Nightly self-improvement · slate 2026-09-14

Generated 2026-09-15T05:00:00.007175-04:00 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

2026-09-14

status completed

Winners

7/10

Brier 0.2394

Learned or changed

of 5 stages

Blocked

wanted to act and could not

Release

v4.31.0

DRIFT

Audit Detailed Deep Slate Review Luna Post-mortem Daily Learning Adaptive Status

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@CHC	5-6	3-7	0.524	hit		Reynaldo López: 14 proj / 9 real outs; David Peterson: 13 proj / 18 real outs
BAL@NYM	4-7	2-1	0.628	miss	Run environment far lower than projected (over-dispersed)	Brandon Young: 16 proj / 15 real outs; Jonah Tong: 15 proj / 18 real outs
CHW@CLE	4-5	7-3	0.506	miss	Genuine toss-up that flipped	Sean Newcomb: 15 proj / 3 real outs; Gavin Williams: 16 proj / 15 real outs
DET@TOR	5-4	6-5	0.494	hit		Troy Melton: 18 proj / 8 real outs; José Soriano: 17 proj / 19 real outs
LAD@CIN	8-3	4-1	0.420	hit		Tarik Skubal: 20 proj / 21 real outs; Nick Lodolo: 13 proj / 21 real outs
MIA@ARI	5-4	7-8	0.483	miss	Run environment far higher than projected (under-dispersed)	Sandy Alcantara: 18 proj / 15 real outs; Corbin Burnes: 9 proj / 14 real outs
NYY@MIN	6-4	8-3	0.430	hit		Will Warren: 16 proj / 17 real outs; Dean Kremer: 14 proj / 17 real outs
SD@COL	8-5	8-7	0.420	hit		Casey Mize: 15 proj / 15 real outs; Tomoyuki Sugano: 14 proj / 15 real outs
SEA@LAA	5-6	4-6	0.528	hit		Kade Anderson: 15 proj / 13 real outs; Reid Detmers: 19 proj / 18 real outs
SF@STL	4-5	1-2	0.505	hit		Landen Roupp: 16 proj / 13 real outs; Quinn Mathews: 15 proj / 20 real outs

Day over day

Rolling evidence window against the prior night. Associations, not causal proof.

Metric	Prior	Current	Change	Result
brier_score	0.2425	0.2411	-0.0014	improved
accuracy	0.6083	0.6333	0.0250	improved
team_run_mae	2.5125	2.5083	-0.0042	improved
total_run_mae	3.7917	3.8167	0.0250	regressed
calibration_error	0.0977	0.1244	0.0267	regressed
starter_er_mae	1.7676	1.7778	0.0102	regressed
starter_outs_mae	3.2365	3.2644	0.0279	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
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 receipt

All seven source-manifest entries were reviewed as declared:

| Source | Status | Declared purpose |
|---|---:|---|
| `prediction_review_20260914.json` | Loaded | Official outcomes and scoring errors |
| `PREDICTION_SYSTEM_DEEP_AUDIT_20260914.json` | Loaded | Probability, provenance, starter, and decision contracts |
| `continuous_learning_daily_20260914.json` | Loaded | Stage quality, calibration, lineage, score distribution, and starter comparisons |
| `detailed_20260914.html` | Loaded | Full game and simulation rendering |
| `20260914_ultra_detailed_player_audit.html` | Loaded | Player-level prediction trace |
| `audit_report_20260914.html` | Loaded | Invariant findings |
| `audit_report_20260914_integrity_revision.html` | **Missing** | Post-Luna remediation proof for immutable revision artifacts |

The missing integrity-revision artifact is the only explicit source-manifest gap. It does not invalidate the game outcomes or structural audits, but it prevents verification of the claimed post-remediation immutable-revision proof.

## Slate coverage

The slate contained 10 games, and all 10 were reviewed:

`ATL@CHC`, `BAL@NYM`, `CHW@CLE`, `DET@TOR`, `LAD@CIN`, `MIA@ARI`, `NYY@MIN`, `SD@COL`, `SEA@LAA`, and `SF@STL`.

Coverage was complete for the ordinary prediction artifacts:

- Expected games: 10
- Canonical artifacts loaded: 10
- Continuous-learning games: 10
- Forensic deep-audit games: 10
- Structural audits completed: 10