Is the draw actually random? We test it.

Chi-square uniformity on every digit position, runs tests and serial correlation on every operator's full draw history — recomputed nightly, p-values published. No other Malaysian results site runs this audit.

Something is worth a second look — see the highlighted rows. 11 operators, 110 statistical tests. The flag threshold is p < 4.5e-4 (Bonferroni-adjusted) — strict on purpose, because running this many tests guarantees a few small p-values by pure chance.
What a flag means — and doesn't. A flagged row says the stored draw history fails a uniformity test, which could be a data-collection artifact or a property of the draws themselves — so we verified the data first. Cross-source check (17 Aug 2026): every sampled flagged-operator date matched the operator's own official archive exactly (128 GD dates, 123 STC dates, 100%), and an independent 593-draw collection taken straight from the official archive reproduces the same digit skew. The non-uniformity is therefore in the published draws, not in our pipeline. That is still not an accusation — non-uniform published numbers can have innocent mechanical explanations — but it is a measured, reproducible fact. The most-regulated operators pass every test through the exact same pipeline.
OperatorDrawsDigit positions 1–4 (p)Runs (p)Serial (p)Last 365d (min p)
Sandakan 4D⚠️1,5451.0e-400.1665.7e-114.0e-100.0040.0715.6e-7
Grand Dragon⚠️2,2951.8e-331.6e-112.8e-208.5e-200.4270.3560.001
Lucky Hari Hari 7:30PM⚠️2,9975.2e-305.0e-97.0e-141.4e-240.6220.2091.5e-9
Lucky Hari Hari 3:30PM⚠️1,3905.1e-150.0512.4e-102.0e-190.7270.5127.4e-12
Sabah 88 4D1,4090.1525.5e-40.3430.9460.2200.5400.070
SportsToto 4D5,6980.0120.6280.4080.9090.2080.3410.322
Special CashSweep1,5500.3210.8150.4740.7930.5250.6220.014
9 Lotto600.0300.8620.7290.0290.8970.9260.029
Da Ma Cai 1+3D5,9990.1430.4230.4350.4090.1120.2140.037
Magnum 4D6,8280.4570.6980.6370.1040.0370.4080.364
Singapore 4D1,4910.7920.4630.6460.7540.6410.8770.182

Position columns: chi-square uniformity of digits 0–9 at each of the four positions over all top-3 prizes. Runs: Wald–Wolfowitz on the 1st-prize sequence. Serial: lag-1 correlation of consecutive 1st prizes. Sorted worst-p first.

How to read the p-values

A p-value answers: if the draw were perfectly fair, how often would we see a pattern at least this extreme by luck alone? Small p = surprising under fairness. It is NOT the probability the draw is rigged.

We run 110 tests, so some small p-values are guaranteed by chance — about 1 in 20 tests lands under 0.05 even on perfect data. That's why the flag line sits at p < 4.5e-4, not 0.05: it corrects for how many chances we gave luck.

What would failure look like? A digit position with a persistently tiny p across recomputes, or a runs/serial failure holding up in the recent window. One-off borderline values are what randomness normally produces — and 'everything passes' is the expected, publishable result.

Year by year: every operator's digits, tested

Chi-square randomness test · 1985–2026 · 151 operator-years checked

19 operator-year(s) stand out and are listed below for review.

This isn't proof of anything either way — it's an ongoing statistical check. A truly fair machine should still occasionally produce an unusual year by chance, and a real bias would show up as p-values repeatedly crashing toward zero.

How this works

For each operator and calendar year we pool the four digit positions of every winning number and run a chi-square goodness-of-fit test against a perfectly uniform distribution (each digit 0–9 equally likely). A high p-value means the digits look random; a very low one means they don't.

Because we run hundreds of tests, roughly 1 in 20 will show p < 0.05 by pure chance — so a low p-value alone means nothing. We apply a Benjamini-Hochberg false-discovery-rate correction across the whole family of tests and only flag (red) the windows that survive it.

Magnum 4D12 flagged
1985p = 0.052026
Da Ma Cai1 flagged
1990p = 0.052026
Sports Totoconsistent with fair
1992p = 0.052026
CashSweepconsistent with fair
2017p = 0.052026
Sabah 88consistent with fair
2017p = 0.052026
STC 4D3 flagged
2017p = 0.052026
Grand Dragon3 flagged
2020p = 0.052026

Windows flagged for review

Survived the FDR correction — worth a closer look (and a check that it isn't a data-capture error on our side).

OperatorYearχ²p-valueDigits tested
Magnum 4D198652.8<0.000114,536
Magnum 4D198744.7<0.000114,608
Magnum 4D198862.8<0.000114,532
Magnum 4D198952.4<0.000114,532
Magnum 4D199055.8<0.000114,712
Magnum 4D199166.3<0.000114,444
Magnum 4D199262.3<0.000114,428
Da Ma Cai199326.10.002014,336
Magnum 4D199331.60.000214,348
Grand Dragon202140.2<0.000133,560
Magnum 4D202123.60.005011,592
Magnum 4D202231.90.000216,460
Grand Dragon202328.30.000933,568
Magnum 4D202437.8<0.000115,080
STC 4D202426.10.002014,996
Grand Dragon202538.5<0.000133,564
Magnum 4D202528.40.000815,180
STC 4D202528.00.000915,180
STC 4D202664.1<0.000111,776

Independent analysis for reference. OngLai4D is not affiliated with any operator. Numbers are drawn from our archive of published results.

Year-by-year check ↓Method leaderboard