Deep slate diagnosis — 20260922

• Games: 15 · hits 6 · misses 9 · winner accuracy 0.4
• Mean total-run bias on losses: 1.556 (negative = under-projected)
• Mean team-run error on losses: 6.222 · mean Brier on losses: 0.317

Loss primary drivers

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

High-level causal issue register

[CRITICAL] Published probability differs materially without a recorded transformation

• Domain: probability_lineage · affected: WAS@DET
• Evidence: 1 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@COL, CHW@KC, CIN@ATL, CLE@BOS, HOU@SEA, LAA@OAK, MIA@CHC, MIL@PHI, MIN@SF, NYM@TEX, SD@LAD, STL@PIT, TB@NYY, TB@NYY, WAS@DET
• Evidence: weighted absolute calibration error 0.1881 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@COL, CHW@KC, CIN@ATL, CLE@BOS, HOU@SEA, LAA@OAK, MIA@CHC, MIL@PHI, MIN@SF, NYM@TEX, SD@LAD, STL@PIT, TB@NYY, TB@NYY, WAS@DET
• Evidence: home/away/total R² = 0.0172/0.0294/0.0090
• 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: CIN@ATL, HOU@SEA, MIA@CHC
• Evidence: 0/3 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: ARI@COL, CHW@KC, CIN@ATL, CLE@BOS, HOU@SEA, LAA@OAK, MIA@CHC, MIL@PHI, MIN@SF, NYM@TEX, SD@LAD, STL@PIT, TB@NYY, TB@NYY, WAS@DET
• Evidence: 16/15 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@COL, CHW@KC, CIN@ATL, CLE@BOS, HOU@SEA, LAA@OAK, MIA@CHC, MIL@PHI, MIN@SF, NYM@TEX, SD@LAD, STL@PIT, TB@NYY, TB@NYY, WAS@DET
• 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: 20260912, 20260913, 20260914, 20260915, 20260916, 20260917, 20260918, 20260919, 20260920, 20260921, 20260922 (10 prior slates loaded)
• Highest-confidence pick loss streak: 2 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.
• single_slate_calibration_instability — 4 consecutive slates (8 in window)
• Target lever(s): confidence.empirical_max_weight, confidence.base_cap, confidence.moderate_cap
• Direction: increase empirical shrinkage
• Guardrail: single_slate_calibration_instability has persisted 4 consecutive slates; this is no longer single-slate noise. Treat it as an actionable pattern and propose a bounded shadow experiment against a target lever.
• 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: regressing · series [0.5333, 0.6, 0.7, 0.4, 0.6667, 0.4444, 0.6667, 0.7333, 0.6, 0.3333, 0.4]
• weighted_calibration_error: improving · series [0.1711, 0.0827, 0.2812, 0.2795, 0.1661, 0.1016, 0.1247, 0.3458, 0.3184, 0.2028, 0.1881]
• margin_bias: regressing · series [-2.667, -1.8, -0.4, -1.667, -1.133, -2.111, -2.0, -0.667, -1.533, -2.667, -1.867]
• mean_team_run_error_on_losses: regressing · series [6.143, 5.0, 6.333, 5.556, 6.4, 3.6, 5.2, 6.5, 6.5, 3.0, 6.222]

Pipeline stage quality

• historical_ml: winner 7/15 (0.4667), team-run MAE 2.9066, margin MAE 3.4277
• variance_applied: winner 6/15 (0.4), team-run MAE 2.9667, margin MAE 3.5333
• enhanced_prediction: winner 7/15 (0.4667), team-run MAE 3.2, margin MAE 4.0
• embedded_savant_run_environment: winner 6/15 (0.4), team-run MAE 3.1667, margin MAE 4.0667
• fundamental_edge_target_review: winner 6/15 (0.4), team-run MAE 3.2, margin MAE 4.1333
• capped_target: winner 6/15 (0.4), team-run MAE 3.2, margin MAE 4.1333
• simulation_result: winner 6/15 (0.4), team-run MAE 3.2, margin MAE 4.1333
• published_final: winner 6/15 (0.4), team-run MAE 3.2, margin MAE 4.1333

Probability lineage

• ARI@COL: raw 0.6408 → calibrated 0.6224 → published 0.56; score coherence followed by matchup-recency guardrail
• CHW@KC: raw 0.5 → calibrated 0.5073 → published 0.5053; published probability matches the calibrated/governed value
• CIN@ATL: raw 0.6937 → calibrated 0.699 → published 0.6; score coherence followed by matchup-recency guardrail
• CLE@BOS: raw 0.5 → calibrated 0.5073 → published 0.5053; published probability matches the calibrated/governed value
• HOU@SEA: raw 0.6546 → calibrated 0.6409 → published 0.6; score coherence followed by matchup-recency guardrail
• LAA@OAK: raw 0.5132 → calibrated 0.5602 → published 0.53; score coherence followed by matchup-recency guardrail
• MIA@CHC: raw 0.6927 → calibrated 0.6546 → published 0.53; score coherence followed by matchup-recency guardrail
• MIL@PHI: raw 0.537 → calibrated 0.6768 → published 0.6; score coherence followed by matchup-recency guardrail
• MIN@SF: raw 0.5367 → calibrated 0.5232 → published 0.5154; score coherence followed by matchup-recency guardrail
• NYM@TEX: raw 0.5722 → calibrated 0.6009 → published 0.6009; published probability matches the calibrated/governed value
• SD@LAD: raw 0.6139 → calibrated 0.5641 → published 0.53; score coherence followed by matchup-recency guardrail
• STL@PIT: raw 0.5625 → calibrated 0.5535 → published 0.5387; score coherence followed by matchup-recency guardrail
• TB@NYY: raw 0.5842 → calibrated 0.5483 → published 0.5349; score coherence followed by matchup-recency guardrail
• TB@NYY: raw 0.5501 → calibrated 0.5286 → published 0.5206; score coherence followed by matchup-recency guardrail
• WAS@DET: raw 0.6101 → calibrated 0.5929 → published 0.5709; no recorded transformation explains the publication delta

Every game, in detail

ARI@COL — ✅ hit

• Predicted ARI 6-5 (conf -0.12, moderate) · actual ARI 7-2
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.1936 · team-run err 4.0 · total bias +2
• Contributing factors:
• COL offense over-projected by 3

CHW@KC — ✅ hit

• Predicted CHW 5-4 (conf -0.01, toss_up) · actual CHW 14-11
• Root cause: toss_up_went_our_way
• Errors: winner correct · Brier 0.244771 · team-run err 16.0 · total bias -16
• Contributing factors:
• total runs under-projected by 16 (9 vs 25)
• CHW offense under-projected by 9
• KC offense under-projected by 7
• Daniel Lynch IV (KC) allowed 8 more ER than projected (9 vs 1)

CIN@ATL — ❌ miss

• Predicted ATL 6-9 (conf +0.20, strong) · actual CIN 4-0
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Run environment far lower than projected (over-dispersed)
• Errors: winner WRONG · Brier 0.36 · team-run err 11.0 · total bias +11
• Contributing factors:
• total runs over-projected by 11 (15 vs 4)
• ATL offense over-projected by 9
• Brandon Williamson (CIN) allowed 5 fewer ER than projected (0 vs 5)
• committed to ATL at 0.20 confidence and it lost

CLE@BOS — ❌ miss

• Predicted BOS 3-4 (conf +0.01, toss_up) · actual CLE 3-2
• Root cause: Genuine toss-up that flipped
• Errors: winner WRONG · Brier 0.255285 · team-run err 2.0 · total bias +2

HOU@SEA — ❌ miss

• Predicted SEA 3-5 (conf +0.20, strong) · actual HOU 7-0
• Root cause: Backed the wrong side with real conviction
• Secondary cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.36 · team-run err 9.0 · total bias +1
• Contributing factors:
• HOU offense under-projected by 4
• SEA offense over-projected by 5
• Cristian Javier (HOU) allowed 3 fewer ER than projected (0 vs 3)
• committed to SEA at 0.20 confidence and it lost

LAA@OAK — ❌ miss

• Predicted LAA 5-4 (conf -0.06, lean) · actual OAK 7-9
• Root 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 (9 vs 16)
• OAK offense under-projected by 5

MIA@CHC — ❌ miss

• Predicted CHC 4-6 (conf +0.06, lean) · actual MIA 8-2
• Root cause: Starting-pitcher projection was wrong
• Secondary cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.2809 · team-run err 8.0 · total bias +0
• Contributing factors:
• MIA offense under-projected by 4
• CHC offense over-projected by 4
• Janson Junk (MIA) allowed 5 fewer ER than projected (0 vs 5)

MIL@PHI — ❌ miss

• Predicted MIL 7-4 (conf -0.20, strong) · actual PHI 4-6
• Root cause: Backed the wrong side with real conviction
• Secondary cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.36 · team-run err 5.0 · total bias +1
• Contributing factors:
• MIL offense over-projected by 3
• committed to MIL at 0.20 confidence and it lost

MIN@SF — ✅ hit

• Predicted MIN 5-3 (conf -0.03, toss_up) · actual MIN 3-2
• Root cause: toss_up_went_our_way
• Errors: winner correct · Brier 0.234816 · team-run err 3.0 · total bias +3
• Contributing factors:
• total runs over-projected by 3 (8 vs 5)

NYM@TEX — ❌ miss

• Predicted TEX 4-6 (conf +0.20, strong) · actual NYM 6-3
• Root cause: Backed the wrong side with real conviction
• Secondary cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.361104 · team-run err 5.0 · total bias +1
• Contributing factors:
• TEX offense over-projected by 3
• committed to TEX at 0.20 confidence and it lost

SD@LAD — ✅ hit

• Predicted LAD 4-6 (conf +0.06, lean) · actual LAD 0-7
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.2209 · team-run err 5.0 · total bias +3
• Contributing factors:
• total runs over-projected by 3 (10 vs 7)
• SD offense over-projected by 4

STL@PIT — ✅ hit

• Predicted PIT 4-5 (conf +0.08, lean) · actual PIT 0-2
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.212833 · team-run err 7.0 · total bias +7
• Contributing factors:
• total runs over-projected by 7 (9 vs 2)
• STL offense over-projected by 4
• PIT offense over-projected by 3

TB@NYY — ✅ hit

• Predicted NYY 3-4 (conf +0.07, lean) · actual NYY 0-2
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.216301 · team-run err 5.0 · total bias +5
• Contributing factors:
• total runs over-projected by 5 (7 vs 2)
• TB offense over-projected by 3
• Carlos Rodón (NYY) allowed 3 more ER than projected (4 vs 1)

TB@NYY — ❌ miss

• Predicted NYY 4-5 (conf +0.04, lean) · actual TB 6-1
• Root cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.271061 · team-run err 6.0 · total bias +2
• Contributing factors:
• NYY offense over-projected by 4

WAS@DET — ❌ miss

• Predicted DET 3-4 (conf +0.14, moderate) · actual WAS 3-1
• Root cause: Backed the wrong side with real conviction
• Errors: winner WRONG · Brier 0.325966 · team-run err 3.0 · total bias +3
• Contributing factors:
• total runs over-projected by 3 (7 vs 4)
• DET offense over-projected by 3
• committed to DET at 0.14 confidence and it lost

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:
• CIN@ATL (miss): lost a 0.20-confidence pick; wider tails lower overconfidence
• HOU@SEA (miss): lost a 0.20-confidence pick; wider tails lower overconfidence
• MIL@PHI (miss): lost a 0.20-confidence pick; wider tails lower overconfidence
• NYM@TEX (miss): lost a 0.20-confidence pick; wider tails lower overconfidence
• WAS@DET (miss): lost a 0.14-confidence pick; wider tails lower overconfidence

player.pitcher.v4 form weights / player.pitcher_prior_games

• Targets: starter_er_mae, starter_outs_mae
• Net: 5 helped vs 4 hurt
• Would help:
• CHW@KC (hit): starter ER off by 8
• CIN@ATL (miss): starter ER off by 5
• HOU@SEA (miss): starter ER off by 3
• MIA@CHC (miss): starter ER off by 5
• TB@NYY (hit): starter ER off by 3
• Would hurt / risk:
• ARI@COL (hit): starter projection was accurate (ER off 1); reweight risks disturbing it
• MIL@PHI (miss): starter projection was accurate (ER off 1); reweight risks disturbing it
• NYM@TEX (miss): starter projection was accurate (ER off 1); reweight risks disturbing it
• STL@PIT (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: 12 helped vs 1 hurt
• Would help:
• ARI@COL (hit): team runs off by 4
• CHW@KC (hit): team runs off by 16
• CIN@ATL (miss): team runs off by 11
• HOU@SEA (miss): team runs off by 9
• LAA@OAK (miss): team runs off by 7
• MIA@CHC (miss): team runs off by 8
• MIL@PHI (miss): team runs off by 5
• NYM@TEX (miss): team runs off by 5
• SD@LAD (hit): team runs off by 5
• STL@PIT (hit): team runs off by 7
• TB@NYY (hit): team runs off by 5
• TB@NYY (miss): team runs off by 6
• Would hurt / risk:
• CLE@BOS (miss): team runs already close (off 2); reweight risks disturbing it