Nightly self-improvement · slate 2026-09-11
Generated 2026-09-12T05:00:00.009283-04:00 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
2026-09-11
status completed
Winners
10/15
Brier 0.2407
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 |
|---|---|---|---|---|---|---|
| BAL@TOR | 4-5 | 7-4 | 0.505 | miss | Genuine toss-up that flipped | Chris Bassitt: 15 proj / 15 real outs; Max Scherzer: 15 proj / 15 real outs |
| CHW@STL | 4-6 | 3-7 | 0.598 | hit | Anthony Kay: 15 proj / 12 real outs; Matthew Liberatore: 15 proj / 13 real outs | |
| CIN@MIL | 6-8 | 0-20 | 0.540 | hit | Andrew Abbott: 14 proj / 6 real outs; Dustin May: 14 proj / 18 real outs | |
| CLE@MIN | 4-5 | 5-2 | 0.505 | miss | Genuine toss-up that flipped | Parker Messick: 18 proj / 16 real outs; Taj Bradley: 17 proj / 18 real outs |
| COL@DET | 4-5 | 2-6 | 0.541 | hit | Mason Adams: 13 proj / 15 real outs; Framber Valdez: 18 proj / 18 real outs | |
| HOU@TB | 3-5 | 1-3 | 0.505 | hit | Miguel Ullola: 9 proj / 3 real outs; Drew Rasmussen: 18 proj / 21 real outs | |
| KC@BOS | 4-6 | 3-2 | 0.653 | miss | Run environment far lower than projected (over-dispersed) | Seth Lugo: 15 proj / 15 real outs; Sonny Gray: 18 proj / 18 real outs |
| LAA@WAS | 4-5 | 3-4 | 0.576 | hit | Yusei Kikuchi: 15 proj / 12 real outs; Cade Cavalli: 17 proj / 18 real outs | |
| LAD@MIA | 6-4 | 6-2 | 0.420 | hit | Blake Snell: 17 proj / 17 real outs; Ryan Gusto: 13 proj / 18 real outs | |
| NYM@NYY | 4-6 | 4-6 | 0.530 | hit | Nolan McLean: 18 proj / 15 real outs; Carlos Rodón: 17 proj / 19 real outs | |
| PHI@ATL | 3-4 | 5-6 | 0.527 | hit | Aaron Nola: 17 proj / 19 real outs; Chris Sale: 19 proj / 21 real outs | |
| PIT@CHC | 4-5 | 2-12 | 0.530 | hit | Wilber Dotel: 15 proj / 6 real outs; Shota Imanaga: 16 proj / 18 real outs | |
| SD@SF | 5-4 | 7-5 | 0.495 | hit | Robbie Ray: 15 proj / 15 real outs; Anthony Molina: 15 proj / 9 real outs | |
| SEA@OAK | 7-6 | 5-6 | 0.494 | miss | Genuine toss-up that flipped | George Kirby: 17 proj / 18 real outs; Jeffrey Springs: 14 proj / 15 real outs |
| TEX@ARI | 6-4 | 1-9 | 0.432 | miss | Starting-pitcher projection was wrong | MacKenzie Gore: 15 proj / 12 real outs; Merrill Kelly: 16 proj / 13 real outs |
Day over day
Rolling evidence window against the prior night. Associations, not causal proof.
| Metric | Prior | Current | Change | Result |
|---|---|---|---|---|
| brier_score | 0.2445 | 0.2421 | -0.0024 | improved |
| accuracy | 0.5750 | 0.6000 | 0.0250 | improved |
| team_run_mae | 2.5708 | 2.4417 | -0.1291 | improved |
| total_run_mae | 4.0750 | 3.7167 | -0.3583 | improved |
| calibration_error | 0.0513 | 0.0872 | 0.0359 | regressed |
| starter_er_mae | 1.9728 | 1.8716 | -0.1012 | improved |
| starter_outs_mae | 3.2062 | 3.1284 | -0.0778 | 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 ### Sources received The manifest contains seven expected artifacts: | Source | Status | Size | Purpose | |---|---:|---:|---| | `prediction review` | Loaded | 377,381 bytes | Official outcomes and scoring errors | | `forensic deep audit` | Loaded | 240,613 bytes | Probability, provenance, starter, and decision contracts | | `continuous learning` | Loaded | 2,544,895 bytes | Stage quality, calibration, lineage, score distribution, starter comparisons | | `detailed report` | Loaded | 352,194 bytes | Full game and simulation rendering | | `ultra detailed player audit` | Loaded | 32,629,527 bytes | Player-level prediction trace | | `structural audit` | Loaded | 1,840,289 bytes | Invariant findings | | `integrity revision audit` | **Missing** | — | Post-remediation proof for immutable revision artifacts | The missing integrity-revision audit is the only explicit source gap. It prevents verification of post-Luna immutable-revision artifacts, but the game-level structural audits report clean saved artifacts. ### Slate coverage All 15 expected games were reviewed: `BAL@TOR`, `CHW@STL`, `CIN@MIL`, `CLE@MIN`, `COL@DET`, `HOU@TB`, `KC@BOS`, `LAA@WAS`, `LAD@MIA`, `NYM@NYY`, `PHI@ATL`, `PIT@CHC`, `SD@SF`, `SEA@OAK`, and `TEX@ARI`. Coverage is complete: - Canonical artifacts loaded: **15/15** - Continuous-learning games: **15/15** - Forensic deep-audit games: **15/15** - Starter comparisons: **15/15** - Structural audits: **15/15** - Overall structu