Lottochaos

Québec 49: 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 11.10.2025 to 10.10.2026, each group plays 10 lines: random Quick Pick versus lines built from the 15 most drawn numbers over the previous 100 draws. The Bonus number are random in both groups so that only the main numbers are compared. 1,050 lines per group in total.

MatchesQuick PickHotExpected by chance
0430465457.8
1455417433.7
2152145139.0
3122118.5
4121.0
5000.0
6000.0

Wins by tier

TierMatchesQuick PickHotExpected by chance
16 Matches000.00
25 Matches + Bonus number000.00
35 Matches000.02
44 Matches121.02
53 Matches122118.53
62 Matches + Bonus number101512.93

Result: the difference between the two groups is not statistically significant (chi² = 5.97, p = 0.113). It stays within what pure chance produces for a sample of this size.

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 29.11.2023 to 10.10.2026.

Matches per prediction

ObservedExpected
0 Matches136 / 130.8
1 Matches105 / 123.9
2 Matches49 / 39.7
3 Matches10 / 5.3
4 Matches0 / 0.3
5 Matches0 / 0.0
6 Matches0 / 0.0

Average: 0.777 matches per prediction – chance gives 0.735.

Deviation from the expectation of chance: chi² = 7.23, p = 0.027. The deviation is noticeable but may be down to chance in this sample; every combination remains equally likely.

Model line for the next draw

102829303140

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

Latest predictions

DatePredictionMatches
10.10.2026 51135373844 0
07.10.2026 92228293440 0
03.10.2026 32938434547 1
30.09.2026 4914272834 1
23.09.2026 141922263642 0
12.09.2026 21316223846 0
09.09.2026 41521273739 0
02.09.2026 112229374046 1
26.08.2026 71112223643 0
22.08.2026 5819222325 1
19.08.2026 2811164449 1
15.08.2026 2920253645 0
12.08.2026 1920244246 0
08.08.2026 2825294247 1
05.08.2026 111328354348 1
01.08.2026 2814202240 0
29.07.2026 51726374246 1
22.07.2026 112529414243 0
18.07.2026 51114223046 0
15.07.2026 51227424647 0
11.07.2026 3610111346 0
08.07.2026 5929333746 0
04.07.2026 5911151943 0
01.07.2026 32030323438 1
27.06.2026 256112025 0
24.06.2026 142930313242 1
20.06.2026 52028373943 0
17.06.2026 111331424345 0
13.06.2026 91718363749 0
10.06.2026 61921344345 0
30.05.2026 61120214146 0
27.05.2026 2412192029 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 6 from 49, 0.7 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.
▲