1+ home run
Research only by policy. It passes the backtest. The live forward test now has 1,473 lineup hitters who batted over 26 dates, past the 200-result and 7-date minimums.
Research only: Tier 3 by policy. Published for context, never as a basis for a decision, until a deliberate product change promotes it. Tier 3: Research only until proven.
The backtest
engine v14 · starters only · walk-forwardSkill is how much lower the Brier score of the served number is than always guessing the league-wide rate known at the time. Every date is predicted using only earlier dates, with calibration curves refit each date the way the site serves them.
Model better
Compared with the hitter's own record, shrunk toward the league. Difference in log loss, model minus player: -0.0032 (95% interval -0.0046 to -0.0018). Negative means the model is more accurate.
1,473 lineup hitters who batted · 26 dates
Counted from the ledgers at this build: predictions frozen before first pitch, under the engine now serving, with the game final and the result in. A hitter who never batted is void and left out. The count restarts with every engine version. This market is research only by policy, so the count changes no status.
The last engine with a finished live sample (engine v8, 2,808 results over 12 dates) showed skill of 2.4% (1.2% to 3.6%).
No clear difference
Log loss, model minus the no-vig closing price: +0.007 (95% interval +0.001 to +0.016; across dates, -0.004 to +0.020) on 867 predictions over 5 dates, engine v8. Positive means the book was more accurate. This comparison changes no status; it answers whether the number is a reason to bet.