see an optometrist or ophthalmologist — never a conclusion, and never a substitute for a full exam. For children this matters more, not less: untreated amblyopia is time-sensitive.

Question 8

WHO PAYS FOR IT?

Voluntary Ko-fi contributions, and commissioned color-accessibility work, which is handled through F-Keys Creative, the studio behind this site. Everything on OpticQuiz itself is free, with no paywall and no account.

What matters more is what is not here. OpticQuiz does not sell eyewear, lenses, color-correcting glasses, or a refraction service, and takes no affiliate commission on any of them.

Why that question is the sharp one A test that is free because it's a funnel into a $400 pair of glasses has a reason to tell you something is wrong with your eyes. Ask that question of every vision test you take, including this one. Here, no result can steer you toward a product we profit from, because there is no such product.