The pre-registered 2026 test
This page was published before the 2026 postseason. Every team’s 2026 regular-season slope, estimated through 2026-08-20, is frozen below and in the repository at commit 6e92608d45d8. When the postseason ends, one regression gets run: each club’s opponent-adjusted October performance against its frozen slope times the pitching-quality jump it actually faces. The anomaly predicts a coefficient near 0.9. The null predicts zero. The result will be reported either way.
| Team | Frozen slope (raw) | After shrinkage | PA | Reading |
|---|---|---|---|---|
| HOU | +1.18 | +0.116 | 4257 | appears to resist good pitching (probably noise) |
| TOR | +0.79 | +0.077 | 4484 | appears to resist good pitching (probably noise) |
| TEX | +0.56 | +0.047 | 4275 | appears to resist good pitching (probably noise) |
| DET | +0.50 | +0.041 | 4492 | appears to resist good pitching (probably noise) |
| CHC | +0.45 | +0.043 | 4715 | appears to resist good pitching (probably noise) |
| PHI | +0.42 | +0.035 | 4450 | appears to resist good pitching (probably noise) |
| NYM | +0.37 | +0.035 | 4503 | appears to resist good pitching (probably noise) |
| ATL | +0.28 | +0.026 | 4416 | indistinguishable from average |
| STL | +0.20 | +0.017 | 4550 | indistinguishable from average |
| PIT | +0.17 | +0.017 | 4646 | indistinguishable from average |
| MIL | +0.14 | +0.015 | 4499 | indistinguishable from average |
| SEA | +0.11 | +0.010 | 4356 | indistinguishable from average |
| CWS | +0.02 | +0.002 | 4360 | indistinguishable from average |
| WSH | +0.02 | +0.001 | 4472 | indistinguishable from average |
| CLE | -0.04 | -0.003 | 4234 | indistinguishable from average |
| ATH | -0.09 | -0.008 | 4264 | indistinguishable from average |
| MIA | -0.10 | -0.010 | 4298 | indistinguishable from average |
| TB | -0.14 | -0.011 | 4255 | indistinguishable from average |
| BAL | -0.18 | -0.017 | 4372 | indistinguishable from average |
| MIN | -0.18 | -0.014 | 4401 | indistinguishable from average |
| SD | -0.19 | -0.018 | 4310 | indistinguishable from average |
| KC | -0.20 | -0.018 | 4389 | indistinguishable from average |
| BOS | -0.23 | -0.021 | 4374 | indistinguishable from average |
| LAA | -0.37 | -0.036 | 4387 | appears to wilt vs good pitching (probably noise) |
| CIN | -0.37 | -0.031 | 4353 | appears to wilt vs good pitching (probably noise) |
| NYY | -0.42 | -0.040 | 4388 | appears to wilt vs good pitching (probably noise) |
| LAD | -0.44 | -0.045 | 4477 | appears to wilt vs good pitching (probably noise) |
| ARI | -0.50 | -0.052 | 4396 | appears to wilt vs good pitching (probably noise) |
| COL | -0.68 | -0.059 | 4434 | appears to wilt vs good pitching (probably noise) |
| SF | -0.92 | -0.096 | 4316 | appears to wilt vs good pitching (probably noise) |
Note the shrinkage column: even the most extreme raw slope in baseball earns a believed effect of about a tenth of its face value. If the October test comes back at zero, that column was the answer all along. If it comes back near 0.9, something genuinely interesting is happening that regular-season reliability cannot see, and the mechanism study begins. Milwaukee’s frozen slope, for the record: mildly positive.