Nightly self-improvement · slate 2026-09-13
Generated 2026-09-14T05:00:00.007467-04:00 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
2026-09-13
status completed
Winners
9/15
Brier 0.2447
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 | 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 |
|---|---|---|---|---|---|---|
| BAL@TOR | 2-4 | 1-8 | 0.530 | hit | Trevor Rogers: 18 proj / 18 real outs; Dylan Cease: 18 proj / 15 real outs | |
| CHW@STL | 5-6 | 1-3 | 0.508 | hit | David Sandlin: 12 proj / 17 real outs; Michael McGreevy: 17 proj / 18 real outs | |
| CIN@MIL | 4-8 | 4-3 | 0.540 | miss | Run environment far lower than projected (over-dispersed) | Chase Burns: 18 proj / 9 real outs; Robert Gasser: 15 proj / 15 real outs |
| CLE@MIN | 4-5 | 9-2 | 0.546 | miss | Small edge on the losing side | Tanner Bibee: 16 proj / 20 real outs; Joe Ryan: 16 proj / 12 real outs |
| COL@DET | 4-7 | 1-8 | 0.646 | hit | Gabriel Hughes: 13 proj / 6 real outs; Jackson Jobe: 14 proj / 19 real outs | |
| HOU@TB | 3-4 | 4-14 | 0.520 | hit | Hayden Wesneski: 17 proj / 9 real outs; Freddy Peralta: 17 proj / 15 real outs | |
| KC@BOS | 2-4 | 1-4 | 0.560 | hit | Noah Cameron: 18 proj / 18 real outs; Payton Tolle: 18 proj / 18 real outs | |
| LAA@WAS | 5-6 | 5-6 | 0.571 | hit | Grayson Rodriguez: 14 proj / 16 real outs; Riley Cornelio: 12 proj / 6 real outs | |
| LAD@MIA | 8-5 | 4-6 | 0.400 | miss | Backed the wrong side with real conviction | Emmet Sheehan: 14 proj / 18 real outs; Eury Pérez: 15 proj / 15 real outs |
| NYM@NYY | 5-6 | 0-2 | 0.564 | hit | Christian Scott: 15 proj / 18 real outs; Cam Schlittler: 19 proj / 18 real outs | |
| PHI@ATL | 4-5 | 9-4 | 0.543 | miss | Small edge on the losing side | Andrew Painter: 16 proj / 15 real outs; Grant Holmes: 15 proj / 12 real outs |
| PIT@CHC | 3-4 | 4-3 | 0.519 | miss | Genuine toss-up that flipped | Bubba Chandler: 15 proj / 16 real outs; Matthew Boyd: 18 proj / 16 real outs |
| SD@SF | 6-4 | 6-4 | 0.461 | hit | Nick Pivetta: 15 proj / 12 real outs; Logan Webb: 16 proj / 15 real outs | |
| SEA@OAK | 6-5 | 7-8 | 0.440 | miss | Backed the wrong side with real conviction | Bryce Miller: 16 proj / 15 real outs; Jacob Lopez: 15 proj / 11 real outs |
| TEX@ARI | 4-3 | 7-6 | 0.494 | hit | Cal Quantrill: 18 proj / 18 real outs; Eduardo Rodriguez: 18 proj / 16 real outs |
Day over day
Rolling evidence window against the prior night. Associations, not causal proof.
| Metric | Prior | Current | Change | Result |
|---|---|---|---|---|
| brier_score | 0.2417 | 0.2425 | 0.0008 | regressed |
| accuracy | 0.6083 | 0.6083 | 0.0000 | flat |
| team_run_mae | 2.5042 | 2.5125 | 0.0083 | regressed |
| total_run_mae | 3.8417 | 3.7917 | -0.0500 | improved |
| calibration_error | 0.0961 | 0.0977 | 0.0016 | regressed |
| starter_er_mae | 1.8661 | 1.7676 | -0.0985 | improved |
| starter_outs_mae | 3.2134 | 3.2365 | 0.0231 | 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.