Every pick, public

How this works

• • Every pick the model generates is written to our database at pick time, with the odds we recommended and our confidence rating.
• • Picks are graded automatically by the same cron job that pulls final scores from the major sports APIs. We don't manually mark losses as pushes.
• • ROI is calculated against the best available line at pick time, betting 1 unit per pick flat-stakes. This is real- world P&L, not a back-test.
• • The full per-factor and calibration breakdown is on /stats if you want to dig deeper.