A dashboard with six series using red, orange, green, teal, blue, and purple looks great to a designer with typical vision. To a user with deuteranopia, the red and green series are nearly identical. The orange and teal also compress. Suddenly four of six series are ambiguous, and the chart communicates nothing.
The fix is not to avoid color. It is to choose palettes where adjacent hues differ in lightness as well as hue, so the distinction survives even when hue perception collapses.
Three rules that cover most cases.
First, never rely on red-green contrast alone. If two series must be adjacent in the legend, make one light and one dark. A pale orange next to a deep blue is distinguishable under every common deficiency type.
Second, test with a simulator before publishing. The Color Oracle desktop app and the Coblis web simulator both render your chart as seen by protanopia, deuteranopia, and tritanopia viewers. Thirty seconds of checking saves a support email.
Third, add a secondary encoding for critical distinctions. Dash versus solid line, circle versus square marker, or a direct label on the line. Color should reinforce the encoding, not be the only carrier of meaning.
Palette starting points.
For categorical data with up to five series, the Okabe-Ito palette (orange, sky blue, bluish green, yellow, blue) is validated across all three deficiency types. For sequential data, a single-hue ramp from light to dark avoids the problem entirely. For diverging data, purple-to-orange through white maintains lightness symmetry.
DataVizForge chart themes now include an Okabe-Ito categorical option in the palette picker. Feedback: sapsap@qq.com.