Nightly self-improvement · slate 2026-09-15
Generated 2026-09-16T05:00:00.010328-04:00 ET from official finals (mlb-api). Every learning stage is listed with what it did. "Blocked" means it wanted to act and a safety rule stopped it; "nothing to learn" means it had no eligible evidence.
Slate
2026-09-15
status completed
Winners
6/15
Brier 0.2476
Learned or changed
0
of 5 stages
Blocked
0
wanted to act and could not
Release
v4.31.0
DRIFT
What each learning stage did
| Stage | Outcome | Detail |
|---|---|---|
| KPI settlement | failed | did not run |
| Bullpen usage model | nothing to learn | No pregame bullpen forecasts to settle. Refit: not attempted. |
| Adaptive parameter proposal (Luna) | nothing to learn | skipped_open_rollback_monitor: |
| Auto-apply | nothing to learn | no_promotable_candidate |
| Code agent | nothing to learn | disabled_by_config: code changes are made on the development host, never in the container. |
Games
| Game | Predicted | Actual | Home win prob | Result | Main driver | Starters |
|---|---|---|---|---|---|---|
| ATL@CHC | 4-5 | 6-3 | 0.535 | miss | Small edge on the losing side | Martín Pérez: 17 proj / 18 real outs; Kevin Gausman: 16 proj / 17 real outs |
| BAL@NYM | 4-5 | 7-5 | 0.530 | miss | Several errors of similar size | Shane Baz: 18 proj / 12 real outs; Sean Manaea: 17 proj / 16 real outs |
| BOS@TEX | 7-6 | 2-4 | 0.485 | miss | Run environment far lower than projected (over-dispersed) | Patrick Sandoval: 15 proj / 8 real outs; Jacob deGrom: 17 proj / 15 real outs |
| CHW@CLE | 5-4 | 6-7 | 0.455 | miss | Small edge on the losing side | Chris Murphy: 14 proj / 2 real outs; Foster Griffin: 16 proj / 12 real outs |
| DET@TOR | 5-4 | 10-1 | 0.438 | hit | Drew Anderson: 15 proj / 15 real outs; Braydon Fisher: 15 proj / 4 real outs | |
| KC@HOU | 3-4 | 2-4 | 0.530 | hit | Michael Wacha: 19 proj / 18 real outs; Hunter Brown: 17 proj / 15 real outs | |
| LAD@CIN | 9-5 | 4-0 | 0.420 | hit | Yoshinobu Yamamoto: 21 proj / 21 real outs; Rhett Lowder: 14 proj / 15 real outs | |
| MIA@ARI | 4-6 | 4-2 | 0.547 | miss | Small edge on the losing side | Janson Junk: 14 proj / 9 real outs; Michael Soroka: 15 proj / 15 real outs |
| MIL@PIT | 5-2 | 5-1 | 0.440 | hit | Jacob Misiorowski: 18 proj / 15 real outs; Lake Bachar: 15 proj / 6 real outs | |
| NYY@MIN | 6-3 | 8-1 | 0.430 | hit | Max Fried: 18 proj / 21 real outs; Bailey Ober: 15 proj / 16 real outs | |
| OAK@TB | 4-8 | 1-2 | 0.667 | hit | Jack Perkins: 12 proj / 20 real outs; Griffin Jax: 15 proj / 15 real outs | |
| PHI@WAS | 6-5 | 3-6 | 0.467 | miss | Starting-pitcher projection was wrong | Cristopher Sánchez: 19 proj / 19 real outs; Jackson Kent: 13 proj / 15 real outs |
| SD@COL | 9-5 | 3-9 | 0.420 | miss | Starting-pitcher projection was wrong | Walker Buehler: 16 proj / 14 real outs; Kyle Freeland: 14 proj / 12 real outs |
| SEA@LAA | 5-4 | 1-2 | 0.470 | miss | Starting-pitcher projection was wrong | Logan Gilbert: 17 proj / 18 real outs; Ryan Johnson: 16 proj / 18 real outs |
| SF@STL | 4-5 | 10-3 | 0.553 | miss | Backed the wrong side with real conviction | Blade Tidwell: 15 proj / 20 real outs; Andre Pallante: 17 proj / 15 real outs |
Day over day
Rolling evidence window against the prior night. Associations, not causal proof.
| Metric | Prior | Current | Change | Result |
|---|---|---|---|---|
| brier_score | 0.2411 | 0.2426 | 0.0014 | regressed |
| accuracy | 0.6333 | 0.6000 | -0.0333 | regressed |
| team_run_mae | 2.5083 | 2.5542 | 0.0459 | regressed |
| total_run_mae | 3.8167 | 3.7917 | -0.0250 | improved |
| calibration_error | 0.1244 | 0.0840 | -0.0404 | improved |
| starter_er_mae | 1.7778 | 1.7739 | -0.0039 | improved |
| starter_outs_mae | 3.2644 | 3.2874 | 0.0230 | regressed |
Calibration produced for the next slate
| Metric | Scale used | Measured | Rows | Status |
|---|---|---|---|---|
| No calibration produced. | ||||
Luna KPI candidates and their promotion gate
Promotion needs 400+ prospective rows over 5+ slates after the proposal, with a gain interval above zero. Only verified pregame forecasts count; local replays never do.
| KPI | Status | Target | Proposed | Rows | Slates | Gain interval | Progress |
|---|---|---|---|---|---|---|---|
| luna_pitcher_release_speed_vs_righthanded_starters | shadow | SO | 20260923 | 330/400 | 5/5 | [-0.0007, 0.0007] | |
| luna_pitcher_barrel_rate_vs_rhh_righthanded_starters | shadow | HR | 20260923 | 330/400 | 5/5 | [-0.0004, 0.0049] | |
| luna_batter_contact_rate_vs_breaking_righthanded_starters | shadow | H | 20260923 | 330/400 | 5/5 | [-0.0045, 0.0044] | |
| luna_batter_contact_rate_vs_breaking_left_handed_starters_for_strikeouts | shadow | SO | 20260924 | 122/400 | 4/5 | [-0.0027, 0.0021] | |
| luna_pitcher_contact_xwoba_allowed_vs_righthanded_starters_for_strikeouts | shadow | SO | 20260924 | 122/400 | 4/5 | [-0.0001, 0.0000] | |
| luna_pitcher_whiff_rate_breaking_vs_left_handed_hitters_for_hits | shadow | H | 20260924 | 122/400 | 4/5 | [-0.0056, -0.0011] | |
| luna_pitcher_home_run_rate_vs_righthanded_starters | shadow | HR | 20260925 | 122/400 | 4/5 | [-0.0010, 0.0056] | |
| luna_batter_ground_ball_rate_vs_fastball_vs_lhp_for_home_runs | shadow | HR | 20260926 | 86/400 | 3/5 | [-0.0001, -0.0000] | |
| luna_pitcher_whiff_rate_vs_right_handed_starters_for_batter_strikeouts | shadow | SO | 20260929 | 51/400 | 2/5 | [–, –] | |
| luna_batter_whiff_rate_vs_breaking_right_handed_starters_for_hits | shadow | H | 20260929 | 51/400 | 2/5 | [–, –] | |
| luna_batter_walk_rate_vs_righthanded_pitchers_for_home_runs | retired | HR | 20260924 | 0/400 | 0/5 | [–, –] |
Operator actions the cycle asked for (0)
Changes to code or a model recipe are never made by the nightly cycle. These are the measured findings that need one.
| Priority | Owner | Finding | Action |
|---|---|---|---|
| No operator actions requested. | |||
Luna post-mortem
No Luna post-mortem this night.