Every method states its probabilities before each draw using only earlier draws, then is scored on what came out. The score is compared with simply treating all numbers as equal. Above zero means it predicted better; below zero means it predicted worse. The engine's forecast is a blend weighted by this track record.
| Ticket | Likelihood vs. an average ticket | Crowding |
|---|
Each strategy's tickets are logged before a draw and scored after it, so nothing can be adjusted in hindsight. Random is the yardstick.
| Strategy | Draws | Tickets | Spent | Won | Return | Winning tickets |
|---|
Three tests on the current format's draws. A flag here would mean something physical is off with the equipment, which is the only situation where past draws could tell you anything about the next one.
| Ran (server time) | Game | Rows seen | New draws | Result |
|---|
Source: New York State open data. Checked every hour by the scheduled task.