Nightly self-improvement · slate 2026-09-05
Generated 2026-09-06T22:17:44.702357-04:00 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
2026-09-05
status completed
Winners
8/15
Brier 0.2424
Learned or changed
0
of 5 stages
Blocked
0
wanted to act and could not
Release
v4.25.0
consistent
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 |
|---|---|---|---|---|---|---|
| ARI@HOU | 5-4 | 4-3 | 0.483 | hit | Brandon Pfaadt: 19 proj / 19 real outs; Ethan Pecko: 12 proj / 13 real outs | |
| ATL@PHI | 5-4 | 2-4 | 0.462 | miss | Several errors of similar size | Martín Pérez: 17 proj / 11 real outs; Zack Wheeler: 16 proj / 18 real outs |
| BOS@BAL | 5-4 | 5-0 | 0.468 | hit | Sonny Gray: 18 proj / 21 real outs; Chris Bassitt: 15 proj / 17 real outs | |
| CHC@MIA | 6-3 | 6-5 | 0.260 | hit | Javier Assad: 15 proj / 14 real outs; Ryan Gusto: 13 proj / 14 real outs | |
| DET@CLE | 3-4 | 6-0 | 0.554 | miss | Backed the wrong side with real conviction | Framber Valdez: 16 proj / 21 real outs; Parker Messick: 18 proj / 18 real outs |
| LAA@PIT | 4-6 | 6-1 | 0.636 | miss | Starting-pitcher projection was wrong | Yusei Kikuchi: 14 proj / 18 real outs; Braxton Ashcraft: 19 proj / 15 real outs |
| MIL@CIN | 8-5 | 3-5 | 0.460 | miss | Run environment far lower than projected (over-dispersed) | Dustin May: 15 proj / 11 real outs; Andrew Abbott: 15 proj / 12 real outs |
| MIN@CHW | 6-5 | 6-4 | 0.470 | hit | Taj Bradley: 17 proj / 15 real outs; Anthony Kay: 16 proj / 13 real outs | |
| NYY@SD | 4-3 | 5-1 | 0.451 | hit | Carlos Rodón: 17 proj / 15 real outs; Robbie Ray: 17 proj / 7 real outs | |
| OAK@SEA | 6-8 | 6-2 | 0.530 | miss | Starting-pitcher projection was wrong | Jeffrey Springs: 12 proj / 21 real outs; George Kirby: 16 proj / 20 real outs |
| SF@NYM | 4-5 | 9-5 | 0.524 | miss | Starting-pitcher projection was wrong | Anthony Molina: 9 proj / 15 real outs; Zac Thornton: 17 proj / 15 real outs |
| STL@COL | 7-6 | 7-10 | 0.491 | miss | Genuine toss-up that flipped | Matthew Liberatore: 16 proj / 13 real outs; Mason Adams: 13 proj / 14 real outs |
| TB@TEX | 6-5 | 6-3 | 0.467 | hit | Drew Rasmussen: 18 proj / 21 real outs; Jacob deGrom: 15 proj / 18 real outs | |
| TOR@KC | 7-5 | 4-3 | 0.430 | hit | Max Scherzer: 15 proj / 17 real outs; Seth Lugo: 15 proj / 15 real outs | |
| WAS@LAD | 3-4 | 5-6 | 0.573 | hit | Cade Cavalli: 17 proj / 18 real outs; Tyler Glasnow: 17 proj / 18 real outs |
Day over day
Rolling evidence window against the prior night. Associations, not causal proof.
| Metric | Prior | Current | Change | Result |
|---|---|---|---|---|
| brier_score | 0.2474 | 0.2459 | -0.0015 | improved |
| accuracy | 0.5833 | 0.5833 | 0.0000 | flat |
| team_run_mae | 2.6792 | 2.5917 | -0.0875 | improved |
| total_run_mae | 3.7083 | 3.6167 | -0.0916 | improved |
| calibration_error | 0.0846 | 0.0868 | 0.0022 | regressed |
| starter_er_mae | 1.9386 | 1.9426 | 0.0040 | regressed |
| starter_outs_mae | 2.8114 | 2.8811 | 0.0697 | 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.