Deep slate diagnosis — 20260908

• Games: 15 · hits 11 · misses 4 · winner accuracy 0.7333
• Mean total-run bias on losses: 2.25 (negative = under-projected)
• Mean team-run error on losses: 6.75 · mean Brier on losses: 0.279

Loss primary drivers

• Small edge on the losing side: 1
• Run environment far lower than projected (over-dispersed): 1
• Starting-pitcher projection was wrong: 2

High-level causal issue register

[CRITICAL] Published probability differs materially without a recorded transformation

• Domain: probability_lineage · affected: ARI@KC, HOU@PHI, TB@ATL
• Evidence: 3 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] Published confidence buckets were poorly aligned on this slate

• Domain: calibration · affected: ARI@KC, CHC@MIL, CIN@LAD, CLE@BAL, COL@NYY, HOU@PHI, LAA@BOS, MIN@DET, NYM@MIA, PIT@CHW, STL@SF, TB@ATL, TEX@SEA, TOR@OAK, WAS@SD
• Evidence: weighted absolute calibration error 0.1825 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: ARI@KC, CHC@MIL, CIN@LAD, CLE@BAL, COL@NYY, HOU@PHI, LAA@BOS, MIN@DET, NYM@MIA, PIT@CHW, STL@SF, TB@ATL, TEX@SEA, TOR@OAK, WAS@SD
• Evidence: home/away/total R² = 0.0118/0.0210/0.0207
• 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@KC, CHC@MIL, CIN@LAD, CLE@BAL, COL@NYY, HOU@PHI, LAA@BOS, MIN@DET, NYM@MIA, PIT@CHW, STL@SF, TB@ATL, TEX@SEA, TOR@OAK, WAS@SD
• 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] Simulation produced no winner-direction intervention on this slate

• Domain: simulation · affected: ARI@KC, CHC@MIL, CIN@LAD, CLE@BAL, COL@NYY, HOU@PHI, LAA@BOS, MIN@DET, NYM@MIA, PIT@CHW, STL@SF, TB@ATL, TEX@SEA, TOR@OAK, WAS@SD
• 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: 20260829, 20260830, 20260831, 20260901, 20260902, 20260903, 20260904, 20260905, 20260906, 20260907, 20260908 (10 prior slates loaded)
• Highest-confidence pick loss streak: 0 slate(s)
• Actionable pattern detected: True

Escalated persistent patterns

• simulation_inert_on_slate — 11 consecutive slates (11 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 11 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 — 11 consecutive slates (11 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 11 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 [0.9412, 0.5, 0.6667, 0.4667, 0.3333, 0.5556, 0.625, 0.5333, 0.6, 0.5455, 0.7333]
• weighted_calibration_error: flat · series [0.3961, 0.0693, 0.1661, 0.2476, 0.21, 0.1474, 0.1316, 0.1, 0.0937, 0.2611, 0.1825]
• margin_bias: improving · series [-1.882, -2.786, -3.5, -2.533, -3.067, -1.778, -0.688, -1.467, -1.467, -0.182, -0.867]
• mean_team_run_error_on_losses: regressing · series [3.0, 7.0, 8.25, 7.125, 6.4, 6.75, 6.0, 5.286, 5.333, 3.8, 6.75]

Pipeline stage quality

• historical_ml: winner 11/15 (0.7333), team-run MAE 1.9069, margin MAE 2.5475
• variance_applied: winner 9/15 (0.6), team-run MAE 1.8333, margin MAE 2.4667
• enhanced_prediction: winner 11/15 (0.7333), team-run MAE 2.1667, margin MAE 2.4667
• embedded_savant_run_environment: winner 11/15 (0.7333), team-run MAE 2.2333, margin MAE 2.7333
• fundamental_edge_target_review: winner 11/15 (0.7333), team-run MAE 2.2333, margin MAE 2.7333
• capped_target: winner 11/15 (0.7333), team-run MAE 2.2333, margin MAE 2.7333
• simulation_result: winner 11/15 (0.7333), team-run MAE 2.2333, margin MAE 2.7333
• published_final: winner 11/15 (0.7333), team-run MAE 2.2333, margin MAE 2.7333

Probability lineage

• ARI@KC: raw 0.656 → calibrated 0.6339 → published 0.5736; no recorded transformation explains the publication delta
• CHC@MIL: raw 0.6488 → calibrated 0.5819 → published 0.54; score coherence followed by matchup-recency guardrail
• CIN@LAD: raw 0.8483 → calibrated 0.6718 → published 0.6718; published probability matches the calibrated/governed value
• CLE@BAL: raw 0.5 → calibrated 0.5071 → published 0.5051; published probability matches the calibrated/governed value
• COL@NYY: raw 0.8126 → calibrated 0.6531 → published 0.6448; no recorded transformation explains the publication delta
• HOU@PHI: raw 0.5638 → calibrated 0.5911 → published 0.5536; no recorded transformation explains the publication delta
• LAA@BOS: raw 0.5134 → calibrated 0.5841 → published 0.5273; score coherence followed by matchup-recency guardrail
• MIN@DET: raw 0.6516 → calibrated 0.583 → published 0.54; score coherence followed by matchup-recency guardrail
• NYM@MIA: raw 0.5787 → calibrated 0.5466 → published 0.5337; score coherence followed by matchup-recency guardrail
• PIT@CHW: raw 0.5355 → calibrated 0.5227 → published 0.5193; published probability matches the calibrated/governed value
• STL@SF: raw 0.5083 → calibrated 0.5725 → published 0.526; score coherence followed by matchup-recency guardrail
• TB@ATL: raw 0.5521 → calibrated 0.5447 → published 0.5215; no recorded transformation explains the publication delta
• TEX@SEA: raw 0.5 → calibrated 0.5343 → published 0.5248; score coherence followed by matchup-recency guardrail
• TOR@OAK: raw 0.6234 → calibrated 0.6173 → published 0.56; score coherence followed by matchup-recency guardrail
• WAS@SD: raw 0.5473 → calibrated 0.5291 → published 0.5211; score coherence followed by matchup-recency guardrail

Every game, in detail

ARI@KC — ✅ hit

• Predicted ARI 5-3 (conf -0.15, moderate) · actual ARI 5-3
• Root cause: strong_correct_call
• Errors: winner correct · Brier 0.181775 · team-run err 0.0 · total bias +0
• Contributing factors:
• Michael Wacha (KC) allowed 3 fewer ER than projected (0 vs 3)

CHC@MIL — ✅ hit

• Predicted MIL 4-6 (conf +0.08, lean) · actual MIL 3-4
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.2116 · team-run err 3.0 · total bias +3
• Contributing factors:
• total runs over-projected by 3 (10 vs 7)

CIN@LAD — ✅ hit

• Predicted LAD 3-9 (conf +0.34, strong) · actual LAD 2-3
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.107692 · team-run err 7.0 · total bias +7
• Contributing factors:
• total runs over-projected by 7 (12 vs 5)
• LAD offense over-projected by 6
• Nick Lodolo (CIN) allowed 3 fewer ER than projected (3 vs 6)

CLE@BAL — ✅ hit

• Predicted CLE 5-4 (conf -0.01, toss_up) · actual CLE 9-5
• Root cause: toss_up_went_our_way
• Errors: winner correct · Brier 0.244885 · team-run err 5.0 · total bias -5
• Contributing factors:
• total runs under-projected by 5 (9 vs 14)
• CLE offense under-projected by 4

COL@NYY — ✅ hit

• Predicted NYY 2-8 (conf +0.29, strong) · actual NYY 3-5
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.126161 · team-run err 4.0 · total bias +2
• Contributing factors:
• NYY offense over-projected by 3
• Gabriel Hughes (COL) allowed 4 fewer ER than projected (1 vs 5)

HOU@PHI — ✅ hit

• Predicted HOU 3-2 (conf -0.11, moderate) · actual HOU 6-5
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.199255 · team-run err 6.0 · total bias -6
• Contributing factors:
• total runs under-projected by 6 (5 vs 11)
• HOU offense under-projected by 3
• PHI offense under-projected by 3
• Andrew Painter (PHI) allowed 3 more ER than projected (5 vs 2)

LAA@BOS — ❌ miss

• Predicted BOS 3-4 (conf +0.05, lean) · actual LAA 6-1
• Root cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.278017 · team-run err 6.0 · total bias +0
• Contributing factors:
• LAA offense under-projected by 3
• BOS offense over-projected by 3

MIN@DET — ❌ miss

• Predicted DET 5-6 (conf +0.08, lean) · actual MIN 3-2
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Run environment far lower than projected (over-dispersed)
• Errors: winner WRONG · Brier 0.2916 · team-run err 6.0 · total bias +6
• Contributing factors:
• total runs over-projected by 6 (11 vs 5)
• DET offense over-projected by 4
• Dean Kremer (MIN) allowed 4 fewer ER than projected (2 vs 6)

NYM@MIA — ✅ hit

• Predicted NYM 6-5 (conf -0.07, lean) · actual NYM 7-5
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.21746 · team-run err 1.0 · total bias -1
• Contributing factors:
• Sandy Alcantara (MIA) allowed 5 more ER than projected (7 vs 2)

PIT@CHW — ❌ miss

• Predicted CHW 3-4 (conf +0.04, toss_up) · actual PIT 9-3
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Run environment far higher than projected (under-dispersed)
• Errors: winner WRONG · Brier 0.269648 · team-run err 7.0 · total bias -5
• Contributing factors:
• total runs under-projected by 5 (7 vs 12)
• PIT offense under-projected by 6
• Sean Burke (CHW) allowed 4 more ER than projected (7 vs 3)

STL@SF — ❌ miss

• Predicted STL 6-5 (conf -0.05, lean) · actual SF 1-2
• Root cause: Run environment far lower than projected (over-dispersed)
• Errors: winner WRONG · Brier 0.276709 · team-run err 8.0 · total bias +8
• Contributing factors:
• total runs over-projected by 8 (11 vs 3)
• STL offense over-projected by 5
• SF offense over-projected by 3

TB@ATL — ✅ hit

• Predicted TB 6-5 (conf -0.04, lean) · actual TB 7-1
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.228955 · team-run err 5.0 · total bias +3
• Contributing factors:
• total runs over-projected by 3 (11 vs 8)
• ATL offense over-projected by 4

TEX@SEA — ✅ hit

• Predicted TEX 5-4 (conf -0.05, lean) · actual TEX 10-5
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.225807 · team-run err 6.0 · total bias -6
• Contributing factors:
• total runs under-projected by 6 (9 vs 15)
• TEX offense under-projected by 5
• Cal Quantrill (TEX) allowed 3 more ER than projected (3 vs 0)
• Bryce Miller (SEA) allowed 3 more ER than projected (4 vs 1)

TOR@OAK — ✅ hit

• Predicted TOR 5-4 (conf -0.12, moderate) · actual TOR 4-2
• Root cause: strong_correct_call
• Errors: winner correct · Brier 0.1936 · team-run err 3.0 · total bias +3
• Contributing factors:
• total runs over-projected by 3 (9 vs 6)

WAS@SD — ✅ hit

• Predicted SD 4-5 (conf +0.04, lean) · actual SD 4-5
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.229392 · team-run err 0.0 · total bias +0
• Contributing factors:
• Casey Mize (SD) allowed 3 more ER than projected (3 vs 0)

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: 0 helped vs 2 hurt
• Would hurt / risk:
• CIN@LAD (hit): strong correct call at 0.34; widening softens a good edge
• COL@NYY (hit): strong correct call at 0.29; widening softens a good edge

player.pitcher.v4 form weights / player.pitcher_prior_games

• Targets: starter_er_mae, starter_outs_mae
• Net: 9 helped vs 1 hurt
• Would help:
• ARI@KC (hit): starter ER off by 3
• CIN@LAD (hit): starter ER off by 3
• COL@NYY (hit): starter ER off by 4
• HOU@PHI (hit): starter ER off by 3
• MIN@DET (miss): starter ER off by 4
• NYM@MIA (hit): starter ER off by 5
• PIT@CHW (miss): starter ER off by 4
• TEX@SEA (hit): starter ER off by 3
• WAS@SD (hit): starter ER off by 3
• Would hurt / risk:
• TOR@OAK (hit): 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 3 hurt
• Would help:
• CIN@LAD (hit): team runs off by 7
• CLE@BAL (hit): team runs off by 5
• COL@NYY (hit): team runs off by 4
• HOU@PHI (hit): team runs off by 6
• LAA@BOS (miss): team runs off by 6
• MIN@DET (miss): team runs off by 6
• PIT@CHW (miss): team runs off by 7
• STL@SF (miss): team runs off by 8
• TB@ATL (hit): team runs off by 5
• TEX@SEA (hit): team runs off by 6
• Would hurt / risk:
• ARI@KC (hit): team runs already close (off 0); reweight risks disturbing it
• NYM@MIA (hit): team runs already close (off 1); reweight risks disturbing it
• WAS@SD (hit): team runs already close (off 0); reweight risks disturbing it