NFL recap
Week 4 · 2026 — graded
Every completed week, the TD Board gets graded against what actually happened — the board exactly as it was published, no retouching. Settled like the bet: a pick who didn't play voids. TDs = rushing or receiving, the outcome the model is calibrated on.
Tier A
22/41
scored · expected 15.7 · 4 void
Tier B
10/39
scored · expected 11.1 · 6 void
Tier Sleeper
5/33
scored · expected 6.2 · 12 void
Overall
37/113
scored · expected 33.0 · 22 void
| Week | A | B | Sleeper | Overall |
|---|---|---|---|---|
| 2026 · Wk 5 | 20/39exp 15.1 | 9/38exp 11.1 | 6/32exp 6.3 | 35/109exp 32.5 |
| 2026 · Wk 4 | 22/41exp 15.7 | 10/39exp 11.1 | 5/33exp 6.2 | 37/113exp 33.0 |
| 2026 · Wk 3 | 11/40exp 15.3 | 12/37exp 10.5 | 9/39exp 7.5 | 32/116exp 33.3 |
| 2026 · Wk 2 | 14/36exp 13.9 | 11/41exp 11.9 | 4/32exp 6.6 | 29/109exp 32.4 |
| 2026 · Wk 1 | 18/34exp 13.4 | 11/40exp 12.1 | 11/35exp 7.6 | 40/109exp 33.1 |
- A79.835.9% TD✓ TD
- B64.527.4% TD✓ TD
- B63.027.4% TD✓ TD
- A91.341.4% TD✓ 2 TDs
- A92.941.4% TD✓ TD
- A80.436.0% TD✓ TD
- A74.635.2% TD✓ TD
- B73.531.9% TD✓ TD
- S32.713.9% TD✓ TD
- A91.541.4% TD✓ TD
- S40.616.4% TD✓ TD
- A65.431.4% TD✓ 2 TDs
- S62.629.9% TD✓ TD
- A92.641.4% TD✓ 3 TDs
- A89.841.4% TD✓ 2 TDs
- A88.341.4% TD✓ TD
- B76.932.9% TD✓ TD
- A79.839.0% TD✓ TD
- B62.527.4% TD✓ TD
- A90.741.4% TD✓ 2 TDs
- A76.233.4% TD✓ TD
- B61.927.4% TD✓ TD
- B50.520.3% TD✓ TD
- S38.516.3% TD✓ TD
- A91.341.4% TD✓ 2 TDs
- A84.136.9% TD✓ TD
- A88.541.1% TD✓ TD
- A76.937.3% TD✓ 2 TDs
- A96.041.4% TD✓ TD
- A79.733.8% TD✓ TD
- A77.236.1% TD✓ TD
- A88.041.4% TD✓ 2 TDs
- B68.729.2% TD✓ 2 TDs
- A95.041.4% TD✓ 2 TDs
- B64.827.4% TD✓ TD
- B56.625.9% TD✓ 3 TDs
- S34.816.0% TD✓ 2 TDs
Each card lists only the ranked threats who scored; the "N/M hit" badge and the tier totals above count everyone, so the misses are in every rate on this page. Graded from the week's stored board — Tier A / B / Sleeper are the top 3 / next 3 / next 3 per game, exactly as published. "Expected" is the sum of the model's calibrated TD probabilities over settled picks: if the calibration is honest, hits should track it over a season. DNPs void, same as the bet. 13 picks have no game log in the grading build (players who dropped off the featured list between builds) — shown as "no log" and left out of the rates rather than counted as misses.