Behind the glass: designing a puzzle
How a 64-square board gets authored.
8% of male players can't tell ruby from teal. How the palette got rebuilt to work for every eye — without dulling the colors.
The email arrived in 2019 and read like a bug report: 'Your red and your cyan are the same color to me.' The player was deuteranopic — red-green colorblind — and on our board, two of the four palette colors collapsed into each other. He wasn't wrong; we'd never tested for it.
The obvious fix — overlaying shapes or letters on the squares — works, but it's a concession. It tells colorblind players they get a different, uglier game. We wanted the board to look identical for everyone and simply read correctly under every type of color vision.
That meant rebuilding the palette as a contrast problem, not a color problem. Two squares that differ only in hue can be identical in luminance — that's exactly what was happening. The fix: enforce luminance distance alongside hue distance. Every adjacent-color pair in the final palette differs by at least 0.35 in perceived lightness, which survives all three major color-vision deficiencies.
The test that matters isn't a simulation (though we run those) — it's handing the game to colorblind players and watching them not hesitate. That's the check the 2019 email forced us to start doing, and it's now part of every palette decision.
Accessibility isn't a mode you add. It's a constraint you design under — and the design gets better for it.
Hue-only pairs collapse for 8% of players — engineered lightness distance survives every vision type.
No symbol overlays or 'special mode' stigma — one palette that reads correctly for everyone.
Simulators help, but shipping means watching actual colorblind players not hesitate.

UI engineer & accessibility lead
Iker rebuilt Ruby Square's palette system in 2022 and now owns everything accessibility-related. He test-plays every build in grayscale.