Colordle Solver

How the Colordle Solver Works

Colordle has thousands of named colors in its database. When you get 42% back on "Mauve" and 38% on "Dusty Rose," you cannot manually cross-reference which one is closer to the target. The solver does that work for you. It filters the entire list to show only colors that would return those same percentage scores when compared to the same target. It uses Delta E CIE2000 in LAB color space, which is the exact same formula Colordle itself uses to calculate similarity. This means the scores the solver computes match the scores the game shows you, down to the decimal point.

The workflow is straightforward and designed to mirror how you naturally play the game. Make a guess in Colordle, note the percentage the game displays, enter both the color name and the percentage into the solver, and click Filter Results. The candidate list immediately shrinks to show only colors that would produce that same percentage against the same target. Make another guess in the game, add it to the solver, and filter again. After two or three guesses with the solver filtering between each one, you are usually looking at fewer than ten candidates, and often just one or two. Pick the most likely candidate from the filtered list as your next guess in the game, and repeat until you hit 100%.

The color search dropdown lets you type partial names to find the color you guessed. If you type "sea," you will see "Sea Green," "Sea Foam," and any other color containing "sea" in its name. Select the one you guessed, enter the percentage, and filter. If you accidentally select the wrong color, just clear the input and search again. The dropdown searches through the entire color database in real time, so you do not need to know the exact spelling or capitalization of any color name.

The tolerance for percentage matching is set to 0.02 points. This means that if the game shows 85.32% and the solver calculates a theoretical score of 85.31%, the color still appears in the filtered list. This small tolerance accounts for rounding differences in the game's display. If you ever see a candidate listed that seems like it should not be there, check the percentage column — it will be within 0.02 of what you entered.

Understanding Delta E and Color Distance

Delta E CIE2000 is the industry standard for measuring perceptual color difference, and it is the formula Colordle uses to calculate the similarity percentage you see after each guess. A Delta E of 1.0 is roughly the smallest difference the human eye can detect under controlled viewing conditions. A Delta E between 1 and 5 is noticeable if you look carefully. A Delta E between 5 and 10 is clearly different. A Delta E above 10 is a distinctly different color. The percentage Colordle shows is derived from this distance using the formula: percentage = max(0, 100 - DeltaE). So a score of 85% means the Delta E between your guess and the target is approximately 15.

The math happens in LAB color space, not RGB. This is a crucial distinction. RGB hex values are not perceptually uniform. Jumping from #FF0000 to #FF1100 represents a numerical change of 17 in the red channel, but visually this difference is nearly invisible. The same numerical jump in a different color range, say from #005500 to #006600 in the green channel, might be clearly perceptible. LAB color space was designed to fix this problem. In LAB space, a Delta E of 5 means the same visual gap regardless of where on the color wheel you are comparing colors. That is why the solver uses LAB — it produces the same percentages Colordle does, because Colordle uses LAB too.

The CIE2000 variant of Delta E is particularly important because earlier versions (CIE76 and CIE94) had known weaknesses. CIE76 overestimates perceptual differences in saturated colors, making bright reds and deep blues appear further apart than they actually look to the eye. CIE94 improved on this but still struggled with certain hue ranges. CIE2000 added additional correction factors for lightness, chroma, and hue interactions, making it the most accurate perceptual distance formula available. Colordle uses CIE2000, and so does our solver.

When you enter your percentage, the solver inverts the formula to find colors with a matching Delta E distance from your guessed color. It is not approximating or estimating — it is calculating the exact same Delta E value the game would calculate for every color in the database and filtering to show only those that match. This brute-force approach is computationally intensive, but modern browsers handle it effortlessly. The solver typically returns results in under 100 milliseconds even on mid-range hardware.

Tips for Getting Better at Colordle

Start with primaries in separate guesses. If you get 20% on Red and 45% on Blue, the target is hue-adjacent to blue but not close to red. This costs you guesses upfront but tells you the hue family faster than anything else. After two or three primary color guesses, you should know roughly which quadrant of the color wheel the target occupies. From there, you can make more targeted guesses within that quadrant.

Percentage context matters more than most people realize. A 40% match on a muted color means something fundamentally different than 40% on a saturated one. Muted colors like "Ash Grey" and "Silver" are close to many other muted colors in LAB space, so a 40% match does not narrow the field as much as you might hope. Saturated colors like "Crimson" and "Cobalt" are further from most other colors, so a 40% match on a saturated color eliminates more candidates. The solver handles this automatically, but when you are guessing manually, keep in mind that low-saturation colors cluster together more tightly than high-saturation ones.

Learn the obscure color names. Colordle pulls from a large list that includes many names most people have never encountered. Names that consistently trip people up include "Celadon" (a pale green), "Amaranth" (a pinkish red), "Gamboge" (a mustard yellow), "Vermilion" (a bright orange-red), "Saffron" (a deep golden yellow), "Periwinkle" (a light blue-purple), "Chartreuse" (a yellow-green), and "Burgundy" (a dark red-purple). If you do not know these names, you cannot guess them. Spend ten or fifteen minutes reviewing a color name reference chart and it will pay dividends within a week of daily play. The unlimited mode on this site is also a good way to encounter and learn unfamiliar color names in a low-pressure environment.

Pay attention to the hex codes the solver shows alongside each candidate. Over time, you will develop an intuitive sense of which hex ranges correspond to which color families. #FFxxxx is almost always a red-family color. #xxFFxx is green. #xxxxFF is blue. The middle digits tell you about mixing: #FF80xx is orange, #FF00FF is magenta, #00FFFF is cyan. This basic hex literacy speeds up your guessing because you can look at the candidate list and immediately recognize which colors are in the right neighborhood without needing to read every name.

Common Mistakes and How to Avoid Them

The most common mistake is entering the wrong color name in the solver. Colordle's database has many similar names: "Navy" and "Navy Blue" are different entries with different hex values. "Sea Green" and "Seafoam Green" are similarly close in name but distinct in the database. Always verify that the color you selected in the dropdown is the exact same one you guessed in the game. A wrong color name will produce a wrong Delta E, which means the filtered candidate list will be inaccurate.

Another common error is entering the percentage with too much or too little precision. The game displays two decimal places (e.g., 85.32%). Enter the full value including decimals. Do not round to the nearest whole number, because rounding from 85.32% to 85% changes the Delta E tolerance window and may include or exclude candidates incorrectly. The solver's 0.02-point tolerance is calibrated for the two-decimal format the game uses.

Finally, do not forget to click Filter Results after adding each guess. The solver does not automatically re-filter when you add a new guess row. This is by design, because automatic filtering would make it difficult to adjust multiple guesses before running the filter. But it does mean you need to remember to click the button each time. If the candidate list seems unchanged after adding a guess, you probably forgot to filter.