To guess RGB from color, estimate the red, green, and blue channels from 0 to 255 and enter either three decimal integers or a CSS #RGB/#RRGGBB value; the guess passes only when every absolute channel error is at most 12 and the three errors total no more than 24. The RGB Guessing Game applies that rule to five mystery sRGB swatches. It is a visual recognition challenge, not an eyedropper for arbitrary images: no image file is uploaded, and the game does not reveal an exact code for the displayed color.
Build a channel-by-channel estimate
RGB represents a color with three channel values. On the 0-to-255 scale, 0 means no contribution from that channel and 255 means its maximum contribution. A practical estimate becomes easier when you separate the judgment into brightness, dominant channel, and balance rather than trying to name the swatch first.
- Estimate brightness: A very dark swatch calls for lower values across all three channels, while a very light swatch calls for higher values.
- Find the dominant channel: If the color looks strongly red, green, or blue, give the corresponding channel a larger value than the other two.
- Balance the remaining channels: If the swatch looks neutral, start with three similar values rather than forcing a strong tint.
- Preserve the visible relationships: If two channels look close and one stands apart, reflect that relationship in the numbers instead of estimating all three independently.
These are starting estimates, not an automated readout. The strongest improvement comes from comparing your numbers with the game’s error feedback and deciding which channel your visual judgment handled least accurately.
Make a guess for each swatch
- Study the active Canvas swatch. Focus on its apparent lightness and the relative strength of its red, green, and blue channels. Do not expect the surrounding interface to expose the target value.
- Choose an input mode. Select the native RGB or HEX format button. RGB is convenient when you are estimating each channel separately; HEX is useful when you can reason more confidently in hexadecimal pairs.
- Enter three channel values. In RGB mode, provide three base-10 whole numbers from 0 through 255. In HEX mode, enter a CSS value using either #RRGGBB or #RGB, including the leading hash. For example, #ABC expands to #AABBCC.
- Submit the estimate. The game converts either notation into the same red, green, and blue integer tuple before comparing it with the current swatch.
- Check both pass conditions. Read the three absolute channel errors, the largest channel error, and the total error. The estimate is accepted only when the largest error is no more than 12 and the total is no more than 24.
- Use the repair attempt after a miss. The first distinct wrong normalized color leaves one repair attempt. Repeating the same normalized color does not consume that second chance, even if you change from RGB to equivalent HEX. A second different wrong color locks the round until you choose Restart.
- Continue through all five fixed rounds. Each accepted swatch opens the next round and awards exactly 200 points. A complete clean run therefore earns 5 × 200 = 1,000 points.
Choose the easier input format
RGB and HEX are not separate scoring systems in this game. Both formats are normalized to one RGB tuple, so an RGB entry and an equivalent HEX entry receive the same comparison and share the same miss signature. MDN’s RGB value reference describes the 0-to-255 integer channel model, while MDN’s hex-color reference covers the supported three- and six-digit CSS forms.
| Input mode | Accepted entry | Normalization and behavior |
|---|---|---|
| Decimal RGB | Three whole numbers from 0 through 255 | The numbers become the red, green, and blue tuple used for scoring. |
| Six-digit HEX | #RRGGBB with the leading hash | Each hexadecimal pair becomes one channel value; letter case does not matter. |
| Short HEX | #RGB with the leading hash | Each digit is duplicated, so #ABC becomes #AABBCC before conversion. |
Choose the notation that reduces your chance of making a transcription error. If your visual estimate naturally produces three channel levels, RGB is usually direct. If you are working from familiar hexadecimal color values, the HEX field accepts both short and full forms without changing the underlying thresholds.
Read the channel errors
For each channel, the game calculates the absolute difference between the target and your guess. A guess must satisfy both inclusive limits: max(channel errors) ≤ 12 and sum(channel errors) ≤ 24. Passing only one condition is not enough.
Consider a hypothetical target of (72, 135, 210) and a guess of (80, 130, 205). The red error is |72 − 80| = 8, the green error is |135 − 130| = 5, and the blue error is |210 − 205| = 5. The largest error is 8 and the total is 8 + 5 + 5 = 18. Because 8 ≤ 12 and 18 ≤ 24, the guess passes.
| Result | Maximum channel error | Total error | Game result |
|---|---|---|---|
| Accepted | No more than 12 | No more than 24 | Award 200 points and open the next fixed round. |
| Rejected by the channel rule | More than 12 | Any total | The guess does not pass, even if two channels are exact. |
| Rejected by the total rule | No more than 12 | More than 24 | The guess does not pass, even though every individual channel is within 12. |
The two limits work together deliberately. The per-channel limit prevents one large mistake from being hidden by two exact channels, while the total limit rejects an estimate whose errors are spread too widely across all three channels.
Understand the Canvas limitation
Normal answer feedback reports all three absolute errors, the largest error, and the total error. It does not disclose the exact target tuple. Because the values are absolute rather than signed, an error of 8 shows the distance from the target but not whether the target’s channel is numerically above or below your guess.
The active swatch is drawn into a Canvas rather than published through normal page text, ARIA labels, data attributes, or inline styles. Future round values are also not rendered into the current DOM. This prevents accidental answer disclosure, but it is not a tamper-proof or anti-cheat boundary: a determined user can sample displayed pixels or inspect downloaded code. The game therefore remains a visual estimation exercise rather than a secure secret-color system.
Keep every submission valid
- Use complete RGB input: Supply exactly three base-10 whole numbers in the range 0 through 255. Fractions, negative values, missing channels, and values above 255 are invalid.
- Include the HEX hash: Enter #RGB or #RRGGBB exactly in the supported CSS notation. Hexadecimal letters are case-insensitive, but the leading hash is required.
- Do not spend a chance on malformed input: Invalid syntax and out-of-range values do not change the scored state or consume the repair attempt.
- Recognize terminal states: Completed and deadlocked rounds do not change further through ordinary actions. Focus moves to the terminal status message, and Restart is used to begin again after a deadlock.
- Check for the shared Boss overlay: While it is visible, input edits, mode changes, submission, and Restart are ignored without changing local or scored state.
- Use either input method: Tab moves between the format buttons, fields, and submit action, and Enter from the form checks an estimate. Touch controls retain mobile-friendly target heights, and the layout becomes one column on a narrow screen.
Score a complete run
A perfect result requires all five swatches to pass the dual error threshold. At 200 points per accepted swatch, the final score is exactly 1,000. Gameplay and best-score handling remain in the browser, with no account, upload, model call, remote color service, or new dependency. The result is a local five-round color-recognition challenge with explicit numerical feedback rather than an automatic color-identification utility.