This is South Africa's PowerBall (5 from 50 + 1 from 16). Looking for the US Powerball? 🇺🇸 US Powerball · 🇺🇸 Mega Millions

SA PowerBall: AI prediction and backtest – what are "statistical" lines worth?

Two honest experiments on the real draws: we compare random Quick Pick lines with "hot" lines built from the most frequent numbers, and we let a learning algorithm (k nearest neighbours) predict each draw, always using only the data known before the draw. Both methods are compared with the exact expectation of pure chance.

Backtest: Quick Pick lines vs hot lines (12 months)

For each of the 105 draws from 21.10.2025 to 20.10.2026, each group plays 10 lines: random Quick Pick versus lines built from the 15 most drawn numbers over the previous 100 draws. The PowerBall are random in both groups so that only the main numbers are compared. 1,050 lines per group in total.

MatchesQuick PickHotExpected by chance
0568625605.5
1406341369.2
2737770.3
3374.9
4000.1
5000.0

Wins by tier

TierMatchesQuick PickHotExpected by chance
15 matches + PowerBall000.00
25 matches000.00
34 matches + PowerBall000.01
44 matches000.10
53 matches + PowerBall000.31
63 matches374.60
72 matches + PowerBall404.40
81 match + PowerBall261523.07
90 match + PowerBall333537.84

Result: the two distributions differ over this period (chi² = 10.09, p = 0.018). Across many repeated tests, a gap like this turns up from time to time; it is not proof of a better method, as every combination remains equally likely.

AI prediction (k nearest neighbours)

For each draw, the model looks for the 7 past moments whose preceding 3 draws most resembled the current ones and plays the numbers most often drawn next. It is reapplied after every draw and never sees the future. 300 draws evaluated from 08.12.2023 to 20.10.2026.

Matches per prediction

ObservedExpected
0 Matches161 / 173.0
1 Matches117 / 105.5
2 Matches20 / 20.1
3 Matches2 / 1.4
4 Matches0 / 0.0
5 Matches0 / 0.0

Average: 0.543 matches per prediction – chance gives 0.500.

Deviation from the expectation of chance: chi² = 2.10, p = 0.350. The model hits exactly as often as numbers picked at random.

Model line for the next draw

19192540

For entertainment only: this line has exactly the same chance as any other.

Latest predictions

DatePredictionMatches
20.10.2026 35131624 0
16.10.2026 312212526 0
13.10.2026 511152932 1
09.10.2026 1011162141 0
06.10.2026 2427323843 0
02.10.2026 425343537 0
25.09.2026 1325274750 0
22.09.2026 913143143 1
18.09.2026 520244043 1
15.09.2026 413273137 1
11.09.2026 512274750 0
08.09.2026 513141924 1
04.09.2026 238404748 1
01.09.2026 2225294850 0
28.08.2026 513344048 0
25.08.2026 510313841 1
18.08.2026 18204648 1
14.08.2026 112132641 0
11.08.2026 19134043 1
07.08.2026 1113194748 0
04.08.2026 217192541 0
31.07.2026 89101347 1
28.07.2026 1627283250 1
24.07.2026 2729313240 0
21.07.2026 2224303239 0
17.07.2026 1461022 1
10.07.2026 515192629 0
07.07.2026 1013163243 1
03.07.2026 1617192425 0
30.06.2026 1320313250 0
26.06.2026 1423243750 1
23.06.2026 1013192444 0
19.06.2026 1924254750 0
16.06.2026 122283947 0
12.06.2026 26123847 0
09.06.2026 315173248 1
05.06.2026 120223141 0

How it works

A nearest-neighbour model looks for points in time whose last draws resembled the current ones and picks the numbers that followed most often. It is reapplied draw by draw without looking into the future; a backtest additionally compares Quick Picks with lines made of frequent numbers.

Example

For 5 from 50, 0.5 matches per line are expected on average; the model lands exactly there.

Limitations

No model can predict random draws. The page shows this transparently with a chi-squared test.

Frequently asked questions about this tool

Can AI predict lottery numbers?
No. With independent, uniformly distributed draws there is no pattern a model could learn.
▲