Deep slate diagnosis — 20260830

• Games: 14 · hits 7 · misses 7 · winner accuracy 0.5
• Mean total-run bias on losses: -0.714 (negative = under-projected)
• Mean team-run error on losses: 7.0 · mean Brier on losses: 0.329

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
• Backed the wrong side with real conviction: 3

High-level causal issue register

[CRITICAL] Published probability differs materially without a recorded transformation

• Domain: probability_lineage · affected: CHW@MIN
• 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] Canonical artifact warning requires reconciliation

• Domain: artifact_integrity · affected: slate-wide
• Evidence: structural audit totals: {'critical': 0, 'error': 0, 'warning': 1, 'info': 100}
• 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] An intermediate stage materially degraded winner direction

• Domain: pipeline_stages · affected: BAL@OAK, PIT@STL
• Evidence: historical_ml → variance_applied changed winner accuracy by -0.1429 (-2 hits)
• Action: List the exact games whose winner flipped at this stage, and keep both the before-stage and after-stage picks in the trace.
• Do not infer: Even if a later stage recovers the pick, this unstable step is still a risk; do not treat it as harmless.

[HIGH] Validated run models explain very little holdout variance

• Domain: model_validation · affected: BAL@OAK, BOS@NYY, CHW@MIN, CIN@CHC, COL@ATL, HOU@NYM, KC@CLE, LAD@DET, MIA@WAS, PHI@LAA, PIT@STL, SD@TB, SEA@TOR, TEX@MIL
• Evidence: home/away/total R² = 0.0429/0.0253/0.0224
• 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: CHW@MIN, CIN@CHC, PHI@LAA
• Evidence: 1/3 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: BAL@OAK, BOS@NYY, CHW@MIN, CIN@CHC, COL@ATL, HOU@NYM, KC@CLE, LAD@DET, MIA@WAS, PHI@LAA, PIT@STL, SD@TB, SEA@TOR, TEX@MIL
• Evidence: 14/14 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: BAL@OAK, BOS@NYY, CHW@MIN, CIN@CHC, COL@ATL, HOU@NYM, KC@CLE, LAD@DET, MIA@WAS, PHI@LAA, PIT@STL, SD@TB, SEA@TOR, TEX@MIL
• 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 (2 prior slates loaded)
• Highest-confidence pick loss streak: 1 slate(s)
• Actionable pattern detected: True

Escalated persistent patterns

• simulation_inert_on_slate — 3 consecutive slates (3 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 3 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 — 3 consecutive slates (3 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 3 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]
• weighted_calibration_error: improving · series [0.2179, 0.3961, 0.0693]
• margin_bias: regressing · series [-1.333, -1.882, -2.786]
• mean_team_run_error_on_losses: regressing · series [5.0, 3.0, 7.0]

Pipeline stage quality

• historical_ml: winner 7/14 (0.5), team-run MAE 2.4184, margin MAE 4.4881
• variance_applied: winner 5/14 (0.3571), team-run MAE 2.3929, margin MAE 4.5
• enhanced_prediction: winner 7/14 (0.5), team-run MAE 2.5, margin MAE 4.7143
• embedded_savant_run_environment: winner 7/14 (0.5), team-run MAE 2.4643, margin MAE 4.6429
• fundamental_edge_target_review: winner 7/14 (0.5), team-run MAE 2.4643, margin MAE 4.6429
• capped_target: winner 7/14 (0.5), team-run MAE 2.4643, margin MAE 4.6429
• simulation_result: winner 7/14 (0.5), team-run MAE 2.4643, margin MAE 4.6429
• published_final: winner 7/14 (0.5), team-run MAE 2.4643, margin MAE 4.6429

Probability lineage

• BAL@OAK: raw 0.568 → calibrated 0.5662 → published 0.5514; score coherence followed by matchup-recency guardrail
• BOS@NYY: raw 0.5085 → calibrated 0.5082 → published 0.53; score coherence followed by matchup-recency guardrail
• CHW@MIN: raw 0.7341 → calibrated 0.7117 → published 0.6867; no recorded transformation explains the publication delta
• CIN@CHC: raw 0.7597 → calibrated 0.7444 → published 0.56; score coherence followed by matchup-recency guardrail
• COL@ATL: raw 0.6332 → calibrated 0.6277 → published 0.58; score coherence followed by matchup-recency guardrail
• HOU@NYM: raw 0.5311 → calibrated 0.5303 → published 0.5416; score coherence followed by matchup-recency guardrail
• KC@CLE: raw 0.5085 → calibrated 0.5082 → published 0.53; score coherence followed by matchup-recency guardrail
• LAD@DET: raw 0.551 → calibrated 0.5498 → published 0.5487; published probability matches the calibrated/governed value
• MIA@WAS: raw 0.5947 → calibrated 0.5913 → published 0.5749; no recorded transformation explains the publication delta
• PHI@LAA: raw 0.67 → calibrated 0.6609 → published 0.6655; score_coherence aligned probability to the published score margin
• PIT@STL: raw 0.5838 → calibrated 0.5804 → published 0.5597; score coherence followed by matchup-recency guardrail
• SD@TB: raw 0.5085 → calibrated 0.5082 → published 0.5337; score coherence followed by matchup-recency guardrail
• SEA@TOR: raw 0.5085 → calibrated 0.5082 → published 0.53; score coherence followed by matchup-recency guardrail
• TEX@MIL: raw 0.6332 → calibrated 0.6277 → published 0.5768; score coherence followed by matchup-recency guardrail

Every game, in detail

BAL@OAK — ✅ hit

• Predicted BAL 8-6 (conf -0.10, moderate) · actual BAL 8-5
• Root cause: strong_correct_call
• Errors: winner correct · Brier 0.201245 · team-run err 1.0 · total bias +1

BOS@NYY — ❌ miss

• Predicted BOS 5-4 (conf -0.06, lean) · actual NYY 1-16
• 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 16.0 · total bias -8
• Contributing factors:
• total runs under-projected by 8 (9 vs 17)
• BOS offense over-projected by 4
• NYY offense under-projected by 12
• Ranger Suarez (BOS) allowed 3 more ER than projected (6 vs 3)
• Will Warren (NYY) allowed 4 fewer ER than projected (1 vs 5)

CHW@MIN — ❌ miss

• Predicted CHW 7-4 (conf -0.37, strong) · actual MIN 1-5
• Root cause: Run environment far lower than projected (over-dispersed)
• Secondary cause: Backed the wrong side with real conviction
• Errors: winner WRONG · Brier 0.471521 · team-run err 7.0 · total bias +5
• Contributing factors:
• total runs over-projected by 5 (11 vs 6)
• CHW offense over-projected by 6
• Zebby Matthews (MIN) allowed 3 fewer ER than projected (1 vs 4)
• committed to CHW at 0.37 confidence and it lost

CIN@CHC — ❌ miss

• Predicted CHC 4-7 (conf +0.12, moderate) · actual CIN 7-5
• Root cause: Backed the wrong side with real conviction
• Secondary cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.3136 · team-run err 5.0 · total bias -1
• Contributing factors:
• CIN offense under-projected by 3
• Shota Imanaga (CHC) allowed 3 more ER than projected (6 vs 3)
• committed to CHC at 0.12 confidence and it lost

COL@ATL — ✅ hit

• Predicted ATL 3-6 (conf +0.16, moderate) · actual ATL 2-3
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.1764 · team-run err 4.0 · total bias +4
• Contributing factors:
• total runs over-projected by 4 (9 vs 5)
• ATL offense over-projected by 3

HOU@NYM — ❌ miss

• Predicted NYM 4-5 (conf +0.08, lean) · actual HOU 6-3
• Root cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.293383 · team-run err 4.0 · total bias +0

KC@CLE — ✅ hit

• Predicted CLE 5-6 (conf +0.06, lean) · actual CLE 1-11
• Root cause: right_winner_wrong_score
• Errors: winner correct · Brier 0.2209 · team-run err 9.0 · total bias -1
• Contributing factors:
• KC offense over-projected by 4
• CLE offense under-projected by 5
• Seth Lugo (KC) allowed 7 more ER than projected (9 vs 2)

LAD@DET — ✅ hit

• Predicted LAD 4-3 (conf -0.10, lean) · actual LAD 6-1
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.203663 · team-run err 4.0 · total bias +0

MIA@WAS — ❌ miss

• Predicted WAS 3-4 (conf +0.15, moderate) · actual MIA 6-2
• Root cause: Backed the wrong side with real conviction
• Secondary cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.330554 · team-run err 5.0 · total bias -1
• Contributing factors:
• MIA offense under-projected by 3
• committed to WAS at 0.15 confidence and it lost

PHI@LAA — ✅ hit

• Predicted PHI 5-2 (conf -0.33, strong) · actual PHI 5-2
• Root cause: strong_correct_call
• Errors: winner correct · Brier 0.111918 · team-run err 0.0 · total bias +0
• Contributing factors:
• Yusei Kikuchi (LAA) allowed 3 fewer ER than projected (2 vs 5)

PIT@STL — ✅ hit

• Predicted PIT 5-4 (conf -0.12, moderate) · actual PIT 5-4
• Root cause: strong_correct_call
• Errors: winner correct · Brier 0.193821 · team-run err 0.0 · total bias +0

SD@TB — ✅ hit

• Predicted TB 3-4 (conf +0.07, lean) · actual TB 4-5
• Root cause: solid_correct_call
• Errors: winner correct · Brier 0.217457 · team-run err 2.0 · total bias -2

SEA@TOR — ❌ miss

• Predicted SEA 5-4 (conf -0.06, lean) · actual TOR 0-7
• 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 +2
• Contributing factors:
• SEA offense over-projected by 5
• TOR offense under-projected by 3
• Max Scherzer (TOR) allowed 4 fewer ER than projected (0 vs 4)

TEX@MIL — ❌ miss

• Predicted MIL 4-5 (conf +0.15, moderate) · actual TEX 7-4
• Root cause: Backed the wrong side with real conviction
• Secondary cause: Small edge on the losing side
• Errors: winner WRONG · Brier 0.332731 · team-run err 4.0 · total bias -2
• Contributing factors:
• TEX offense under-projected by 3
• committed to MIL at 0.15 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: 4 helped vs 1 hurt
• Would help:
• CHW@MIN (miss): lost a 0.37-confidence pick; wider tails lower overconfidence
• CIN@CHC (miss): lost a 0.12-confidence pick; wider tails lower overconfidence
• MIA@WAS (miss): lost a 0.15-confidence pick; wider tails lower overconfidence
• TEX@MIL (miss): lost a 0.15-confidence pick; wider tails lower overconfidence
• Would hurt / risk:
• PHI@LAA (hit): strong correct call at 0.33; widening softens a good edge

player.pitcher.v4 form weights / player.pitcher_prior_games

• Targets: starter_er_mae, starter_outs_mae
• Net: 6 helped vs 3 hurt
• Would help:
• BOS@NYY (miss): starter ER off by 4
• CHW@MIN (miss): starter ER off by 3
• CIN@CHC (miss): starter ER off by 3
• KC@CLE (hit): starter ER off by 7
• PHI@LAA (hit): starter ER off by 3
• SEA@TOR (miss): starter ER off by 4
• Would hurt / risk:
• MIA@WAS (miss): starter projection was accurate (ER off 1); reweight risks disturbing it
• SD@TB (hit): starter projection was accurate (ER off 0); reweight risks disturbing it
• TEX@MIL (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 4 hurt
• Would help:
• BOS@NYY (miss): team runs off by 16
• CHW@MIN (miss): team runs off by 7
• CIN@CHC (miss): team runs off by 5
• COL@ATL (hit): team runs off by 4
• HOU@NYM (miss): team runs off by 4
• KC@CLE (hit): team runs off by 9
• LAD@DET (hit): team runs off by 4
• MIA@WAS (miss): team runs off by 5
• SEA@TOR (miss): team runs off by 8
• TEX@MIL (miss): team runs off by 4
• Would hurt / risk:
• BAL@OAK (hit): team runs already close (off 1); reweight risks disturbing it
• PHI@LAA (hit): team runs already close (off 0); reweight risks disturbing it
• PIT@STL (hit): team runs already close (off 0); reweight risks disturbing it
• SD@TB (hit): team runs already close (off 2); reweight risks disturbing it