How AutoPick analyses lottery numbers
AutoPick turns decades of verified draw history into research-backed number selections. Here is the full pipeline, from raw results to an explained line.
1. Import verified draw history
Every historical draw for the game is imported from official sources, checked against the game rules in force on the day, and stored with the exact numbers, draw instance and prize breakdown. Disputed or unverified records are excluded from analysis.
2. Profile every number
Each number is measured for frequency across rolling windows, current absence gap, average and longest gap, and how often it appears alongside every other number as a pair or triple.
3. Fingerprint every draw
Each historical result is broken down into its structure: sum, spread, odd/even balance, low/high balance, gap pattern, consecutive runs and repeats from the previous draw. This is the draw DNA used for comparison.
4. Build a candidate pool
For an upcoming draw, AutoPick generates tens of thousands of valid combinations that satisfy the game's ticket rules, then filters out structurally implausible ones before scoring begins.
5. Score and rank with an ensemble
Multiple independent research models score each candidate. Their weights are set by how well each model has performed in walk-forward testing, so the ensemble reflects measured performance rather than opinion.
6. Apply shared-jackpot analysis
Combinations that large numbers of players commonly choose — birthday-heavy ranges, straight lines, obvious sequences — carry a higher risk of splitting a prize. AutoPick scores that risk and prefers lines that would be shared less.
7. Publish with explanations
The strongest lines are published as the week's research lines, each with a quality score, a confidence band and a plain-English explanation of the factors behind it.
8. Measure the outcome honestly
After the draw, every research line is scored against the actual result and against a random baseline. That feedback updates model weights, so the engine keeps learning from real outcomes.