Colour History Analyser: Random or Unusual?
Paste the last results you have seen, as digits, colours or two sides. The analyser sets your counts, streaks and switches beside what random results look like, and says in plain words whether the list is consistent with random or unusual for it.
What the analyser does
Apps of this kind show a history of past results, and many people read it for a pattern. This tool asks one narrower question that has a straight answer: does the list look like random results, or would random results rarely look like this? It counts how often each result came up, finds the longest streak, counts how often the result switched, and tests a popular rule, “bet the opposite after a few in a row”.
Every check ends in one of three plain verdicts: consistent with random, unusual for random, or too few results to say. No list of past results can say what comes next.
How to use it
- Choose what you are pasting: digits 0 to 9, colours (G, R, V) or two sides such as B and S.
- Paste or type the results. Spaces, commas, new lines or nothing between them all work. If your list starts with the newest result, tick the box. In digits mode, big means 5 to 9 and small 0 to 4, the common description of these rounds; if your app shows big and small differently, paste its own B and S results in two-sides mode.
- Read the table of checks and the counts under it, then press the last button to draw 1,000 random lists of the same length and see how many of those look unusual.
The examples under the box were made by this site’s own random-number generator, not copied from any app. Most are plain random draws, picked because they show something: a typical list, a long streak, a short list. Three come from generators built to be biased, so you can see what an unusual result looks like.
A worked example: 100 random digits
The opening example is 100 digits from the generator. These are the five checks, with the figures from the tool itself.
| Check | This list | Random lists | Verdict |
|---|---|---|---|
| Count of each digit | χ² = 10.4 on 9 degrees of freedom | usually below 16.9 | Consistent (p = 0.32) |
| Big against small | 47 big, 53 small | about 50 each | Consistent (p = 0.55) |
| Longest streak | 7 small in a row | usually 7 | Consistent (p = 0.54) |
| Runs | 48 | about 51 | Consistent (p = 0.64) |
| Bet the opposite after 3 in a row | 14 of 24 (58%) | 50% | Consistent (p = 0.54) |
A p-value is the chance that random results would sit at least this far from the average. Below 0.05, one chance in 20, the analyser calls a check unusual. Nothing here comes close.
Seven small results in a row looks striking on a chart, yet 100 random results hold a streak of 7 or more on one side or the other 54% of the time. The popular rule won 14 of its 24 bets, 58%, which is not 50% and is still consistent with random: for 24 bets, random lists usually land between 29% and 71%.
When a list looks unusual
Three examples are built to be odd:
- Lopsided digits. The digit 7 is made three times as likely as the others. It turns up 25 times in 120, where 12 are expected: χ² = 26.5, p = 0.002. Only the counts check flags it.
- A generator that likes to repeat. It repeats the last result 72% of the time. The list has 30 runs where about 41 are expected, p = 0.018. Only the runs check flags it.
- Results that swap almost every round. 48 runs where about 31 are expected, p below 0.001. “Bet the opposite after 2” won 11 of 12 bets, p = 0.006, but a rule can only look good on a list whose results already swap. It cannot make the next ones swap.
Unusual for random means random results would rarely look like this. It does not say why, or that anything is wrong with an app. Odd on one check is also common: about one random list in six is flagged by at least one of five checks. The sensible next step is more results, not a conclusion.
What it cannot tell you
- The next result. A past list cannot say what comes next, and a pattern in it is not a signal.
- How an app works. We cannot see inside any app. A list can be consistent with random and come from something else, or be unusual and come from a random source.
- Anything from a very short list. Under about 50 digits the counts check says it has too few results, and the rule needs 10 bets.
- Your app’s own shares. Digits and two sides are judged against equal chances. For colours, 40% green, 40% red and 20% violet is an illustration: use the shares your own app states.
- More than the standard tests give. The streak and runs checks are exact. The counts check uses the usual chi-square approximation, and the rule check treats overlapping bets as separate, so both are close rather than exact.
Where to go next
The streak odds calculator turns a streak you have seen into a chance. The prediction simulator shows long streaks arising in random rounds, and colour tricks and formulas goes through why chart patterns do not predict. For how a round is played, read the Win Go guide, and see the rest of the free tools.
Common questions
Can this analyser tell me what the next result will be?
No. Nothing in a list of past results can say what the next result will be, because each round is a fresh draw. The analyser only compares the list you paste with what random results look like and tells you how unusual it is. The next result is not part of that.
What does “consistent with random” mean in the results?
It means random results would often look like your list, or stranger. It is a statement about the pasted list only. It says nothing about how any app makes its results, because we cannot see inside any app.
Why does a random list sometimes look odd?
Random does not mean evenly spread. Streaks, lopsided counts and neat-looking patterns are normal in random lists, and the more checks you run, the likelier one of them looks odd. In 1,000 random lists of 100 digits, about one in six is flagged by at least one of the five checks.
How many results do I need to paste?
At least 10 for any check. The digit counts need about 50 digits so that every digit is expected at least five times, and the opposite-after rule needs 10 bets. A longer list lets each check notice smaller differences, so paste as many as you have.
Does it matter if my list starts with the newest result?
Counts, streaks and runs come out the same either way. The popular rule looks at which result came next, so it needs the true order: tick the box if your list starts with the newest result, as most apps show it.
Does the analyser send what I paste anywhere?
No. The list is read and checked inside this page, on your own device. If you press the copy-link button, the list goes into the link itself, after the # sign, which a website never receives.