Why 'It Looks Fine to Me' Isn't Good Enough for Text Color

"It looks fine to me" is the most common reason low-contrast text ships. It's also the least reliable test available, because the person saying it is looking at a calibrated monitor, in good light, with eyesight that isn't the eyesight every visitor has. A light gray paragraph on a white card can be genuinely, provably hard to read for a meaningful share of people looking at the same pixels - not a matter of taste, but a measurable ratio between two colors that either clears a real threshold or doesn't.

Why "it looks fine to me" isn't a test

Screens vary. Brightness settings vary. Ambient light varies - the same page that's readable at a desk under office lighting can be nearly invisible on a phone outdoors in direct sun. Eyesight varies too, and not only for people who identify as having low vision: contrast sensitivity declines gradually with age for almost everyone, long before anyone would call it a disability. A designer checking their own work on their own screen is testing exactly one combination of monitor, lighting, and eyesight out of a very wide range their actual visitors have.

The Web Content Accessibility Guidelines exist because "looks fine to me" doesn't scale past one person's setup. WCAG 2.x defines contrast as a precise, calculable ratio between the relative luminance of two colors - a number from 1:1 (identical, invisible) to 21:1 (pure black on pure white) - and sets minimum ratios for text to be considered legible at two levels: AA, the commonly required baseline, and AAA, a stricter level most sites don't fully reach but can aim for where it costs nothing.

What the actual thresholds are

Four numbers cover the whole spec: normal text needs at least 4.5:1 at AA and 7:1 at AAA. Large text - defined by WCAG as 18pt (24px) regular weight or 14pt (about 19px) bold - needs only 3:1 at AA and 4.5:1 at AAA, because bigger, bolder strokes are inherently easier to distinguish from a background at a lower contrast ratio. None of these numbers are arbitrary; they come from research on what a person with moderately reduced vision can actually resolve.

Who this actually affects

Low vision and color vision deficiencies are the obvious groups, but they're not the only ones. Anyone using a phone outdoors, anyone on an older or cheaper screen with a narrower color range, anyone reading for long stretches where low contrast causes real eye strain even with normal vision - all of them benefit from the same fix. And for a business, there's a second, separate reason this isn't optional: contrast requirements show up directly in accessibility law and in procurement checklists agencies and larger clients run against vendor sites, so a real ratio matters even to someone who's never thought about the guidelines by name.

Contrast is a relationship, not a property of one color

It's easy to think of a hex value as "light" or "dark" on its own, but the WCAG ratio only exists between two colors - a foreground and whatever sits behind it. The same mid-gray that reads as comfortably dark against white can read as uncomfortably light against black, and swapping either color changes the number even if the other one never moves. That's why a swatch shown alone, without its actual background, can't really be judged for legibility at all - the badge only means something once both colors in the actual pairing are known.

This is also why the formula weights red, green, and blue unevenly - 21.26% red, 71.52% green, 7.22% blue - rather than averaging them. Human vision is far more sensitive to green light than to blue, so two colors with identical average brightness can look meaningfully different in perceived lightness if their green content differs. A calculation that just averaged the three channels would call colors "equally bright" that most people would describe as clearly different, which is exactly the gap between eyeballing and measuring this whole guide is about.

Why guessing is worse than it seems

Contrast ratio isn't intuitive to estimate by eye, because human perception of brightness isn't linear the way the underlying math is - two colors that look roughly similar in "darkness" can differ by a factor of several in actual relative luminance, and two colors that look quite different can sit closer together than expected. That gap between perception and the real number is exactly why a calculated ratio, not a glance, is the only reliable check.

This site's contrast checker runs the WCAG formula directly - pick or type two colors and the ratio and all four pass/fail badges update immediately, calculated entirely in your browser. No account, no upload, nothing more than the math the spec itself defines.