Project TD Tuning — Version 12

# V12 Context V12 didn’t change anything. No new factors, no recalibration, no threshold adjustments. We just let it run. All the way through. Every season we have data for. 25 years of football, start to finish, no one pulling the plug. The number that came out the other end: 63.8% across 7,017 games. That’s not a cherry-picked streak or a good run. That’s the full picture — good weeks and bad ones, playoff slates and Week 17 garbage time, dominant eras and parity years — all of it, averaged together. The summary table tells the real story. Scan it and you’ll see the model isn’t steady. There are green weeks scattered across decades, red weeks too, and long stretches of comfortable 60s. What held is the floor. It never collapsed. That’s what we wanted to know. We’re showing week-by-week summaries here instead of the full game-by-game breakdown — at this scale, the individual picks would bury the signal. We’ll get back to the per-game detail as we move forward. For now, the pattern is what matters.