Deep slate diagnosis — 20260904

• Games: 16 · hits 10 · misses 6 · winner accuracy 0.625
• Mean total-run bias on losses: 2.0 (negative = under-projected)
• Mean team-run error on losses: 6.0 · mean Brier on losses: 0.318

Loss primary drivers

• Run environment far lower than projected (over-dispersed): 1
• Run environment far higher than projected (under-dispersed): 1
• Starting-pitcher projection was wrong: 3
• Backed the wrong side with real conviction: 1

High-level causal issue register

[CRITICAL] Published probability differs materially without a recorded transformation

• Domain: probability_lineage · affected: ARI@HOU, SF@NYM
• Evidence: 2 game(s) have an unexplained ≥2pp calibrated-to-published delta
• Action: Block any promotion until the missing probability transformation is recorded and reconciled.
• Do not infer: Do not assume the probability governor was bypassed until you have traced the pick-side math.

[HIGH] Validated run models explain very little holdout variance

• Domain: model_validation · affected: ARI@HOU, ATL@PHI, BOS@BAL, CHC@MIA, DET@CLE, DET@CLE, LAA@PIT, MIL@CIN, MIN@CHW, NYY@SD, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC, WAS@LAD
• Evidence: home/away/total R² = 0.0187/0.0264/0.0260
• 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.

[INFO] Every settled game received structural QA

• Domain: coverage · affected: ARI@HOU, ATL@PHI, BOS@BAL, CHC@MIA, DET@CLE, DET@CLE, LAA@PIT, MIL@CIN, MIN@CHW, NYY@SD, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC, WAS@LAD
• Evidence: 16/16 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] Simulation produced no winner-direction intervention on this slate

• Domain: simulation · affected: ARI@HOU, ATL@PHI, BOS@BAL, CHC@MIA, DET@CLE, DET@CLE, LAA@PIT, MIL@CIN, MIN@CHW, NYY@SD, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC, WAS@LAD
• 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, 20260902, 20260903, 20260904 (7 prior slates loaded)
• Highest-confidence pick loss streak: 0 slate(s)
• Actionable pattern detected: True

Escalated persistent patterns

• simulation_inert_on_slate — 8 consecutive slates (8 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 8 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 — 8 consecutive slates (8 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 8 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: improving · series [0.4667, 0.9412, 0.5, 0.6667, 0.4667, 0.3333, 0.5556, 0.625]
• weighted_calibration_error: improving · series [0.2179, 0.3961, 0.0693, 0.1661, 0.2476, 0.21, 0.1474, 0.1316]
• margin_bias: improving · series [-1.333, -1.882, -2.786, -3.5, -2.533, -3.067, -1.778, -0.688]
• mean_team_run_error_on_losses: flat · series [5.0, 3.0, 7.0, 8.25, 7.125, 6.4, 6.75, 6.0]

Pipeline stage quality

• historical_ml: winner 10/16 (0.625), team-run MAE 2.4463, margin MAE 2.2201
• variance_applied: winner 8/16 (0.5), team-run MAE 2.4688, margin MAE 2.1875
• enhanced_prediction: winner 10/16 (0.625), team-run MAE 2.3438, margin MAE 2.0625
• embedded_savant_run_environment: winner 10/16 (0.625), team-run MAE 2.4062, margin MAE 2.0625
• fundamental_edge_target_review: winner 10/16 (0.625), team-run MAE 2.4062, margin MAE 2.0625
• capped_target: winner 10/16 (0.625), team-run MAE 2.4062, margin MAE 2.0625
• simulation_result: winner 10/16 (0.625), team-run MAE 2.4062, margin MAE 2.0625
• published_final: winner 10/16 (0.625), team-run MAE 2.4062, margin MAE 2.0625

Probability lineage

• ARI@HOU: raw 0.6275 → calibrated 0.6219 → published 0.5825; no recorded transformation explains the publication delta
• ATL@PHI: raw 0.585 → calibrated 0.5815 → published 0.5599; score coherence followed by matchup-recency guardrail
• BOS@BAL: raw 0.6303 → calibrated 0.6241 → published 0.53; score coherence followed by matchup-recency guardrail
• CHC@MIA: raw 0.7374 → calibrated 0.7104 → published 0.7245; score_coherence aligned probability to the published score margin
• DET@CLE: raw 0.5085 → calibrated 0.5083 → published 0.5251; score coherence followed by matchup-recency guardrail
• DET@CLE: raw 0.568 → calibrated 0.5662 → published 0.5544; score coherence followed by matchup-recency guardrail
• LAA@PIT: raw 0.67 → calibrated 0.6591 → published 0.662; score_coherence aligned probability to the published score margin
• MIL@CIN: raw 0.6628 → calibrated 0.6526 → published 0.54; score coherence followed by matchup-recency guardrail
• MIN@CHW: raw 0.5216 → calibrated 0.5211 → published 0.5381; score coherence followed by matchup-recency guardrail
• NYY@SD: raw 0.6275 → calibrated 0.6219 → published 0.6434; score_coherence aligned probability to the published score margin
• OAK@SEA: raw 0.551 → calibrated 0.5499 → published 0.5485; published probability matches the calibrated/governed value
• SF@NYM: raw 0.5265 → calibrated 0.5258 → published 0.5468; no recorded transformation explains the publication delta
• STL@COL: raw 0.551 → calibrated 0.5499 → published 0.5485; published probability matches the calibrated/governed value
• TB@TEX: raw 0.585 → calibrated 0.5815 → published 0.5599; score coherence followed by matchup-recency guardrail
• TOR@KC: raw 0.551 → calibrated 0.5499 → published 0.5571; score coherence followed by matchup-recency guardrail
• WAS@LAD: raw 0.5619 → calibrated 0.5603 → published 0.5615; published probability matches the calibrated/governed value

Every game, in detail

ARI@HOU — ✅ hit

• Predicted HOU 4-6 (conf +0.17, moderate) · actual HOU 1-3
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.174304 · team-run err 6.0 · total bias +6
• Contributing factors:
• total runs over-projected by 6 (10 vs 4)
• ARI offense over-projected by 3
• HOU offense over-projected by 3
• Cristian Javier (HOU) allowed 3 fewer ER than projected (1 vs 4)

ATL@PHI — ✅ hit

• Predicted ATL 4-3 (conf -0.12, moderate) · actual ATL 5-2
• Root cause: strong_correct_call
• Errors: winner correct · Brier 0.193658 · team-run err 2.0 · total bias +0

BOS@BAL — ❌ miss

• Predicted BAL 4-6 (conf +0.06, lean) · actual BOS 1-0
• Root cause: Run environment far lower than projected (over-dispersed)
• Errors: winner WRONG · Brier 0.2809 · team-run err 9.0 · total bias +9
• Contributing factors:
• total runs over-projected by 9 (10 vs 1)
• BOS offense over-projected by 3
• BAL offense over-projected by 6
• Ranger Suarez (BOS) allowed 3 fewer ER than projected (0 vs 3)
• Shane Baz (BAL) allowed 3 fewer ER than projected (1 vs 4)

CHC@MIA — ✅ hit

• Predicted CHC 6-3 (conf -0.45, strong) · actual CHC 6-1
• Root cause: strong_correct_call
• Errors: winner correct · Brier 0.07589 · team-run err 2.0 · total bias +2
• Contributing factors:
• Janson Junk (MIA) allowed 3 fewer ER than projected (1 vs 4)

DET@CLE — ✅ hit

• Predicted CLE 3-4 (conf +0.05, lean) · actual CLE 6-7
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.225537 · team-run err 6.0 · total bias -6
• Contributing factors:
• total runs under-projected by 6 (7 vs 13)
• DET offense under-projected by 3
• CLE offense under-projected by 3
• Logan Allen (CLE) allowed 3 more ER than projected (3 vs 0)

DET@CLE — ✅ hit

• Predicted CLE 3-4 (conf +0.11, moderate) · actual CLE 3-4
• Root cause: strong_correct_call
• Errors: winner correct · Brier 0.198558 · team-run err 0.0 · total bias +0

LAA@PIT — ✅ hit

• Predicted PIT 5-7 (conf +0.32, strong) · actual PIT 0-1
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.114226 · team-run err 11.0 · total bias +11
• Contributing factors:
• total runs over-projected by 11 (12 vs 1)
• LAA offense over-projected by 5
• PIT offense over-projected by 6
• Ryan Johnson (LAA) allowed 6 fewer ER than projected (1 vs 7)
• Jared Jones (PIT) allowed 5 fewer ER than projected (0 vs 5)

MIL@CIN — ✅ hit

• Predicted MIL 8-5 (conf -0.08, lean) · actual MIL 10-7
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.2116 · team-run err 4.0 · total bias -4
• Contributing factors:
• total runs under-projected by 4 (13 vs 17)
• Shane Drohan (MIL) allowed 5 more ER than projected (6 vs 1)

MIN@CHW — ❌ miss

• Predicted MIN 6-5 (conf -0.08, lean) · actual CHW 1-4
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Run environment far lower than projected (over-dispersed)
• Errors: winner WRONG · Brier 0.289555 · team-run err 6.0 · total bias +6
• Contributing factors:
• total runs over-projected by 6 (11 vs 5)
• MIN offense over-projected by 5
• Erick Fedde (CHW) allowed 5 fewer ER than projected (0 vs 5)

NYY@SD — ❌ miss

• Predicted NYY 6-3 (conf -0.29, strong) · actual SD 2-3
• Root cause: Backed the wrong side with real conviction
• Secondary cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.414021 · team-run err 4.0 · total bias +4
• Contributing factors:
• total runs over-projected by 4 (9 vs 5)
• NYY offense over-projected by 4
• Walker Buehler (SD) allowed 3 fewer ER than projected (1 vs 4)
• committed to NYY at 0.29 confidence and it lost

OAK@SEA — ❌ miss

• Predicted SEA 5-6 (conf +0.10, lean) · actual OAK 7-6
• Root cause: Starting-pitcher projection was wrong
• Errors: winner WRONG · Brier 0.30085 · team-run err 2.0 · total bias -2
• Contributing factors:
• Logan Gilbert (SEA) allowed 6 more ER than projected (7 vs 1)

SF@NYM — ✅ hit

• Predicted NYM 4-5 (conf +0.09, lean) · actual NYM 6-10
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.205362 · team-run err 7.0 · total bias -7
• Contributing factors:
• total runs under-projected by 7 (9 vs 16)
• NYM offense under-projected by 5

STL@COL — ✅ hit

• Predicted STL 6-5 (conf -0.10, lean) · actual STL 7-6
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.203854 · team-run err 2.0 · total bias -2

TB@TEX — ❌ miss

• Predicted TEX 3-4 (conf +0.12, moderate) · actual TB 7-6
• Root cause: Run environment far higher than projected (under-dispersed)
• Secondary cause: Backed the wrong side with real conviction
• Errors: winner WRONG · Brier 0.313526 · team-run err 6.0 · total bias -6
• Contributing factors:
• total runs under-projected by 6 (7 vs 13)
• TB offense under-projected by 4
• committed to TEX at 0.12 confidence and it lost

TOR@KC — ❌ miss

• Predicted KC 5-7 (conf +0.11, moderate) · actual TOR 9-2
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Backed the wrong side with real conviction
• Errors: winner WRONG · Brier 0.310311 · team-run err 9.0 · total bias +1
• Contributing factors:
• TOR offense under-projected by 4
• KC offense over-projected by 5
• Jameson Taillon (TOR) allowed 3 fewer ER than projected (1 vs 4)
• Daniel Lynch IV (KC) allowed 4 more ER than projected (5 vs 1)
• committed to KC at 0.11 confidence and it lost

WAS@LAD — ✅ hit

• Predicted LAD 4-5 (conf +0.12, moderate) · actual LAD 3-5
• Root cause: strong_correct_call
• Errors: winner correct · Brier 0.192284 · team-run err 1.0 · total bias +1

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: 3 helped vs 2 hurt
• Would help:
• NYY@SD (miss): lost a 0.29-confidence pick; wider tails lower overconfidence
• TB@TEX (miss): lost a 0.12-confidence pick; wider tails lower overconfidence
• TOR@KC (miss): lost a 0.11-confidence pick; wider tails lower overconfidence
• Would hurt / risk:
• CHC@MIA (hit): strong correct call at 0.45; widening softens a good edge
• LAA@PIT (hit): strong correct call at 0.32; widening softens a good edge

player.pitcher.v4 form weights / player.pitcher_prior_games

• Targets: starter_er_mae, starter_outs_mae
• Net: 10 helped vs 3 hurt
• Would help:
• ARI@HOU (hit): starter ER off by 3
• BOS@BAL (miss): starter ER off by 3
• CHC@MIA (hit): starter ER off by 3
• DET@CLE (hit): starter ER off by 3
• LAA@PIT (hit): starter ER off by 6
• MIL@CIN (hit): starter ER off by 5
• MIN@CHW (miss): starter ER off by 5
• NYY@SD (miss): starter ER off by 3
• OAK@SEA (miss): starter ER off by 6
• TOR@KC (miss): starter ER off by 4
• Would hurt / risk:
• ATL@PHI (hit): starter projection was accurate (ER off 0); reweight risks disturbing it
• SF@NYM (hit): starter projection was accurate (ER off 1); reweight risks disturbing it
• TB@TEX (miss): starter projection was accurate (ER off 1); reweight risks disturbing it

recent. / season. run-mean weighting

• Targets: team_run_mae, total_run_mae, accuracy
• Net: 10 helped vs 6 hurt
• Would help:
• ARI@HOU (hit): team runs off by 6
• BOS@BAL (miss): team runs off by 9
• DET@CLE (hit): team runs off by 6
• LAA@PIT (hit): team runs off by 11
• MIL@CIN (hit): team runs off by 4
• MIN@CHW (miss): team runs off by 6
• NYY@SD (miss): team runs off by 4
• SF@NYM (hit): team runs off by 7
• TB@TEX (miss): team runs off by 6
• TOR@KC (miss): team runs off by 9
• Would hurt / risk:
• ATL@PHI (hit): team runs already close (off 2); reweight risks disturbing it
• CHC@MIA (hit): team runs already close (off 2); reweight risks disturbing it
• DET@CLE (hit): team runs already close (off 0); reweight risks disturbing it
• OAK@SEA (miss): team runs already close (off 2); reweight risks disturbing it
• STL@COL (hit): team runs already close (off 2); reweight risks disturbing it
• WAS@LAD (hit): team runs already close (off 1); reweight risks disturbing it