Nightly self-improvement · slate 2026-09-08
Generated 2026-09-09T05:00:00.008565-04:00 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
2026-09-08
status completed
Winners
11/15
Brier 0.2188
Learned or changed
0
of 5 stages
Blocked
0
wanted to act and could not
Release
v4.26.0
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@KC | 5-3 | 5-3 | 0.426 | hit | Corbin Burnes: 4 proj / 9 real outs; Michael Wacha: 19 proj / 18 real outs | |
| CHC@MIL | 4-6 | 3-4 | 0.540 | hit | David Peterson: 12 proj / 13 real outs; Jacob Misiorowski: 18 proj / 19 real outs | |
| CIN@LAD | 3-9 | 2-3 | 0.672 | hit | Nick Lodolo: 12 proj / 14 real outs; Tarik Skubal: 18 proj / 22 real outs | |
| CLE@BAL | 5-4 | 9-5 | 0.495 | hit | Tanner Bibee: 17 proj / 13 real outs; Brandon Young: 17 proj / 11 real outs | |
| COL@NYY | 2-8 | 3-5 | 0.645 | hit | Gabriel Hughes: 13 proj / 15 real outs; Cam Schlittler: 19 proj / 21 real outs | |
| HOU@PHI | 3-2 | 6-5 | 0.446 | hit | Hayden Wesneski: 17 proj / 18 real outs; Andrew Painter: 16 proj / 15 real outs | |
| LAA@BOS | 3-4 | 6-1 | 0.527 | miss | Small edge on the losing side | Reid Detmers: 19 proj / 18 real outs; Patrick Sandoval: 15 proj / 18 real outs |
| MIN@DET | 5-6 | 3-2 | 0.540 | miss | Starting-pitcher projection was wrong | Dean Kremer: 14 proj / 15 real outs; Drew Anderson: 15 proj / 16 real outs |
| NYM@MIA | 6-5 | 7-5 | 0.466 | hit | Sean Manaea: 19 proj / 15 real outs; Sandy Alcantara: 19 proj / 13 real outs | |
| PIT@CHW | 3-4 | 9-3 | 0.519 | miss | Starting-pitcher projection was wrong | Bubba Chandler: 15 proj / 18 real outs; Sean Burke: 17 proj / 4 real outs |
| STL@SF | 6-5 | 1-2 | 0.474 | miss | Run environment far lower than projected (over-dispersed) | Quinn Mathews: 15 proj / 18 real outs; Landen Roupp: 17 proj / 18 real outs |
| TB@ATL | 6-5 | 7-1 | 0.478 | hit | Freddy Peralta: 17 proj / 18 real outs; AJ Smith-Shawver: 13 proj / 10 real outs | |
| TEX@SEA | 5-4 | 10-5 | 0.475 | hit | Cal Quantrill: 17 proj / 18 real outs; Bryce Miller: 17 proj / 11 real outs | |
| TOR@OAK | 5-4 | 4-2 | 0.440 | hit | José Soriano: 17 proj / 21 real outs; Jack Perkins: 12 proj / 15 real outs | |
| WAS@SD | 4-5 | 4-5 | 0.521 | hit | Riley Cornelio: 11 proj / 9 real outs; Casey Mize: 15 proj / 17 real outs |
Day over day
Rolling evidence window against the prior night. Associations, not causal proof.
| Metric | Prior | Current | Change | Result |
|---|---|---|---|---|
| brier_score | 0.2572 | 0.2542 | -0.0031 | improved |
| accuracy | 0.5333 | 0.5500 | 0.0167 | improved |
| team_run_mae | 2.5958 | 2.5792 | -0.0166 | improved |
| total_run_mae | 3.6750 | 3.8917 | 0.2167 | regressed |
| calibration_error | 0.0900 | 0.0641 | -0.0259 | improved |
| starter_er_mae | 1.9958 | 2.0207 | 0.0249 | regressed |
| starter_outs_mae | 3.0586 | 3.0083 | -0.0503 | 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.