Color Mixing Game rules are surprisingly rigid: five fixed targets, slider controls that change meaning between rounds, a Delta E 1976 acceptance threshold of 3.0, a strict two-distinct-miss lock, and an exact 1,000-point perfect score. Each round shows more than two color patches — the target hex value, the target's CIELAB coordinates, the active model name, the current mix's hex, and the current mix's RGB channels are all visible at once, so two visually similar colors can still be distinguished by their numbers. Players move three sliders: red, green, and blue from 0 to 255 in additive rounds, or cyan, magenta, and yellow from 0% to 100% in idealized subtractive rounds. Pressing Check mix scores the current controls. A correct mix immediately adds 200 points and loads the next fixed target; five correct rounds therefore yield an exact 1,000. The run lives entirely in the browser, with no account, no upload, and a deterministic Restart that returns every variable to its starting state.
The Rule Set at a Glance
Before starting a run, it helps to see the entire rulebook on one screen. The table below collects every numeric constant and the part of the game it governs, so you can return to a single reference whenever a rule seems unclear mid-round rather than guess at the numbers.
| Rule | What It Governs | Value |
|---|---|---|
| Round count | Targets per game | 5 |
| RGB slider range | Additive rounds | 0–255 per channel |
| CMY slider range | Idealized subtractive rounds | 0%–100% per channel |
| Acceptance threshold | Maximum colour difference | Delta E 1976 ≤ 3.0 |
| Points per correct mix | Round scoring | 200 |
| Perfect score | Five correct rounds | 1,000 |
| Distinct misses per round | Wrong-mix budget | 2 |
| Repeat wrong-mix | Duplicate handling | Deduplicated, does not consume the second chance |
| Locked round state | After two distinct misses | Sliders and Check mix disabled |
| Restart recovery | Reset to deterministic start | First target, zeroed sliders, zero score, empty signature set |
How to Play a Round
The control surface is small, but each round has a precise order of operations. These steps describe a single round in Color Mixing Game; the same procedure repeats for all five targets, switching the slider semantics between RGB and CMY as the active model changes.
- Read the round banner. Note the active model (RGB or idealized CMY), the target hex value printed above the target swatch, and the target's CIELAB coordinates printed below the swatch.
- Compare the current mix to the target using the printed numbers — the target hex, the live mix hex, and the current RGB channels — rather than trusting visual judgement alone, which can be thrown off by uncalibrated displays.
- Adjust the three sliders with touch, mouse, or arrow keys. Use Tab to focus a slider, then nudge it by one unit with the arrow keys. Touch inputs are sized at least 44 pixels tall for finger play.
- Press Check mix. The current controls are evaluated against the target through the full sRGB to CIELAB pipeline and scored with Delta E 1976.
- If the result is a match, you collect 200 points and the next fixed target appears. If not, change at least one slider value to create a different signature, because identical wrong submissions are deduplicated and do not consume the second chance.
- If the round locks after two distinct wrong submissions, the only recovery is Restart, which resets the entire run to the deterministic starting state.
Scoring and the Two-Miss Lock
Scoring is simple on the surface: each correct mix returns exactly 200 points, and the run caps at 1,000 because there are exactly five fixed targets. The interesting rule sits underneath, in how the game counts wrong answers. A wrong mix is recorded as a signature — the set of three slider values submitted at that moment. The first unique wrong signature leaves one chance. If the next submission uses the identical triple, the game deduplicates it and the second chance remains available; a different signature is what actually consumes the second chance and locks the round.
Once a round locks, all three sliders and the Check button become disabled, so the attempt history, the slider values, and the round score cannot drift while you consider what to do next. The single recovery path is Restart. Pressing Restart is a deterministic full reset: the run returns to the same first target, all three sliders reset to their starting position, the score returns to zero, and the signature set is empty again. The first target after Restart is not chosen at random — it is a fixed fixture, which means a consistent reset is part of the rule set, not a side effect of the browser's memory.
Players who want a target-by-target walkthrough that follows the same five rounds under these rules can read the five-target RGB and CMY walkthrough guide.
What Counts as Close Enough (Delta E 1976)
Visual judgement is not what scores a mix. The game decodes each encoded sRGB channel with the standard piecewise transfer function (the gamma curve that turns stored sRGB bytes into linear display light), multiplies the resulting linear triplet by the published sRGB-to-XYZ matrix for the D65 illuminant, normalizes the XYZ triple against the D65 white point (x = 0.3127, y = 0.3290), and then transforms that result into CIELAB using the published epsilon (216/24389) and kappa (24389/27) constants. The conversion chain follows the algorithm described in the W3C CSS Color 4 conversion code, and the underlying sRGB definition, the 2.4 transfer exponent, and the D65 chromaticity are cross-checked against the International Color Consortium sRGB registry.
Delta E 1976 — the Euclidean distance between the two Lab triplets — is then computed as the square root of the summed squared differences in L, a, and b. The acceptance rule is blunt: a mix counts as correct when Delta E 1976 is 3.0 or lower. Anything above 3.0 is a miss. RGB channel distance is never used in place of Delta E, so two sRGB values that look close on a screen may still be a miss if their Lab coordinates are far apart, and the reverse can also be true. This is also why the game prints hex values and Lab numbers on every patch: two colours that look similar to one player on an uncalibrated display may have a numeric Lab distance that clearly crosses the threshold.
RGB Rounds vs. Idealized CMY Rounds
The five-target run mixes two completely different teaching models, and the rules treat them on purpose. The table below summarises how each mode controls the swatch and where the two models differ.
| Aspect | RGB Round | Idealized CMY Round |
|---|---|---|
| Slider count | 3 | 3 |
| Range per slider | 0–255 | 0%–100% |
| Channel semantics | Red, green, blue display channels | Cyan, magenta, yellow teaching channels |
| Mix effect | Raising a slider adds that display-light channel | Raising a slider removes its complementary display channel |
| Physics fidelity | Models additive screen light | Explicitly idealized; not a prediction of paint, ink, paper, binders, or pigment chemistry |
| Numerical output | sRGB primaries applied directly to the swatch | Cyan removes red, magenta removes green, yellow removes blue, then channels are rounded to integer sRGB |
The five targets — Signal Orange, Ideal Green, Electric Cyan, Ideal Violet, and Screen Magenta — are split between these two modes, and their exact slider solutions are fixed, inspectable fixtures rather than generated guesses. Because the CMY calculation is a deliberately simplified teaching rule rather than a spectral simulation, the same slider values that hit one target in CMY mode will not behave like real pigment mixtures on paper or canvas; physical mixtures depend on spectral properties that this browser model does not represent.
Edge Cases: Restart, Invalid Inputs, and Privacy
A handful of behaviours sit underneath the obvious rule set and are worth knowing before you start a run, because they affect whether the game can complete the task.
- Restart is deterministic. The first target, starting slider values, score, and signature set all return to their initial state on every Restart, so the same inputs always lead to the same outcome across resets on the same device.
- Invalid inputs are ignored atomically. Fractional slider values, NaN, values outside the active mode range, missing channels, and invalid arrays are no-ops. They do not throw errors, do not advance the round, and do not record a miss signature.
- Locked and completed states are stable. Once a round locks (two distinct misses) or completes (a correct mix), redundant inputs are no-ops — neither the score nor the attempt history can drift while you decide what to do next.
- The run stays in your browser. Slider values, attempts, colour conversions, and the score are processed locally; nothing is uploaded, and no account is required. A separate best-score is held on this device only.
With those mechanics in mind, the active rulebook for any run is short enough to memorize: read the banner, adjust three sliders, press Check, collect 200 points for a Delta E under 3.0, and avoid two distinct misses — with Restart always available as a clean, full reset whenever a round locks.
Related reading: Color Sudoku Rules: Five Boards, One Solution Each.