2026-10-16
Registered 2026-10-11T23:13:03.799Z · history through 2026-10-09
AI selection 1
614313350
Euro numbers
312
AI selection 2
115434850
Euro numbers
12
Forecast audit · Model v1.0.0
Two Euro numbers.
One result to measure.
AI selection 2 matched main number 35 and both Euro numbers, 4 and 8. The original line was registered on 11 October, using results through 9 October.
Official result
One jackpot-winning line · €120,000,000
1117202235★48
The winning line was played in North Rhine-Westphalia, Germany. Official ONCE prize breakdown · Regional report citing WestLotto.
Every available line, including misses
| Method | Main numbers | Euro numbers | Matches |
|---|---|---|---|
| AI selection 1 | 413344448 | 410 | 0+1 |
| AI selection 2 | 831333536 | 48 | 1+2 |
| Random control 1 | 1420294049 | 16 | 1+0 |
| Random control 2 | 219214142 | 610 | 0+0 |
| Calendar hypothesis | 2333344147 | 57 | 0+0 |
| History filter | 711202845 | 38 | 2+1 |
Outlined balls matched the draw. Random controls are separate baseline lines, not additional AI selections. AI and random rows come from the timestamped forecast log; the calendar row comes from its locked record. The history filter is a reconstruction from earlier results, with no separate public pre-draw timestamp. Personal and manual lines cannot be evaluated without the numbers actually selected.
AI selection 2 matched the 1+2 prize category (€10.00 in the official euro table). The history-filter line matched 2+1 (€8.90). These are comparisons with the prize table, not claims that a ticket was bought or a prize paid. Operator currency amounts may differ.
The AI method, in plain English
- Look backwards. For each number, count appearances in the last 50 draws, its frequency over the eligible history, and how many draws have passed since its last appearance.
- Learn small adjustments. A regularised logistic model combines those features. It starts near equal weights and limits adjustments to avoid treating noisy patterns as certainty. Euro-number history resets when the number pool changes.
- Make a weighted random selection. Select five different main numbers and two different Euro numbers without replacement. More weight means more chance of being selected by this generator, not a proven greater chance in the actual lottery.
- Make it repeatable. A fixed seed uses the model version, target date, last available result date, mode and line index. AI selections 1 and 2 use the same model with different line indices.
- Freeze the line before the draw. Keep the original values and later compare them with the official result. Friday's existing forecasts remain unchanged.
AI selection 2 is not the seven highest-scoring numbers. Euro number 8 had a lower model score than the equal-weight baseline and still entered through the weighted random selection. The method reproduced the original line using only the 997 earlier draws; today's result was excluded.
What the model saw before this draw
| Selected number | Last 50 draws | Eligible history | Draws since last hit | Model inclusion score |
|---|---|---|---|---|
| Main 8 | 2/50 | 104/997 | 22 | 10.09% |
| Main 31 | 4/50 | 94/997 | 10 | 10.09% |
| Main 33 | 6/50 | 95/997 | 5 | 10.00% |
| Main 35 | 7/50 | 114/997 | 8 | 9.80% |
| Main 36 | 7/50 | 87/997 | 3 | 10.02% |
| Euro 4 | 10/50 | 74/475 | 5 | 16.91% |
| Euro 8 | 7/50 | 70/475 | 7 | 15.74% |
The equal-weight baseline is 10% per main number and 16.67% per Euro number. These model scores are experimental marginal estimates used as sampling weights. They are not validated physical probabilities or the chance of winning with the complete line.
What this result tells us
One fixed Euro pair has a 1 in 66 chance (about 1.52%) of matching both Euro numbers in a fair draw. Across the two registered AI pairs for this draw, the chance of at least one pair matching was 2 in 66 (about 3.03%): both pairs are distinct. The exact 1+2 category has a chance of about 1 in 188 per line.
Both AI lines must count in the evaluation. Their main-number matches were 0 and 1; their Euro-number matches were 1 and 2. One successful line does not establish improved future odds, and none matched 5+2. Keep the method fixed and track every future result against random controls.
The earlier calendar result of 13.4% more historical individual-number matches belongs to a different method. The calendar line scored 0+0 here; it cannot be credited with the AI result.
The next registered AI selections
16 October retains the original forecasts through 9 October. The 20 October forecasts include the newly published 13 October result. These are dated forecast records, not a live list of upcoming draws; use the ticket builder for the current next two dates.
2026-10-20
Registered 2026-10-13T19:15:03.162Z · history through 2026-10-13
AI selection 1
1330384142
Euro numbers
59
AI selection 2
710114349
Euro numbers
35
Check the record
Download the comparison and model factors · Timestamped AI and random forecast log · Locked calendar study · Model source
Forecast timestamps are self-published and are not independent third-party certification. Each valid line retains the fair-draw jackpot chance of 1 in 139,838,160.