We tested every 4D myth against 41 years of real draws. Here's the scoreboard.

Every 4D player carries a theory. Pattern numbers like 8888 come out less. Chinese New Year draws are different. Magnum's machine has habits. The number you dreamed about is due. This August we stopped nodding along and tested all of them — against every prize published in Malaysia since 1985, over 612,000 slots across 41 years. Every test in this article runs live on this site, recomputed nightly, so you can check the working.

How we kept ourselves honest

Run enough tests and some will look significant by luck alone — that's how fake discoveries are manufactured. So every batch uses a Bonferroni-corrected threshold (flag only if p is smaller than 0.05 divided by the number of tests), and we publish effect sizes next to every p-value, because with 612,000 data points a statistically “significant” difference can still be too small to matter. The full method is on the myths page.

Myth 1: pattern numbers avoid the draw

Players believe both directions — that 8888 never comes out, and that sequences like 1234 are secretly common. Combinatorics says exactly 10 quadruple numbers, 7 ascending runs and 90 palindromes of each shape exist among the 10,000, so we know precisely how often each family should appear.

Palindromes: 5,609 observed against 5,507 expected. Ascending runs: 418 against 428. The overall test lands at p = 0.48 — the draw treats 8888 exactly like it treats 4715. The myth fails in both directions at once.

Myth 2: some days are luckier

Month of year, the Chinese New Year window, the Raya window, month-end — eight tests, eight clean nulls, with the largest deviation anywhere a rounding-error 0.79 percentage points. But here's the trap that makes this myth feel true: the CNY window really does contain 8% more draws per date, because the special-draw calendar clusters there. Count draws instead of normalising per prize slot and you'd “discover” a CNY effect that is only the calendar. Details on the time-patterns page.

Myth 3: each operator's machine has its own habits

Magnum vs Toto vs Da Ma Cai: the digit distributions differ by at most 0.25 percentage points from each other — statistically indistinguishable (p = 0.016, above the corrected threshold). Whatever machine superstition you hold about the big three, the numbers don't support it.

Myth 4: dream numbers win

We scored every dream symbol's numbers from the Magnum, Toto and Da Ma Cai 万字图 dictionaries against all 612,000 prizes — with a baseline matched to each number's digit shape, so the comparison is fair.

Only 2 of 67 symbols even clear an uncorrected p<0.05, where chance alone predicts 3.35 — there are fewer lucky dreams than luck would produce. And the test had the power to find a real edge: anything above ~5.5% could not have hidden. The tradition is a lovely one; the archive just doesn't pay it. Full table on the dream test page.

Myth 5: streaks and eras

The hottest number of the 1990s, 4317, hit 23 times — but a fair draw's expected best over that many draws is 24.1. The 2000s champion underperformed chance too. Decade “champions” are what a maximum over thousands of numbers looks like, not a property of the number. Same story for streaks today: every number currently “in form” is on exactly the streak length coincidence predicts.

The one thing that did flag

Honesty cuts both ways: one test refused to pass. The digit distributions of the six Malaysian-licensed operators are clean — but two offshore daily operators fail uniformity tests hard, a finding we cross-verified against the operators' own published archives (128 of 128 and 123 of 123 sampled dates match) and reproduced with an independently collected sample. That story, with every caveat it deserves, lives on the randomness audit.