Scorecard
How every one of our models has done against the books. Yes, we show the losing ones too. Every record here updates itself each morning after the night's results come in.
NFL · 2026 season · week 9
Every NFL game this season, for each of our models: did the team it favored win, and did the side its spread liked cover the closing spread.
| Straight up · this week | Straight up · season | Against the spread · this week | Against the spread · season | ||
|---|---|---|---|---|---|
| QB-adjusted | 0–0 | 40–3156% | 0–0 | 36–32–353% | Games → |
| Basic | 0–0 | 42–2959% | 0–0 | 37–30–355% | Games → |
| Long memorysince Oct 15 | 0–0 | 26–1662% | 0–0 | 20–21–149% | Games → |
| Sim | 1–0100% | 46–2763% | 0–10% | 43–28–361% | Games → |
| The bet | — | — | 0–10%−1.0u | 25–20–256%+2.7u | Games → |
Props leans vs the line
Every prop where our projection was over or under the posted line, graded on the result, this season.
| Record | Win % | Graded | |
|---|---|---|---|
| NFL | 1292–1059 | 55% | 2351 |
| NBA | 1211–1060 | 53% | 2271 |
Spread records are graded at the closing line: the last price before the game started. Prop leans are graded against the line posted when we made the projection, and first basket at the best US price. Pushes count as neither a win nor a loss. A record is not a pick: it grades what we published, after the games. Each season counts only predictions saved before the game; a model marked “since” started grading mid-season. Last graded Sat, 6:23 AM EST.