Nightly self-improvement · slate 2026-09-20
Generated 2026-09-21T05:00:00.009094-04:00 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
2026-09-20
status completed
Winners
9/15
Brier 0.2209
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 | candidate_registered: |
| 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 | 5-3 | 4-2 | 0.420 | hit | Martín Pérez: 17 proj / 13 real outs; Hunter Brown: 17 proj / 15 real outs | |
| BOS@TB | 4-5 | 1-5 | 0.552 | hit | Patrick Sandoval: 14 proj / 21 real outs; Griffin Jax: 15 proj / 15 real outs | |
| CHC@CIN | 8-4 | 9-1 | 0.440 | hit | David Peterson: 15 proj / 18 real outs; Rhett Lowder: 13 proj / 9 real outs | |
| DET@CHW | 5-4 | 1-8 | 0.440 | miss | Backed the wrong side with real conviction | Troy Melton: 18 proj / 11 real outs; Davis Martin: 15 proj / 13 real outs |
| KC@PIT | 2-3 | 3-4 | 0.573 | hit | Michael Wacha: 19 proj / 17 real outs; Lake Bachar: 15 proj / 3 real outs | |
| MIA@SD | 4-6 | 3-7 | 0.580 | hit | Sandy Alcantara: 16 proj / 18 real outs; Walker Buehler: 14 proj / 20 real outs | |
| MIL@BAL | 7-1 | 3-0 | 0.371 | hit | Jacob Misiorowski: 18 proj / 9 real outs; Brandon Young: 17 proj / 18 real outs | |
| MIN@LAA | 4-5 | 8-0 | 0.505 | miss | Genuine toss-up that flipped | Dean Kremer: 14 proj / 18 real outs; Ryan Johnson: 17 proj / 14 real outs |
| NYY@ARI | 6-4 | 4-8 | 0.455 | miss | Small edge on the losing side | Will Warren: 16 proj / 13 real outs; Corbin Burnes: 12 proj / 8 real outs |
| OAK@CLE | 4-7 | 0-1 | 0.580 | hit | Jack Perkins: 14 proj / 18 real outs; Gavin Williams: 16 proj / 21 real outs | |
| PHI@NYM | 4-5 | 7-2 | 0.517 | miss | Genuine toss-up that flipped | Cristopher Sánchez: 19 proj / 19 real outs; Jonah Tong: 16 proj / 14 real outs |
| SEA@COL | 9-6 | 2-1 | 0.440 | hit | Kade Anderson: 15 proj / 21 real outs; Tomoyuki Sugano: 14 proj / 15 real outs | |
| SF@LAD | 4-8 | 1-3 | 0.580 | hit | Matt Wilkinson: 13 proj / 15 real outs; Jack Dreyer: 15 proj / 3 real outs | |
| TOR@TEX | 2-4 | 7-2 | 0.550 | miss | Backed the wrong side with real conviction | Spencer Miles: 18 proj / 9 real outs; Jacob deGrom: 17 proj / 13 real outs |
| WAS@STL | 6-5 | 3-5 | 0.488 | miss | Genuine toss-up that flipped | Jake Irvin: 15 proj / 16 real outs; Quinn Mathews: 17 proj / 21 real outs |
Day over day
Rolling evidence window against the prior night. Associations, not causal proof.
| Metric | Prior | Current | Change | Result |
|---|---|---|---|---|
| brier_score | 0.2434 | 0.2393 | -0.0041 | improved |
| accuracy | 0.6083 | 0.6000 | -0.0083 | regressed |
| team_run_mae | 2.3917 | 2.5042 | 0.1125 | regressed |
| total_run_mae | 3.2833 | 3.3583 | 0.0750 | regressed |
| calibration_error | 0.0875 | 0.0885 | 0.0010 | regressed |
| starter_er_mae | 1.6653 | 1.6855 | 0.0202 | regressed |
| starter_outs_mae | 3.2782 | 3.4839 | 0.2057 | 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.