How the score is built
Each outcome is fit as logit P = μ + βbatter + πpitcher across every plate appearance in the season, so hitter and pitcher skill are solved together rather than one being assumed. The adjusted column is this player against a league-average opponent; the run value is that rate priced by its wOBA weight (a walk is worth .690 runs, a home run 2.015), and those values sum to the adjusted wOBA above.
| Outcome | Actual | Adjusted | League | Run value | Regressed, PA |
|---|
| Walk / HBP | 6.5% | 8.6% | 9.8% | 0.0594 | 137 |
|---|
| Strikeout | 19.6% | 20.1% | 22.2% | 0.0000 | 50 |
|---|
| Single | 10.9% | 13.1% | 14.0% | 0.1152 | 246 |
|---|
| Double | 5.4% | 4.2% | 4.1% | 0.0518 | 3972 |
|---|
| Triple | 1.1% | 0.4% | 0.3% | 0.0064 | 922 |
|---|
| Home run | 0.0% | 2.2% |
|---|
Regressed is how many plate appearances of league-average performance are blended in before a rate is believed — estimated from the spread of the league itself, not chosen. Strikeout rate separates hitters quickly and is barely touched; doubles are mostly noise and are pulled hard toward average. It is why a hot 30-PA cameo does not top this list.
The opponent effect is usually small over a full season, and that is a finding rather than a shortcoming: individual pitchers differ enormously (wOBA allowed spans roughly .263 to .342), but across hundreds of plate appearances almost everyone faces a near-representative sample. The adjustment moves part-time players far more than regulars.
