Nightly self-improvement · slate 2026-09-19
Generated 2026-09-20T05:00:00.006584-04:00 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
2026-09-19
status completed
Winners
11/15
Brier 0.2507
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@HOU | 4-3 | 6-3 | 0.438 | hit | Grant Holmes: 15 proj / 18 real outs; Hayden Wesneski: 16 proj / 16 real outs | |
| BOS@TB | 4-5 | 1-2 | 0.529 | hit | Brayan Bello: 13 proj / 24 real outs; Freddy Peralta: 17 proj / 17 real outs | |
| CHC@CIN | 7-4 | 5-2 | 0.440 | hit | Matthew Boyd: 18 proj / 15 real outs; Nick Lodolo: 14 proj / 21 real outs | |
| DET@CHW | 6-4 | 1-3 | 0.440 | miss | Run environment far lower than projected (over-dispersed) | Jackson Jobe: 15 proj / 18 real outs; Sean Burke: 15 proj / 18 real outs |
| KC@PIT | 4-5 | 5-6 | 0.541 | hit | Noah Cameron: 17 proj / 18 real outs; Bubba Chandler: 16 proj / 12 real outs | |
| MIA@SD | 5-8 | 6-7 | 0.580 | hit | Eury Pérez: 15 proj / 14 real outs; Casey Mize: 15 proj / 9 real outs | |
| MIL@BAL | 4-5 | 1-0 | 0.530 | miss | Run environment far lower than projected (over-dispersed) | Robert Gasser: 15 proj / 21 real outs; Trevor Rogers: 18 proj / 24 real outs |
| MIN@LAA | 3-4 | 5-6 | 0.554 | hit | Joe Ryan: 15 proj / 15 real outs; Reid Detmers: 18 proj / 15 real outs | |
| NYY@ARI | 5-2 | 3-5 | 0.289 | miss | Backed the wrong side with real conviction | Cam Schlittler: 19 proj / 18 real outs; Brandon Pfaadt: 19 proj / 13 real outs |
| OAK@CLE | 4-6 | 6-12 | 0.580 | hit | Jacob Lopez: 14 proj / 8 real outs; Tanner Bibee: 17 proj / 14 real outs | |
| PHI@NYM | 5-6 | 3-10 | 0.529 | hit | Andrew Painter: 16 proj / 16 real outs; Christian Scott: 15 proj / 14 real outs | |
| SEA@COL | 10-5 | 3-5 | 0.279 | miss | Starting-pitcher projection was wrong | Bryce Miller: 15 proj / 16 real outs; Jose Quintana: 13 proj / 12 real outs |
| SF@LAD | 1-7 | 4-10 | 0.565 | hit | Yunior Marte: 15 proj / 14 real outs; Tarik Skubal: 19 proj / 15 real outs | |
| TOR@TEX | 2-4 | 2-6 | 0.611 | hit | José Soriano: 18 proj / 15 real outs; Cal Quantrill: 18 proj / 4 real outs | |
| WAS@STL | 5-4 | 8-5 | 0.457 | hit | Andrew Alvarez: 17 proj / 18 real outs; Michael McGreevy: 17 proj / 12 real outs |
Day over day
Rolling evidence window against the prior night. Associations, not causal proof.
| Metric | Prior | Current | Change | Result |
|---|---|---|---|---|
| brier_score | 0.2413 | 0.2434 | 0.0021 | regressed |
| accuracy | 0.5833 | 0.6083 | 0.0250 | improved |
| team_run_mae | 2.4333 | 2.3917 | -0.0416 | improved |
| total_run_mae | 3.2333 | 3.2833 | 0.0500 | regressed |
| calibration_error | 0.0572 | 0.0875 | 0.0303 | regressed |
| starter_er_mae | 1.6860 | 1.6653 | -0.0207 | improved |
| starter_outs_mae | 3.3721 | 3.2782 | -0.0939 | improved |
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.