What this page shows
The tables above record how many times each ball has been drawn in three of the draws South Africans bet on most through Lucky Numbers: UK49s Lunchtime, UK49s Teatime and the SA Daily Lotto. They are a tally of what has already happened — nothing more. They are not a prediction tool, not a betting system, and not a set of numbers to play.
We publish them because the figures are widely searched and widely misread. The most useful thing draw-frequency data actually proves is the opposite of what most "hot numbers" pages claim: over a long enough run of draws, every number lands at almost exactly the same rate. The small gaps you see between the hottest and coldest balls are what random chance looks like over thousands of draws, not evidence that any ball is "lucky".
How to read the frequency tables
Each draw has two short tables: the most-drawn balls and the least-drawn balls over the window shown beneath the draw name. The window is the full published history for each draw — for the two UK49s draws that is roughly ten thousand draws each, and for the SA Daily Lotto it is every draw since it launched in March 2019.
Across that many draws the spread is tiny. In UK49s Lunchtime the most-drawn ball (47) has appeared about 1,376 times and the least-drawn (8) about 1,190 — a difference of well under one in seven over nearly thirty years. In a draw where six balls are pulled from forty-nine, perfectly even luck would put every number within a small band of the others, and that is exactly what the data shows. The gap is noise, not a signal.
UK49s Lunchtime and Teatime
The two UK49s draws use the same 1–49 pool but are separate draws on separate machines, so their "hot" lists differ — Lunchtime's most-drawn ball is 47, Teatime's is 37. That difference is itself the lesson: if hot numbers carried any predictive weight, the same balls would dominate both draws. They do not. The lists diverge precisely because each draw is independent random selection, and the rankings reshuffle a little with every new result.
Treat the UK49s tables as trivia about the past. A ball that has been drawn slightly more often over 10,000 draws is not more likely to come up in tonight's draw, and a "cold" ball is not "due". The machine has no memory.
If you bet UK49s through a South African operator, our UK49s Lucky Numbers guide explains how the bet types and payouts actually work — which matters far more to your money than which balls are warm.
SA Daily Lotto
The SA Daily Lotto draws five balls from 1–36 every night, so its counts are smaller and its history shorter than UK49s. The same rule applies: ball 10 leading the all-time count does not make it a better pick than ball 19 at the bottom of the list. With a five-from-thirty-six draw every night, the counts will keep drifting and the order will keep changing — that drift is randomness at work, not a trend to follow.
For how the Daily Lotto bet works in South Africa, see our SA Daily Lotto guide.
Why "hot" and "cold" numbers cannot help you
Believing a number is "due" because it has not appeared recently is the gambler's fallacy — the single most common mistake in lottery and Lucky Numbers betting. Each draw is an independent event: the balls do not know or care what came before. A coin that lands heads ten times running is still 50/50 on the eleventh flip, and a lottery ball is no different.
Choosing numbers by frequency does not raise your chance of winning, and it does not lower it either — every combination is equally likely, so a frequency strategy simply has no effect on the odds while feeling like it does. The only honest use of this page is to understand the maths and, ideally, to gamble less because of it.
We tested it: 206 draws, and the numbers came out fair
Saying "the gaps are just noise" is easy. In July 2026 we checked it properly, and here is the whole method so you can repeat it.
We collected every UK49s Lunchtime draw from 1 January to 25 July 2026 — 206 draws — straight from the published results archive, and checked each one for integrity first (six distinct main balls, in range, correctly ordered; all 206 passed). That gives 1,236 main-ball observations spread across 49 numbers, so if the draw is fair each number should land about 25.2 times.
Then we ran a chi-squared goodness-of-fit test, which is the standard statistical tool for exactly this question: are these 49 counts consistent with every number being equally likely?
| |
|---|
| Sample | 206 draws, 1 Jan – 25 Jul 2026 |
| Main-ball observations | 1,236 |
| Expected per number | 25.2 |
| Chi-squared statistic | 50.29 (48 degrees of freedom) |
| Threshold for "not random" at 5% | 65.17 |
| p-value | 0.38 |
The result is 50.29 against a threshold of 65.17, giving a p-value of 0.38 — comfortably consistent with a fair, uniform draw. In plain terms: there is no evidence whatsoever that any ball is favoured. The variation is exactly what randomness produces.
Why the same data still looks like a pattern
Here is the trap, on the very same 206 draws:
- Hottest: 17 (drawn 39 times), then 24 (35), and 31, 38 and 40 (34 each).
- Coldest: 37 and 2 (16 each), 4 (17), 21 (18), 23 (20).
A gap of 39 against 16 looks enormous. A "hot numbers" page would sell it as a signal. But for any single number chosen in advance, pure chance alone produces a range of roughly 15 to 35 around that 25.2 average, 95% of the time.
And that is the subtle part: you are not looking at one number, you are looking at the highest and lowest of 49 at once. Scanning 49 numbers and reporting the extremes will routinely throw up values outside the single-number band even when nothing is wrong — which is precisely why the chi-squared test, which weighs all 49 together, is the honest way to ask the question. It found nothing.
So the hot/cold table is real data and a real tally. It is simply not evidence of anything, and the maths above is what proves it rather than asserts it.
How we compile this data
The all-time counts in the tables are transcribed from public, regularly updated statistics pages — lottonumbers.com for the two UK49s draws and lottery.co.za for the SA Daily Lotto — on the date shown under "Last updated". We publish the raw counts exactly as the source reports them and never adjust a figure to imply a pattern. Because new draws land daily, the numbers are a snapshot of the date we refreshed them, and we re-check them monthly.
The chi-squared test in the section above uses a separate dataset we collected ourselves: the full published archive of UK49s Lunchtime draws for 1 January to 25 July 2026, taken from the 2026 results archive. Every draw was checked for integrity before use, and the test counts only the six main balls, not the booster. We re-run it when we refresh the page.
Frequently asked questions
Do hot numbers win more often?
No. A "hot" number has been drawn more often in the past purely by chance, and that history has no bearing on the next draw. Every number has the same chance of being drawn every time, regardless of how often it has appeared before.
Are cold numbers "due" to come up?
No. The idea that a number not seen for a while is "due" is the gambler's fallacy. Each draw is independent, so a cold number is exactly as likely — no more, no less — than any other number on the next draw.
Can I use this data to improve my chances?
No method of picking numbers changes the odds of a random draw. Frequency, birthdays, patterns on the ticket — all have the same expected outcome. This page is for understanding the maths, not for building a system. If a site promises a winning method from frequency data, it is misleading you.
Why are the windows "all draws" instead of the last 90?
The tables report the full all-time frequency that the source pages publish, with the exact window recorded under each one. Over a longer window the figures are actually more revealing — they show how closely every number converges, which is the whole point. For a shorter, self-contained window we collected 206 UK49s Lunchtime draws from 2026 ourselves and tested them directly; that analysis is in the chi-squared section above.
Has anyone actually tested whether these draws are fair?
We have. On 206 UK49s Lunchtime draws from January to July 2026, a chi-squared goodness-of-fit test returns 50.29 on 48 degrees of freedom, p = 0.38 — well inside what a fair, uniform draw produces. There is no evidence any ball is favoured. The full method and the raw counts are in the section above so you can reproduce it.
Is betting on Lucky Numbers a good way to make money?
No. Lucky Numbers and lottery-style draws carry a built-in margin and every result is random, so over time the house keeps an edge. Treat any stake as money spent on entertainment, never as income. If it stops being fun, our responsible gambling guide explains the support available.
More statistics and guides
For the betting mechanics behind these draws, start with the Lucky Numbers hub, or read the UK49s and SA Daily Lotto guides. For sport instead of numbers, see our PSL statistics by season.
These tables record past draws only. Each draw is independent and random — past frequency does not predict future results, and no number is ever "due". 18+ only. Gambling involves financial risk; play for entertainment, never as income. If gambling is no longer fun, contact the National Responsible Gambling Programme on 0800 006 008 or SMS HELP to 076 675 0710.
