Methodology

Transparency matters, so here is exactly how every article on this site is produced.

1. Data source

All winning numbers come from the official New York Open Data (Socrata) datasets, the same public feeds used by news organizations. We never transcribe numbers by hand.

2. Statistical window

Unless stated otherwise, statistics use the most recent 50 draws of each game: frequency counts (hot numbers), longest absence streaks (cold / overdue numbers), odd–even and low–high splits, sum distributions, consecutive-pair rates and repeat rates.

3. AI predicted sets

Each article includes candidate number sets built by different strategies:

• Hot Streak — samples numbers weighted by recent frequency.
• Balanced Mix — enforces the historically dominant odd/even and low/high structure and a typical sum range.
• Overdue Watch — boosts the longest-absent numbers over a hot-number core.
• Pattern Play — mirrors recent structural tendencies such as consecutive pairs and repeats.
• Wildcard — a uniform random line for contrast.

4. The honest part

Lottery drawings are independent random events. Past frequency does not change future odds, and no method — statistical or AI — can genuinely predict a random draw. Our sets are a fun, structured way to pick numbers, not a promise of results. Please read our full disclaimer and always play responsibly.