Fantasy · MLB prop projections
Independent pitcher-strikeout projections (affine-corrected mean, Negative-Binomial, own-history probability calibration) · market lines shown as reference only · slate 2026-09-02
Predictions for 2026-09-02.numbers frozen Sep 03 02:34 UTC (stable across rebuilds) · refreshed hourly · graded automatically after games go final
Accuracy record
Out-of-sample calibration: stated 53.1% vs realized 56.0% (gap +3.0 pts, n=1058)
Raw track (source model, no anchor/ceiling): stated 60.7% vs realized 55.3% (gap -5.3 pts, n=994)
Highest-confidence predictions today
| player | prop | predicted | corrected | lean | P | tier | ref line | deviation |
|---|
| Casey Mize | K | 4.39 | 4.39 | LESS | 53.0% | D | 4.5 | -0.11 |
| Brandon Williamson | K | 3.75 | 3.75 | MORE | 53.0% | D | 3.5 | +0.25 |
| Grant Holmes | K | 3.79 | 3.79 | LESS | 53.0% | D | 4.0 | -0.21 |
| Jacob Lopez | K | 5.32 | 5.32 | LESS | 53.0% | D | 5.5 | -0.18 |
| Cody Bradford | K | 4.24 | 4.24 | LESS | 53.0% | D | 4.5 | -0.26 |
| Trevor Rogers | K | 4.96 | 4.96 | LESS | 53.0% | D | 5.5 | -0.54 |
Full board — every pitcher scored
| player | prop | game | predicted | corrected | P(more) | lean | P | tier | raw P(more) | ref line | deviation | RW proj | RW | bench |
|---|
| Casey Mize | K | 824470 | 4.39 | 4.39 | 47.0% | LESS | 53.0% | D | 43.8% | 4.5 | -0.11 | 5.2 | ≠ | — |
| Brandon Williamson | K | 824470 | 3.75 | 3.75 | 53.0% | MORE | 53.0% | D | 50.1% | 3.5 | +0.25 | 3.5 | ≠ | 3.10 |
| Grant Holmes | K | 822686 | 3.79 | 3.79 | 47.0% | LESS | 53.0% | D | 33.4% | 4.0 | -0.21 | 5.2 | ≠ | 3.06 |
| Jacob Lopez | K | 822851 | 5.32 | 5.32 | 47.0% | LESS | 53.0% | D | 43.1% | 5.5 | -0.18 | 4.4 | = | 6.99 |
| Cody Bradford | K | 822851 | 4.24 | 4.24 | 47.0% | LESS | 53.0% | D | 41.1% | 4.5 | -0.26 | — | · | 2.52 |
| Trevor Rogers | K | 824312 | 4.96 | 4.96 | 47.0% | LESS | 53.0% | D | 37.7% | 5.5 | -0.54 | — | · | 6.59 |
| Tomoyuki Sugano | K | 824312 | 3.93 | 3.93 | 53.0% | MORE | 53.0% | D | 53.3% | 3.5 | +0.43 | 3.5 | ≠ | 3.27 |
| Andrew Painter | K | 825037 | 3.94 | 3.94 | 53.0% | MORE | 53.0% | D | 53.4% | 3.5 | +0.44 | — | · | 4.84 |
| Taylor Clarke | K | 825037 | 3.50 | 4.00 | 53.0% | MORE | 53.0% | D | 95.8% | 0.5 | +3.50 | — | · | — |
| Patrick Sandoval | K | 824717 | 5.08 | 5.08 | 47.0% | LESS | 53.0% | D | 39.5% | 5.0 | +0.08 | — | · | 5.81 |
| Bryce Miller | K | 824717 | 4.56 | 4.56 | 47.0% | LESS | 53.0% | D | 46.6% | 4.0 | +0.56 | — | · | 3.56 |
| Justin Hagenman | K | 822931 | 0.00 | 2.25 | 47.0% | LESS | 53.0% | D | 0.0% | 3.0 | -0.75 | — | · | — |
| Griffin Jax | K | 822931 | 5.36 | 5.36 | 53.0% | MORE | 53.0% | D | 59.0% | 4.5 | +0.86 | 5.0 | = | 6.72 |
| Joey Cantillo | K | 824390 | 4.60 | 4.60 | 47.0% | LESS | 53.0% | D | 47.2% | 4.5 | +0.10 | — | · | 4.63 |
| Jared Jones | K | | 5.29 | 5.29 | 47.0% | LESS | 53.0% | D | 42.8% | 5.0 | +0.29 | — | · | — |
| Landen Roupp | K | 823339 | 5.03 | 5.03 | 47.0% | LESS | 53.0% | D | 38.8% | 5.0 | +0.03 | — | · | 4.11 |
| David Peterson | K | 824634 | 4.94 | 4.94 | 53.0% | MORE | 53.0% | D | 52.8% | 4.0 | +0.94 | — | · | 6.59 |
| Drew Anderson | K | 823660 | 4.40 | 4.40 | 47.0% | LESS | 53.0% | D | 43.9% | 4.0 | +0.40 | — | · | 4.74 |
| Eury Pérez | K | 824068 | 4.62 | 4.62 | 47.0% | LESS | 53.0% | D | 32.4% | 5.0 | -0.38 | — | · | 5.24 |
| Noah Cameron | K | 824068 | 4.74 | 4.74 | 47.0% | LESS | 53.0% | D | 49.5% | 4.5 | +0.24 | — | · | 4.71 |
| Jacob Misiorowski | K | 824634 | 8.03 | 8.03 | 53.0% | MORE | 53.0% | D | 52.3% | 7.5 | +0.53 | — | · | 6.85 |
| Hayden Wesneski | K | 824147 | 4.81 | 4.81 | 47.0% | LESS | 53.0% | D | 35.3% | 5.0 | -0.19 | — | · | 4.55 |
| Cam Schlittler | K | 823983 | 7.51 | 7.51 | 47.0% | LESS | 53.0% | D | 35.0% | 8.0 | -0.49 | — | · | 8.28 |
| Reid Detmers | K | 823983 | 7.22 | 7.22 | 53.0% | MORE | 53.0% | D | 55.2% | 6.5 | +0.72 | — | · | — |
| Yoshinobu Yamamoto | K | 823906 | 6.03 | 6.03 | 47.0% | LESS | 53.0% | D | 39.6% | 6.5 | -0.47 | — | · | — |
RW = RotoWire independent projection: = same lean · ≠ opposite lean. Probabilities are computed line-free: affine-corrected mean, fitted dispersion, and a calibration layer fitted on this system's own graded history. The reference line is shown for comparison only; the lean and deviation both follow the corrected mean, so they always agree.
Statistical projections with quantified uncertainty, graded daily against official MLB results above. Point predictions are raw model output; probabilities reflect only what settled data has supported.