Nightly self-improvement · slate 2026-09-04
Generated 2026-09-05T02:00:00.006202-04:00 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
2026-09-04
status completed
Winners
10/16
Brier 0.2315
Learned or changed
0
of 5 stages
Blocked
0
wanted to act and could not
Release
v4.18.0
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 |
|---|---|---|---|---|---|---|
| ARI@HOU | 4-6 | 1-3 | 0.583 | hit | Merrill Kelly: 15 proj / 18 real outs; Cristian Javier: 13 proj / 15 real outs | |
| ATL@PHI | 4-3 | 5-2 | 0.440 | hit | Chris Sale: 19 proj / 18 real outs; Cristopher Sánchez: 19 proj / 18 real outs | |
| BOS@BAL | 4-6 | 1-0 | 0.530 | miss | Run environment far lower than projected (over-dispersed) | Ranger Suarez: 15 proj / 21 real outs; Shane Baz: 17 proj / 19 real outs |
| CHC@MIA | 6-3 | 6-1 | 0.275 | hit | Shota Imanaga: 16 proj / 18 real outs; Janson Junk: 15 proj / 15 real outs | |
| DET@CLE | 3-4 | 6-7 | 0.525 | hit | Keider Montero: 18 proj / 17 real outs; Logan Allen: 9 proj / 18 real outs | |
| DET@CLE | 3-4 | 3-4 | 0.554 | hit | ||
| LAA@PIT | 5-7 | 0-1 | 0.662 | hit | Ryan Johnson: 15 proj / 18 real outs; Jared Jones: 15 proj / 18 real outs | |
| MIL@CIN | 8-5 | 10-7 | 0.460 | hit | Shane Drohan: 16 proj / 12 real outs; Rhett Lowder: 15 proj / 12 real outs | |
| MIN@CHW | 6-5 | 1-4 | 0.462 | miss | Starting-pitcher projection was wrong | Zebby Matthews: 18 proj / 15 real outs; Erick Fedde: 14 proj / 15 real outs |
| NYY@SD | 6-3 | 2-3 | 0.357 | miss | Backed the wrong side with real conviction | Max Fried: 17 proj / 17 real outs; Walker Buehler: 16 proj / 15 real outs |
| OAK@SEA | 5-6 | 7-6 | 0.548 | miss | Starting-pitcher projection was wrong | Kade Morris: 9 proj / 8 real outs; Logan Gilbert: 17 proj / 14 real outs |
| SF@NYM | 4-5 | 6-10 | 0.547 | hit | Matt Wilkinson: 12 proj / 9 real outs; Nolan McLean: 18 proj / 18 real outs | |
| STL@COL | 6-5 | 7-6 | 0.452 | hit | Andre Pallante: 18 proj / 15 real outs; Ryan Feltner: 14 proj / 15 real outs | |
| TB@TEX | 3-4 | 7-6 | 0.560 | miss | Run environment far higher than projected (under-dispersed) | Nick Martinez: 18 proj / 19 real outs; Trevor Williams: 14 proj / 3 real outs |
| TOR@KC | 5-7 | 9-2 | 0.557 | miss | Starting-pitcher projection was wrong | Jameson Taillon: 14 proj / 6 real outs; Daniel Lynch IV: 15 proj / 15 real outs |
| WAS@LAD | 4-5 | 3-5 | 0.561 | hit | Jackson Kent: 13 proj / 15 real outs; Blake Snell: 16 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 | 0.2520 | 0.2474 | -0.0046 | improved |
| accuracy | 0.5500 | 0.5833 | 0.0333 | improved |
| team_run_mae | 2.7667 | 2.6792 | -0.0875 | improved |
| total_run_mae | 3.8500 | 3.7083 | -0.1417 | improved |
| calibration_error | 0.0989 | 0.0846 | -0.0143 | improved |
| starter_er_mae | 1.9192 | 1.9386 | 0.0194 | regressed |
| starter_outs_mae | 3.0192 | 2.8114 | -0.2078 | 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
Luna (gpt-5.6-luna) post-mortem, status received. Full analysis.
# 1. Evidence receipt and gaps ## Source manifest Six of seven declared artifacts were loaded: | Source | Status | Purpose | |---|---:|---| | `prediction_review_20260904.json` — 399,800 bytes | Loaded | Official outcomes and scoring errors | | `PREDICTION_SYSTEM_DEEP_AUDIT_20260904.json` — 255,680 bytes | Loaded | Probability, provenance, starter, and decision contracts | | `continuous_learning_daily_20260904.json` — 2,555,562 bytes | Loaded | Pipeline quality, calibration, probability lineage, score distribution, and starter comparisons | | `detailed_20260904.html` — 380,084 bytes | Loaded | Full-game and simulation rendering | | `20260904_ultra_detailed_player_audit.html` — 33,148,291 bytes | Loaded | Player-level prediction trace | | `audit_report_20260904.html` — 1,949,987 bytes | Loaded | Invariant findings | | `audit_report_20260904_integrity_revision.html` | **Missing** | Post-Luna remediation proof for immutable revision artifacts | The missing integrity-revision artifact is a receipt gap, but the supplied data does not attribute any game-level prediction error to an artifact defect. Every game’s structural audit passed. ## Slate coverage All 16 listed games were reviewed: ARI@HOU, ATL@PHI, BOS@BAL, CHC@MIA, DET@CLE Game 1, DET@CLE Game 2, LAA@PIT, MIL@CIN, MIN@CHW, NYY@SD, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC, WAS@LAD. Coverage claims: - Expected games: 16. - Games reviewed: 16. - Structural audits: 16/16. - Canonical artifacts loaded: 16. - Continuous-l