Deep slate diagnosis — 20260901

• Games: 15 · hits 7 · misses 8 · winner accuracy 0.4667
• Mean total-run bias on losses: -2.375 (negative = under-projected)
• Mean team-run error on losses: 7.125 · mean Brier on losses: 0.375

Loss primary drivers

• Small edge on the losing side: 1
• Run environment far higher than projected (under-dispersed): 2
• Starting-pitcher projection was wrong: 4
• Backed the wrong side with real conviction: 1

High-level causal issue register

[HIGH] Canonical artifact warning requires reconciliation

• Domain: artifact_integrity · affected: slate-wide
• Evidence: structural audit totals: {'critical': 0, 'error': 0, 'warning': 1, 'info': 108}
• Action: Fix the flagged stat or presentation contract before treating this slate as clean evidence.
• Do not infer: A warning flags a display or contract issue; it does not prove the model picked the wrong side.

[HIGH] Published margins are compressed relative to realized margins

• Domain: score_distribution · affected: CHW@HOU, DET@MIN, MIL@CHC, NYM@TB, PHI@ARI, STL@LAD, TOR@CLE
• Evidence: 13/15 predictions were ≤2 runs; 9/15 actual margins were ≥4; mean predicted margin 1.67 vs actual 4.20
• Action: Check whether predicted margins are too narrow across several recent out-of-sample slates (measuring tail coverage and margin MAE) before widening the score spread.
• Do not infer: A single blowout game is not a reason to widen every score prediction.

[HIGH] Published confidence buckets were poorly aligned on this slate

• Domain: calibration · affected: ATL@WAS, BAL@COL, CHW@HOU, DET@MIN, MIA@KC, MIL@CHC, NYM@TB, NYY@LAA, OAK@TEX, PHI@ARI, SD@CIN, SEA@BOS, SF@PIT, STL@LAD, TOR@CLE
• Evidence: weighted absolute calibration error 0.2476 across 15 games
• Action: Watch the same confidence buckets over the full rolling window before changing any calibration weight.
• Do not infer: A single slate has far too few games to establish a stable calibration curve.

[HIGH] Validated run models explain very little holdout variance

• Domain: model_validation · affected: ATL@WAS, BAL@COL, CHW@HOU, DET@MIN, MIA@KC, MIL@CHC, NYM@TB, NYY@LAA, OAK@TEX, PHI@ARI, SD@CIN, SEA@BOS, SF@PIT, STL@LAD, TOR@CLE
• Evidence: home/away/total R² = 0.0361/0.0227/0.0280
• Action: Treat exact run scores as low-confidence estimates, and look for stronger features or a better target before tuning the score tails.
• Do not infer: A small positive R² only means the model beats guessing the average; it does not make the exact scores reliable.

[MEDIUM] Most high-raw picks lost before/after shrinkage

• Domain: calibration · affected: BAL@COL, MIL@CHC, SD@CIN, STL@LAD
• Evidence: 0/4 games with raw pick probability ≥0.65 won
• Action: Track how often high-raw-probability picks lose across multiple slates, and keep the raw to calibrated to published lineage.
• Do not infer: This one slate is not enough to justify a new per-bucket shrinkage rule.

[INFO] Every settled game received structural QA

• Domain: coverage · affected: ATL@WAS, BAL@COL, CHW@HOU, DET@MIN, MIA@KC, MIL@CHC, NYM@TB, NYY@LAA, OAK@TEX, PHI@ARI, SD@CIN, SEA@BOS, SF@PIT, STL@LAD, TOR@CLE
• Evidence: 15/15 games structurally audited
• Action: Keep the per-game audit receipt in every report.
• Do not infer: Full audit coverage does not cancel the findings raised inside the audited games.

[INFO] Material calibrated-to-published movements have recorded lineage

• Domain: probability_lineage · affected: ATL@WAS, CHW@HOU, DET@MIN, MIA@KC, MIL@CHC, PHI@ARI, SD@CIN, SEA@BOS, STL@LAD
• Evidence: 9 material movement(s), all explained by named score-coherence and/or recency transformations
• Action: Keep showing every probability layer and the named transformation between them.
• Do not infer: A published probability above the calibrated one is not a governor bypass when a recorded transformation explains it.

[INFO] Simulation produced no winner-direction intervention on this slate

• Domain: simulation · affected: ATL@WAS, BAL@COL, CHW@HOU, DET@MIN, MIA@KC, MIL@CHC, NYM@TB, NYY@LAA, OAK@TEX, PHI@ARI, SD@CIN, SEA@BOS, SF@PIT, STL@LAD, TOR@CLE
• Evidence: flip candidates 0; published flips 0
• Action: Measure the simulation's contribution across rolling slates, separating score realization from winner flipping.
• Do not infer: Zero winner flips is not a defect and does not mean the simulation did no score-reconciliation work.

Cross-slate trends (persistent-pattern memory)

• Window: 20260828, 20260829, 20260830, 20260831, 20260901 (4 prior slates loaded)
• Highest-confidence pick loss streak: 1 slate(s)
• Actionable pattern detected: True

Escalated persistent patterns

• simulation_inert_on_slate — 5 consecutive slates (5 in window)
• Target lever(s): simulation.allow_score
• Direction: confirm the simulation module is running, not silently no-oping
• Guardrail: simulation_inert_on_slate has persisted 5 consecutive slates; this is no longer single-slate noise. Treat it as an actionable pattern and raise the code/diagnostic action named; do not force a parameter move.
• weak_run_model_signal — 5 consecutive slates (5 in window)
• Target lever(s): training.xgb.n_estimators, training.xgb.max_depth, training.xgb.learning_rate
• Direction: investigate features/target via isolated training backtest (manual)
• Guardrail: weak_run_model_signal has persisted 5 consecutive slates; this is no longer single-slate noise. Treat it as an actionable pattern and raise the code/diagnostic action named; do not force a parameter move.

Metric direction across the window

• winner_accuracy: regressing · series [0.4667, 0.9412, 0.5, 0.6667, 0.4667]
• weighted_calibration_error: regressing · series [0.2179, 0.3961, 0.0693, 0.1661, 0.2476]
• margin_bias: regressing · series [-1.333, -1.882, -2.786, -3.5, -2.533]
• mean_team_run_error_on_losses: regressing · series [5.0, 3.0, 7.0, 8.25, 7.125]

Pipeline stage quality

• historical_ml: winner 7/15 (0.4667), team-run MAE 3.0977, margin MAE 4.1769
• variance_applied: winner 6/15 (0.4), team-run MAE 3.0667, margin MAE 4.1333
• enhanced_prediction: winner 6/15 (0.4), team-run MAE 3.2667, margin MAE 4.9333
• embedded_savant_run_environment: winner 6/15 (0.4), team-run MAE 3.1667, margin MAE 4.8667
• fundamental_edge_target_review: winner 7/15 (0.4667), team-run MAE 3.1333, margin MAE 4.8
• capped_target: winner 7/15 (0.4667), team-run MAE 3.1333, margin MAE 4.8
• simulation_result: winner 7/15 (0.4667), team-run MAE 3.1333, margin MAE 4.8
• published_final: winner 7/15 (0.4667), team-run MAE 3.1333, margin MAE 4.8

Probability lineage

• ATL@WAS: raw 0.5085 → calibrated 0.5082 → published 0.5317; score coherence followed by matchup-recency guardrail
• BAL@COL: raw 0.7515 → calibrated 0.7373 → published 0.7461; score_coherence aligned probability to the published score margin
• CHW@HOU: raw 0.5085 → calibrated 0.5082 → published 0.53; score coherence followed by matchup-recency guardrail
• DET@MIN: raw 0.5032 → calibrated 0.505 → published 0.53; score coherence followed by matchup-recency guardrail
• MIA@KC: raw 0.6275 → calibrated 0.6227 → published 0.6464; score_coherence aligned probability to the published score margin
• MIL@CHC: raw 0.7934 → calibrated 0.6442 → published 0.6986; score_coherence aligned probability to the published score margin
• NYM@TB: raw 0.5085 → calibrated 0.5082 → published 0.5269; score coherence followed by matchup-recency guardrail
• NYY@LAA: raw 0.534 → calibrated 0.5332 → published 0.54; score coherence followed by matchup-recency guardrail
• OAK@TEX: raw 0.551 → calibrated 0.5499 → published 0.5487; published probability matches the calibrated/governed value
• PHI@ARI: raw 0.585 → calibrated 0.5816 → published 0.5602; score coherence followed by matchup-recency guardrail
• SD@CIN: raw 0.689 → calibrated 0.6778 → published 0.6; score coherence followed by matchup-recency guardrail
• SEA@BOS: raw 0.6328 → calibrated 0.6273 → published 0.53; score coherence followed by matchup-recency guardrail
• SF@PIT: raw 0.5615 → calibrated 0.56 → published 0.5445; score coherence followed by matchup-recency guardrail
• STL@LAD: raw 0.6793 → calibrated 0.6696 → published 0.58; score coherence followed by matchup-recency guardrail
• TOR@CLE: raw 0.546 → calibrated 0.545 → published 0.5552; no recorded transformation explains the publication delta

Every game, in detail

ATL@WAS — ❌ miss

• Predicted ATL 7-5 (conf -0.06, lean) · actual WAS 5-9
• Root cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.282705 · team-run err 6.0 · total bias -2
• Contributing factors:
• WAS offense under-projected by 4

BAL@COL — ❌ miss

• Predicted BAL 7-4 (conf -0.49, strong) · actual COL 2-4
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Run environment far lower than projected (over-dispersed)
• Errors: winner WRONG · Brier 0.556717 · team-run err 5.0 · total bias +5
• Contributing factors:
• total runs over-projected by 5 (11 vs 6)
• BAL offense over-projected by 5
• Gabriel Hughes (COL) allowed 4 fewer ER than projected (1 vs 5)
• committed to BAL at 0.49 confidence and it lost

CHW@HOU — ✅ hit

• Predicted CHW 6-5 (conf -0.06, lean) · actual CHW 5-1
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.2209 · team-run err 5.0 · total bias +5
• Contributing factors:
• total runs over-projected by 5 (11 vs 6)
• HOU offense over-projected by 4
• Sean Burke (CHW) allowed 4 fewer ER than projected (0 vs 4)
• Ronel Blanco (HOU) allowed 4 fewer ER than projected (0 vs 4)

DET@MIN — ❌ miss

• Predicted DET 4-3 (conf -0.06, lean) · actual MIN 2-15
• Root cause: Run environment far higher than projected (under-dispersed)
• Errors: winner WRONG · Brier 0.2809 · team-run err 14.0 · total bias -10
• Contributing factors:
• total runs under-projected by 10 (7 vs 17)
• MIN offense under-projected by 12
• Troy Melton (DET) allowed 3 more ER than projected (5 vs 2)

MIA@KC — ❌ miss

• Predicted KC 3-5 (conf +0.29, strong) · actual MIA 6-3
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Backed the wrong side with real conviction
• Errors: winner WRONG · Brier 0.417801 · team-run err 5.0 · total bias -1
• Contributing factors:
• MIA offense under-projected by 3
• Tyler Phillips (MIA) allowed 4 fewer ER than projected (0 vs 4)
• Randy Dobnak (KC) allowed 5 more ER than projected (6 vs 1)
• committed to KC at 0.29 confidence and it lost

MIL@CHC — ❌ miss

• Predicted CHC 6-8 (conf +0.40, strong) · actual MIL 9-4
• Root cause: Backed the wrong side with real conviction
• Secondary cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.488059 · team-run err 7.0 · total bias +1
• Contributing factors:
• MIL offense under-projected by 3
• CHC offense over-projected by 4
• Robert Gasser (MIL) allowed 3 fewer ER than projected (2 vs 5)
• committed to CHC at 0.40 confidence and it lost

NYM@TB — ✅ hit

• Predicted TB 5-6 (conf +0.05, lean) · actual TB 2-6
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.223781 · team-run err 3.0 · total bias +3
• Contributing factors:
• total runs over-projected by 3 (11 vs 8)
• NYM offense over-projected by 3
• Sean Manaea (NYM) allowed 3 more ER than projected (5 vs 2)
• Freddy Peralta (TB) allowed 3 fewer ER than projected (1 vs 4)

NYY@LAA — ✅ hit

• Predicted NYY 6-4 (conf -0.08, lean) · actual NYY 7-3
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.2116 · team-run err 2.0 · total bias +0

OAK@TEX — ✅ hit

• Predicted TEX 4-5 (conf +0.10, lean) · actual TEX 5-8
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.203649 · team-run err 4.0 · total bias -4
• Contributing factors:
• total runs under-projected by 4 (9 vs 13)
• TEX offense under-projected by 3

PHI@ARI — ✅ hit

• Predicted PHI 4-3 (conf -0.12, moderate) · actual PHI 7-1
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.193429 · team-run err 5.0 · total bias -1
• Contributing factors:
• PHI offense under-projected by 3

SD@CIN — ❌ miss

• Predicted SD 7-4 (conf -0.20, strong) · actual CIN 3-4
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Backed the wrong side with real conviction
• Errors: winner WRONG · Brier 0.36 · team-run err 4.0 · total bias +4
• Contributing factors:
• total runs over-projected by 4 (11 vs 7)
• SD offense over-projected by 4
• Nick Lodolo (CIN) allowed 4 fewer ER than projected (2 vs 6)
• committed to SD at 0.20 confidence and it lost

SEA@BOS — ❌ miss

• Predicted BOS 3-5 (conf +0.06, lean) · actual SEA 9-6
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Run environment far higher than projected (under-dispersed)
• Errors: winner WRONG · Brier 0.2809 · team-run err 7.0 · total bias -7
• Contributing factors:
• total runs under-projected by 7 (8 vs 15)
• SEA offense under-projected by 6
• Jedixson Paez (BOS) allowed 6 more ER than projected (8 vs 2)

SF@PIT — ✅ hit

• Predicted PIT 5-6 (conf +0.09, lean) · actual PIT 12-13
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.20746 · team-run err 14.0 · total bias -14
• Contributing factors:
• total runs under-projected by 14 (11 vs 25)
• SF offense under-projected by 7
• PIT offense under-projected by 7
• Logan Webb (SF) allowed 7 more ER than projected (9 vs 2)
• Paul Skenes (PIT) allowed 4 more ER than projected (5 vs 1)

STL@LAD — ❌ miss

• Predicted LAD 5-7 (conf +0.16, moderate) · actual STL 13-8
• Root cause: Run environment far higher than projected (under-dispersed)
• Secondary cause: Backed the wrong side with real conviction
• Errors: winner WRONG · Brier 0.3364 · team-run err 9.0 · total bias -9
• Contributing factors:
• total runs under-projected by 9 (12 vs 21)
• STL offense under-projected by 8
• committed to LAD at 0.16 confidence and it lost

TOR@CLE — ✅ hit

• Predicted CLE 4-5 (conf +0.11, moderate) · actual CLE 1-6
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.197815 · team-run err 4.0 · total bias +2
• Contributing factors:
• TOR offense over-projected by 3
• Gavin Williams (CLE) allowed 3 fewer ER than projected (0 vs 3)

Cross-game lever tension

Each candidate fix, with the games it would help and the games it would hurt. A lever that repairs one game and breaks two is visible here before it is proposed.

score.variance_multiplier / score.variance_floor (widen run distributions)

• Targets: brier_score, calibration_error
• Net: 5 helped vs 0 hurt
• Would help:
• BAL@COL (miss): lost a 0.49-confidence pick; wider tails lower overconfidence
• MIA@KC (miss): lost a 0.29-confidence pick; wider tails lower overconfidence
• MIL@CHC (miss): lost a 0.40-confidence pick; wider tails lower overconfidence
• SD@CIN (miss): lost a 0.20-confidence pick; wider tails lower overconfidence
• STL@LAD (miss): lost a 0.16-confidence pick; wider tails lower overconfidence

player.pitcher.v4 form weights / player.pitcher_prior_games

• Targets: starter_er_mae, starter_outs_mae
• Net: 10 helped vs 0 hurt
• Would help:
• BAL@COL (miss): starter ER off by 4
• CHW@HOU (hit): starter ER off by 4
• DET@MIN (miss): starter ER off by 3
• MIA@KC (miss): starter ER off by 5
• MIL@CHC (miss): starter ER off by 3
• NYM@TB (hit): starter ER off by 3
• SD@CIN (miss): starter ER off by 4
• SEA@BOS (miss): starter ER off by 6
• SF@PIT (hit): starter ER off by 7
• TOR@CLE (hit): starter ER off by 3

recent. / season. run-mean weighting

• Targets: team_run_mae, total_run_mae, accuracy
• Net: 13 helped vs 1 hurt
• Would help:
• ATL@WAS (miss): team runs off by 6
• BAL@COL (miss): team runs off by 5
• CHW@HOU (hit): team runs off by 5
• DET@MIN (miss): team runs off by 14
• MIA@KC (miss): team runs off by 5
• MIL@CHC (miss): team runs off by 7
• OAK@TEX (hit): team runs off by 4
• PHI@ARI (hit): team runs off by 5
• SD@CIN (miss): team runs off by 4
• SEA@BOS (miss): team runs off by 7
• SF@PIT (hit): team runs off by 14
• STL@LAD (miss): team runs off by 9
• TOR@CLE (hit): team runs off by 4
• Would hurt / risk:
• NYY@LAA (hit): team runs already close (off 2); reweight risks disturbing it