Lotto Prediction Lab

Explore how Lotto numbers have been drawn in recent South African draws, and see for yourself how much a ranking depends on the settings you choose.

Please read first: Lotto draws are random, and the balls have no memory. Nothing in this lab changes the chance of any future draw, and a number that has been drawn often, or not at all lately, is no more likely to appear next time. This is an educational statistics tool. You must be 18 or older to play. Play responsibly.

About Lotto and our dataset

Lotto is South Africa’s best-known lottery game, normally drawn on Wednesdays and Saturdays. In each draw, 6 main numbers are chosen from 1 to 52. We keep a record of past results, and everything on this page is worked out from those records. The lab scores the main numbers only.

With 6 numbers drawn from 52, there are 20,358,520 possible combinations. Every one of them has the same chance of being drawn, which is 1 in 20,358,520 to match all 6 numbers.

The Lotto dataset used on this page
GameLotto
Number pool1 to 52
Main numbers drawn6
Historical sample100 draws (we use up to 100)
Sample period to
Latest draw
Recent windowThe latest 40 draws
Records last updated5 Nov 2026
Data sourceResults published by the lottery operator, stored in our archive

This sample of 100 draws spans about 51 weeks of draws, so earlier history isn’t included. The lab only uses these latest draws. Older results may be on the results page.

Illustrative ranking (not a tip)

81022242932

This is an illustration of how the scoring works, not a tip or a recommendation to play these numbers. Every combination has exactly the same chance of being drawn. These are the 6 highest-scoring numbers with the starting settings (50% frequency, 35% recent form, 15% gap since last drawn). The percentages are weights you can change, not probabilities.

Try the weight lab

Each slider decides how much one measure counts towards a number’s score. Move them and the highest-scoring numbers re-rank straight away. Notice how different settings give different lists. That is the whole point: the ranking reflects your choices, not anything hidden in the draws.

Highest-scoring numbers for the current settings

81022242932

Understanding Lotto Historical Statistics

These are the four things the lab looks at. Each one is a count or comparison taken from past draws:

The measures used in the lab and what each one means
MeasureWhat it isWhat its slider does
Number frequencyHow many times a number appeared in the 100 draws we analysed.A higher weight lifts the numbers that appeared most often.
Recent frequencyHow many times it appeared in the latest 40 draws only.A higher weight lifts numbers that showed up lately.
Last appearanceHow many draws have passed since the number last came up.A higher weight lifts numbers that have been missing the longest.
Frequency distributionHow the counts compare across all 52 numbers, shown in the bars and tables further down.No slider. It is there to read, not to score.

A number’s score is its frequency, recent and last-appearance measures, each scaled from 0 to 1, multiplied by your weights and added up. The lab then lists the 6 highest scores. Patterns in past draws don’t establish what will happen in future ones.

The exact scaling

Frequency and recent frequency are divided by the highest count among the numbers, so the busiest number scores 1. Last appearance is divided by half the sample (at least 5 draws) and capped at 1, so a longer gap gives a higher score.

Do some numbers really come up more than others?

Short version: the counts look about like chance. If every draw were purely random, each number would show up about 11.5 times in 100 draws, give or take about 3.2. Here the most-drawn number appeared 18 times and the least-drawn 4, and 49 of 52 numbers landed inside that normal range. For random draws you would typically expect around 19 in 20. That is a rough guide, not a formal test.

Some numbers will always look busier or quieter in a sample like this. That is simply what randomness looks like, and it isn’t a pattern you can use.

Most drawn in this sample

Five most-drawn numbers in the sample
NumberTimes drawn
2218
816
1016
2416
3216

Least drawn in this sample

Five least-drawn numbers in the sample
NumberTimes drawn
274
315
256
386
37

Odd/even and high/low splits

Every draw can be described by how many of its numbers are odd or even, and how many are low (1 to 26) or high (27 to 52). Here is how the draws in our sample divide up.

Odd and even

Odd and even split across sampled draws
SplitDrawsShare
0 odd, 6 even22.0%
1 odd, 5 even1717.0%
2 odd, 4 even2828.0%
3 odd, 3 even3131.0%
4 odd, 2 even1616.0%
5 odd, 1 even55.0%
6 odd, 0 even11.0%

Low and high

Low and high split across sampled draws
SplitDrawsShare
0 low, 6 high22.0%
1 low, 5 high1111.0%
2 low, 4 high2323.0%
3 low, 3 high3131.0%
4 low, 2 high2424.0%
5 low, 1 high99.0%
6 low, 0 high00.0%

Mixed splits are more common than extreme ones simply because there are more ways to make them. That says nothing about which combination will be drawn.

Lab backtest: how would it have done on the last 5 draws?

Here is a fair test. For each of the last few draws, we step back to just before it, run the lab on only the draws that came earlier, and see how many of its top 6 numbers turned up. A tick means one did.

These picks weren’t published at the time. We worked them out afterwards, using the same method as the lab above.

Lab selection compared with the actual draw for the last 5 draws
DrawLab’s selectionNumbers drawnMatched
4 Nov 2026 #2695
8 10 22 (matched) 23 24 40
71922293246
1 of 6
31 Oct 2026 #2694
8 10 22 (matched) 23 24 40
51520222839
1 of 6
28 Oct 2026 #2693
8 (matched) 10 22 23 24 (matched) 40
81118244148
2 of 6
24 Oct 2026 #2692
8 (matched) 10 (matched) 22 23 24 40
289103034
2 of 6
17 Oct 2026 #2690
8 10 (matched) 22 23 24 40
246101546
1 of 6

Across the last 50 draws we could test this way, the lab matched 0.76 numbers per draw on average. Picking 6 numbers completely at random from 52 would match about 0.69 (a rough yardstick, not a statistical test). When the two are close, the method is adding nothing beyond luck, and a small gap either way is just ordinary variation.

How often each number was drawn

22
18
8
16
10
16
24
16
32
16
40
15
12
14
4
13
18
13
29
13
45
13
6
12
7
12
23
12
26
12
39
12
43
12
46
12
Show the counts as a table
Times each number was drawn in the sample
NumberTimes drawn
111
211
37
413
59
612
712
816
99
1016
119
1214
1311
1410
1511
167
177
1813
1911
2011
217
2218
2312
2416
256
2612
274
287
2913
3011
315
3216
339
3410
358
3611
377
386
3912
4015
417
4211
4312
449
4513
4612
479
4812
4912
5010
5110
528

Past 100 draws

The most recent draws in our archive, 10 to a page. Always check the official result for anything that matters.

Past Lotto draws, page 3 of 10
Draw Numbers
19 Aug 2026 #2673
101921405051
15 Aug 2026 #2672
21319203452
12 Aug 2026 #2671
61127283445
8 Aug 2026 #2670
71013182442
5 Aug 2026 #2669
172329313340
1 Aug 2026 #2668
91112243941
29 Jul 2026 #2667
41522244152
25 Jul 2026 #2666
91326282943
22 Jul 2026 #2665
2428323435
18 Jul 2026 #2664
102540414748

About Lotto Historical Data

TechChamp maintains a historical archive of Lotto results for informational and educational purposes. The statistical information on this page is calculated from the draw records available in our database.

Number frequencies and historical distributions describe previously published results. They do not indicate which numbers will appear in future draws. Each valid combination has the same probability under the applicable draw rules.

Official operator publications remain the authoritative source for confirming results.

Common questions

Does this predict the next Lotto draw?

No. The lab ranks numbers using past draws, and past draws don’t influence future ones. It’s a way to look at the data, not a forecast.

What are the chances of matching all 6 numbers?

With 6 numbers drawn from 52, there are 20,358,520 possible combinations, so any one ticket has a 1 in 20,358,520 chance of matching all 6 main numbers. Past results don’t change that.

Does the lab include the bonus ball?

No. The lab only scores the main numbers. The bonus ball is shown in the results table where our records include it.

Where does the Lotto data come from?

From the results the lottery operator publishes, which we store in our archive. If a result here ever differs from the official one, the official result is correct.

How many draws does the lab look at?

Up to the latest 100 draws in our archive. The exact number and the dates they cover are shown on the page, and older results may be on the results page.

What does “last appearance” mean, and does a long gap matter?

It counts how many draws have passed since a number last came up. Many players wonder whether a long gap means a number is “due”. It doesn’t: a long gap carries no information about the next draw.

Do the sliders improve my chances?

No. Every valid combination has exactly the same chance each time, however the lab ranks the numbers.

Why do different settings give different lists?

Because the list depends entirely on how much weight you give each measure. That is why none of the lists should be treated as advice.

Were the backtest selections real predictions?

No. We worked them out afterwards using only the draws that came before each one, so you can see how the lab’s method behaves. We did not publish them in advance, and how they scored says nothing about future draws.

Where can I find older results?

The Lotto results page keeps previous draws, and you can use the ticket checker there to compare your numbers with a past draw.

Keep exploring

Play responsibly

The lottery is for adults aged 18 and over. Think of a ticket as a bit of entertainment, not a way to make money. Only spend what you can comfortably afford to lose, and don’t chase losses. If gambling is affecting you or someone close to you, please talk to a qualified support organisation in your area. Read our responsible play page