Deep slate report 20260917

Which reports actually fed the diagnosis. A source shown as missing was not silently assumed.

High-level causal issue register

Source	Status	Evidence supplied	Path	Bytes
prediction review	loaded	official outcomes and scoring errors	/app/output/reports/prediction_review_20260917.json	229581
forensic deep audit	loaded	probability, provenance, starter and decision contracts	/app/output/reports/PREDICTION_SYSTEM_DEEP_AUDIT_20260917.json	150866
continuous learning	loaded	pipeline-stage quality, calibration, probability lineage, score distribution, and official starter comparisons	/app/output/learning/continuous_learning_daily_20260917.json	1555702
detailed report	loaded	full game and simulation rendering	/app/output/reports/detailed_20260917.html	214325
ultra detailed player audit	loaded	player-level prediction trace	/app/output/reports/20260917_ultra_detailed_player_audit.html	20306685
structural audit	loaded	all invariant findings	/app/output/reports/audit_report_20260917.html	1143113
integrity revision audit	missing	post-Luna remediation proof for immutable revision artifacts	/app/output/reports/audit_report_20260917_integrity_revision.html	n/a

Deterministic findings across data integrity, model validation, calibration, score distribution, pipeline stages, simulation, and coverage. Critical is reserved for broken evidence or an unexplained contract violation; high marks material model/process risk. Each row pairs the action to take (Do this) with the wrong conclusion to avoid (Don't conclude).

Cross-slate trends (persistent-pattern memory)

Escalated persistent patterns

A finding that has repeated at or beyond the escalation streak is no longer single-slate noise. It carries its target registered lever and an action-oriented guardrail.

Finding streaks across the window

Metric direction across the window

Pipeline stage regression

Every stored stage is scored against the same official finals. A later recovery remains visible; it does not erase an unstable intermediate flip.

Probability lineage and calibration

All columns are pick-side probabilities, so away favorites are not accidentally compared using home-side semantics.

Score distribution and margin compression

Narrow predicted margins next to wide actual margins is the signature the score-variance tail is too tight.

Per-game evidence coverage

Every game, in depth (9)

Severity	Domain	Finding and evidence	Affected	Action & guardrail
HIGH	model_validation	Validated run models explain very little holdout variance home/away/total R² = 0.0027/0.0232/0.0056	BOS@TEX, DET@CHW, KC@HOU, LAD@CIN, MIL@PIT, MIN@LAA, OAK@TB, PHI@NYM, SD@COL	Do thisTreat exact run scores as low-confidence estimates, and look for stronger features or a better target before tuning the score tails. Don't concludeA small positive R² only means the model beats guessing the average; it does not make the exact scores reliable.
INFO	coverage	Every settled game received structural QA 9/9 games structurally audited	BOS@TEX, DET@CHW, KC@HOU, LAD@CIN, MIL@PIT, MIN@LAA, OAK@TB, PHI@NYM, SD@COL	Do thisKeep the per-game audit receipt in every report. Don't concludeFull audit coverage does not cancel the findings raised inside the audited games.
INFO	probability_lineage	Material calibrated-to-published movements have recorded lineage 6 material movement(s), all explained by named score-coherence and/or recency transformations	DET@CHW, KC@HOU, LAD@CIN, MIL@PIT, OAK@TB, SD@COL	Do thisKeep showing every probability layer and the named transformation between them. Don't concludeA published probability above the calibrated one is not a governor bypass when a recorded transformation explains it.
INFO	simulation	Simulation produced no winner-direction intervention on this slate flip candidates 0; published flips 0	BOS@TEX, DET@CHW, KC@HOU, LAD@CIN, MIL@PIT, MIN@LAA, OAK@TB, PHI@NYM, SD@COL	Do thisMeasure the simulation's contribution across rolling slates, separating score realization from winner flipping. Don't concludeZero winner flips is not a defect and does not mean the simulation did no score-reconciliation work.

Finding	Persistence	Target lever(s)	Direction	Guardrail
simulation_inert_on_slate	11 consecutive · 11 in window	simulation.allow_score	confirm the simulation module is running, not silently no-oping	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 · 11 in window	training.xgb.n_estimators, training.xgb.max_depth, training.xgb.learning_rate	investigate features/target via isolated training backtest (manual)	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.

Finding	Consecutive slates	Incidence	Present now	Fix type
high_raw_probability_cluster	0	1/11	no	param
pipeline_winner_instability	0	2/11	no	code/diagnostic
score_margin_compression	0	1/11	no	param
simulation_inert_on_slate	11	11/11	yes	code/diagnostic
single_slate_calibration_instability	0	7/11	no	param
weak_run_model_signal	11	11/11	yes	code/diagnostic

Metric	Direction	Series (old → new)
winner_accuracy	regressing	[0.5455, 0.7333, 0.6667, 0.6, 0.6667, 0.5333, 0.6, 0.7, 0.4, 0.6667, 0.4444]
weighted_calibration_error	improving	[0.2611, 0.1825, 0.1069, 0.1171, 0.2091, 0.1711, 0.0827, 0.2812, 0.2795, 0.1661, 0.1016]
margin_bias	regressing	[-0.182, -0.867, -2.133, -1.0, -2.933, -2.667, -1.8, -0.4, -1.667, -1.133, -2.111]
team_run_error_on_losses	improving	[3.8, 6.75, 5.6, 6.5, 5.0, 6.143, 5.0, 6.333, 5.556, 6.4, 3.6]

Stage	Winner hits	Accuracy	Team-run MAE	Total-run MAE	Margin MAE
historical_ml	4/9	44.4%	2.086	2.174	3.582
variance_applied	4/9	44.4%	2.056	2.111	3.667
enhanced_prediction	4/9	44.4%	1.889	2	3.556
embedded_savant_run_environment	4/9	44.4%	1.944	2.111	3.444
fundamental_edge_target_review	4/9	44.4%	1.944	2.111	3.444
capped_target	4/9	44.4%	1.944	2.111	3.444
simulation_result	4/9	44.4%	1.944	2.111	3.444
published_final	4/9	44.4%	1.944	2.111	3.444

Transition	Δ team-run MAE	Δ margin MAE	Games whose winner/tie changed
historical_ml → variance_applied	-0.0307	0.0843	BOS@TEX: TEX → TIE
variance_applied → enhanced_prediction	-0.1667	-0.1111	BOS@TEX: TIE → TEX
enhanced_prediction → embedded_savant_run_environment	0.0555	-0.1112
embedded_savant_run_environment → fundamental_edge_target_review	0	0
fundamental_edge_target_review → capped_target	0	0
capped_target → simulation_result	0	0
simulation_result → published_final	0	0

Game	Outcome	Raw model	Calibrated	Governor	Score coherence	Published	Explanation
BOS@TEX	miss	0.5782	0.5633	0.5633	0.5633	0.5538	no recorded transformation explains the publication delta
DET@CHW	miss	0.5527	0.5837	0.5837	0.5807	0.53	score coherence followed by matchup-recency guardrail
KC@HOU	hit	0.6073	0.5606	0.5606	0.5606	0.5295	score coherence followed by matchup-recency guardrail
LAD@CIN	hit	0.6566	0.7106	0.7106	0.7106	0.56	score coherence followed by matchup-recency guardrail
MIL@PIT	miss	0.6087	0.6562	0.6562	0.6543	0.56	score coherence followed by matchup-recency guardrail
MIN@LAA	miss	0.5399	0.5228	0.5228	0.5228	0.5165	score coherence followed by matchup-recency guardrail
OAK@TB	hit	0.6499	0.6502	0.6502	0.6502	0.58	score coherence followed by matchup-recency guardrail
PHI@NYM	miss	0.562	0.5612	0.5612	0.5612	0.5442	score coherence followed by matchup-recency guardrail
SD@COL	hit	0.7208	0.6972	0.6972	0.6972	0.54	score coherence followed by matchup-recency guardrail

Sides analyzed	Broke on pitching	Broke on offense	Both	Variance	Mean starter ERA error	Mean lineup runs vs proj
18	2	3	3	10	2.16	-0.22

Slate	Games	Winner acc	Mean starter ERA error	Mean lineup runs vs proj	Sides broke on pitching	Sides broke on offense
20260916	15	0.667	2.09	-0.23	4	5
20260915	15	0.400	-1.58	-0.70	1	12
20260914	10	0.700	-0.48	-0.20	2	6
20260913	15	0.600	-0.07	0.03	2	8
20260912	15	0.533	1.35	0.57	6	7
20260911	15	0.667	0.99	0.00	4	9
20260910	5	0.600	-3.19	-1.70		8
20260909	15	0.667	3.36	1.07	8	7
20260908	15	0.733	-0.15	-0.23	7	15
20260907	11	0.545	-0.86	-0.36	5	4
20260906	15	0.600	0.74	-0.30	6	10
20260905	15	0.533	-1.17	-0.07	2	6

Game	Pred margin	Actual margin	Pred − actual
BOS@TEX	1	1	0
DET@CHW	1	2	-1
KC@HOU	1	4	-3
LAD@CIN	4	6	-2
MIL@PIT	2	3	-1
MIN@LAA	1	1	0
OAK@TB	3	9	-6
PHI@NYM	1	3	-2
SD@COL	3	7	-4

BOS@TEXMISS

Predicted: TEX 3-4 (conf +0.11, moderate)Actual: BOS 4-3Brier 0.307total bias +0

Root cause: Backed the wrong side with real conviction

Evidence-backed diagnosis

published system edge 0.11 (moderate); this is not the half-scale distance from 50%
team-run absolute error 2; total bias +0
pregame home-minus-away evidence: season -0.484, recent +0.262, pitcher -0.061, composite -0.154
confidence pipeline reached 0.563 before later publication/recency controls; published home probability was 0.554
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

committed to TEX at 0.11 confidence and it lost

Team	Starter	Projected	Actual	ER err
BOS	Sonny Gray	2 ER / 18 outs	3 ER / 19 outs	1
TEX	Tyler Alexander	1 ER / 15 outs	0 ER / 3 outs	-1

Decision internals: run factors home×0.937 / away×0.998; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

Every value is actual / projected. Run impact approximates how far each hitter's real line moved run expectancy from its projection; it orders the movers and is not a published stat.

Top deviations that drove the game:

Trevor Story (BOS SS): 2-for-4, 1 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Willson Contreras (BOS 1B): 1-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.26]
Corey Seager (TEX SS): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.26]
Nick Sogard (BOS 2B): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Josh Jung (TEX DH): 1-for-4, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.62]

Pitcher versus batter: what actually broke this game

Per-pitch matchup fields are not populated in the current feeds, so this is reconstructed from the settled box: each starter's workload-adjusted ERA, the bullpen share behind him, and the opposing lineup's real production against its projection, with the hitters who beat that projection named.

AWAY pitching

Sonny Gray (BOS) threw 19 outs, 3 ER = 4.26 ERA against a projected 3 ERA (1.26 ERA error)

Bullpen behind him: 0 ER over 3 outs.

The HOME lineup he faced produced 3 runs on 4 hits and 1 HR, against a projected 4 runs / 8 hits (-1 runs vs projection).

Who did the damage to him:

Josh Jung: 1-for-4, 1 HR, 1 RBI [+0.62 runs vs projection] season xwOBA 0.347, elite - real projection error
Danny Jansen: 2-for-3, 0 HR, 1 RBI [+0.47 runs vs projection] season xwOBA 0.262, weak - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

HOME pitching

Tyler Alexander (TEX) threw 3 outs, 0 ER = 0 ERA against a projected 1.8 ERA (-1.8 ERA error)

Bullpen behind him: 2 ER over 3 outs.

The AWAY lineup he faced produced 2 runs on 5 hits and 1 HR, against a projected 4 runs / 8 hits (-2 runs vs projection).

Who did the damage to him:

Trevor Story: 2-for-4, 1 HR, 3 RBI [+1.4 runs vs projection] season xwOBA 0.248, weak - a good day, not a model error
Jahmai Jones: 0-for-2, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.327, above average - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material. exp BB and exp SO are the continuous rate expectations. The integer beside each is a simulation draw reconciled to the opposing simulated batting box, so it is not the line that ranks pitchers: on the 2026-09-07 slate the integers published 39 walks and 121 strikeouts against 51 and 130 real, while these expectations landed on 50.4 and 132.6.)

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
BOS	Sonny Gray (Starter)	19 / 18	4.26 / 3	3 / 2	3 / 6	1 / 0	3 / 1	1.73	6 / 5	6.05	venue away: 2.62 vs 2.91 ERA (ok) · opp run factor 0.94 (average) · 27 prior appearances this season
BOS	Garrett Whitlock (Middle Reliever (High Leverage))	3 / 3	0 / 18 (-18.00)	0 / 2	0 / 2	0 / 0	1 / 2	0.11	0 / 0	1.37
TEX	Tyler Alexander (Starter)	3 / 15	0 / 1.8	0 / 1	0 / 3	0 / 1	1 / 0	0.85	0 / 1	4.43	venue home: 4.5 vs 0 ERA (ok) · opp run factor 1 (average) · 68 prior appearances this season
TEX	Chase Silseth (Middle Reliever (High Leverage))	2 / 6	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0.61	2 / 1	1.29
TEX	Jakob Junis (Setup Man)	1 / 6	54 / 9 (+45.00)	2 / 2	0 / 1	0 / 1	1 / 2	0.25	0 / 1	0.66

Hitters: actual / projected (sorted by run impact vs projection)

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Trevor Story (SS)	BOS	4 / 4	2 / 1	0 / 0	0 / 0	1 / 0	3 / 1	1 / 0	0 / 0	2 / 1	1.4	0.248 weak	good day
Willson Contreras (1B)	BOS	3 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	1 / 1	-1.26	0.377 elite	bad day
Corey Seager (SS)	TEX	4 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	1 / 1	-1.26	0.352 elite	bad day
Nick Sogard (2B)	BOS	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	-0.8	0.318 average	no signal
Josh Jung (DH)	TEX	4 / 4	1 / 1	0 / 1	0 / 0	1 / 0	1 / 0	1 / 1	0 / 0	0 / 1	0.62	0.347 elite	projection too low
Roman Anthony (LF)	BOS	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	0 / 1	-0.47	0.348 elite	—
Jarren Duran (CF)	BOS	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.297 below average	—
Ezequiel Duran (3B)	TEX	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	-0.47	0.306 average	—
Brandon Nimmo (RF)	TEX	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	3 / 1	-0.47	0.345 elite	—
Danny Jansen (C)	TEX	3 / 3	2 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 1	0.47	0.262 weak	—
Jahmai Jones (DH)	BOS	2 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0.33	0.327 above average	—
Caleb Durbin (3B)	BOS	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	0 / 0	-0.14	0.293 below average	—
Wyatt Langford (LF)	TEX	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	2 / 1	0 / 1	-0.14	0.316 average	—
Jake Burger (1B)	TEX	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	1 / 0	0 / 1	-0.14	0.315 average	—
Nicky Lopez (2B)	TEX	1 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	-0.14	0.257 weak	—
Wilyer Abreu (RF)	BOS	3 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 0	0 / 1	0.02	0.351 elite	—
Adley Rutschman (C)	BOS	4 / 4	1 / 1	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 1	0	0.331 above average	—
Evan Carter (CF)	TEX	2 / 3	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	1 / 1	0	0.279 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.484	positive favors home; negative favors away
Recent form	0.262	positive favors home; negative favors away
Starting-pitcher matchup	-0.061	positive favors home; negative favors away
Composite	-0.154	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.91-4.29	n/a	Base ML historical model prediction
variance_applied	4-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	3-4	0.554	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run)
embedded_savant_run_environment	3-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-4	n/a	Capped at max 12 realistic runs
simulation_result	3-4	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.91-4.29	4-4	-0.38	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-4	3-4	1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run)
embedded_savant_run_environment	3-4	3-4	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-4	3-4	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-4	3-4	0	Capped at max 12 realistic runs
simulation_result	3-4	3-4	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

These are pregame model adjustments, not claims about actual postgame player error.

Team	Player	Role	Net effect	Largest change
BOS	Sonny Gray	pitcher / Starter	better	predicted_SO 2
TEX	Danny Jansen	batter / C	worse	predicted_RBI 1.67
TEX	Jake Burger	batter / 1B	better	predicted_RBI 1.45
BOS	Caleb Durbin	batter / 3B	worse	predicted_AB -1.377
BOS	Trevor Story	batter / SS	worse	predicted_SO -1.339
TEX	Corey Seager	batter / SS	worse	predicted_AB -1.27
BOS	Roman Anthony	batter / LF	better	predicted_SO -1.217
BOS	Jarren Duran	batter / CF	worse	predicted_SO -1.143

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	BOS pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	BOS
info	CURRENT_PITCHER_DUPLICATION	TEX pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	TEX
info	CURRENT_FINALIZED_MATCH_BOX	BOS model-prediction batting totals differ from final simulation box score (AB: current=32, sim=31; H: current=8, sim=4; 2B: current=2, sim=0; HR: current=1, sim=3; RBI: current=4, sim=3; R: current=4, sim=3; BB: current=3, sim=2; SO: current=8, sim=3).	BOS
info	CURRENT_PLAYER_MATCH_BOX	BOS current finalized hitter lines are model predictions and differ from the simulation box score: Roman Anthony: HR current=0 box=1, RBI current=0 box=1, R current=0 box=1, BB current=1 box=0; Adley Rutschman: 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0, SO current=1 box=0; Trevor Story: H current=1 box=0, RBI current=1 box=0, SO current=1 box=0; Jarren Duran: H current=1 box=0, SO current=1 box=0; Jahmai Jones: H current=0 box=1, HR current=0 box=1, RBI current=0 box=1, R current=0 box=1; Wilyer Abreu: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Caleb Durbin: AB current=3 box=2, H current=1 box=0, R current=1 box=0, BB current=0 box=1; Willson Contreras: BB current=1 box=0	BOS
info	CURRENT_FINALIZED_MATCH_BOX	TEX model pitcher projections differ from final simulation pitching box score (HR: current=2, sim=3).	TEX
info	CURRENT_FINALIZED_MATCH_BOX	TEX model-prediction batting totals differ from final simulation box score (2B: current=1, sim=5; HR: current=1, sim=0; BB: current=2, sim=3; SO: current=8, sim=5).	TEX
info	CURRENT_PLAYER_MATCH_BOX	TEX current finalized hitter lines are model predictions and differ from the simulation box score: Josh Jung: R current=1 box=0, SO current=1 box=0; Ezequiel Duran: 2B current=0 box=1, RBI current=1 box=0; Wyatt Langford: H current=1 box=2, 2B current=0 box=1, BB current=1 box=0; Corey Seager: AB current=4 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Jake Burger: 2B current=0 box=1, RBI current=1 box=2, BB current=0 box=1, SO current=1 box=0; Brandon Nimmo: H current=1 box=2, 2B current=0 box=1, RBI current=1 box=0, R current=0 box=1; Nicky Lopez: H current=1 box=0, BB current=0 box=1; Danny Jansen: AB current=3 box=4, RBI current=0 box=2, R current=0 box=1, SO current=1 box=2	TEX
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

DET@CHWMISS

Predicted: DET 5-4 (conf -0.06, lean)Actual: CHW 1-3Brier 0.281total bias +5

Root cause: Starting-pitcher projection was wrong

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 5; total bias +5
pregame home-minus-away evidence: season -0.442, recent -0.770, pitcher +0.084, composite -0.431
confidence pipeline reached 0.584 before later publication/recency controls; published home probability was 0.470
score-volatility actions: supported_blowout_margin_expansion
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 5 (9 vs 4)
DET offense over-projected by 4
Erick Fedde (CHW) allowed 4 fewer ER than projected (1 vs 5)

Team	Starter	Projected	Actual	ER err
DET	Framber Valdez	3 ER / 18 outs	3 ER / 15 outs	0
CHW	Erick Fedde	5 ER / 14 outs	1 ER / 14 outs	-4

Decision internals: run factors home×1.055 / away×0.972; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

Every value is actual / projected. Run impact approximates how far each hitter's real line moved run expectancy from its projection; it orders the movers and is not a published stat.

Top deviations that drove the game:

Riley Greene (DET LF): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.64]
Dillon Dingler (DET C): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.62]

Pitcher versus batter: what actually broke this game

AWAY pitching

Framber Valdez (DET) threw 15 outs, 3 ER = 5.4 ERA against a projected 4.5 ERA (0.9 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 2 runs on 3 hits and 0 HR, against a projected 2 runs / 5 hits (0 runs vs projection).

Who did the damage to him:

Brenton Doyle: 2-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.236, weak - a good day, not a model error
Jake Rogers: 0-for-3, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.269, weak - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

HOME pitching

No settled pitching line for this club. The forecast named Erick Fedde (CHW) at a projected 9.64 ERA, but the per-player feed carries no pitching row for him in this game, so the actual-versus-projected comparison cannot be made on this side. That is a source-coverage gap in the settled box, not a forecasting failure.

Bullpen behind him: 0 ER over 0 outs.

The AWAY lineup he faced produced 1 runs on 8 hits and 0 HR, against a projected 6 runs / 8 hits (-5 runs vs projection).

Who did the damage to him:

Kevin McGonigle: 2-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.357, elite - a good day, not a model error
Spencer Torkelson: 1-for-3, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.312, average - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
DET	Framber Valdez (Starter)	15 / 18	5.4 / 4.5	3 / 3	3 / 5	1 / 0	4 / 5	2.22	8 / 8	3.39	venue away: 4.02 vs 4 ERA (ok) · opp run factor 1.06 (average) · 29 prior appearances this season

Hitters: actual / projected (sorted by run impact vs projection)

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Riley Greene (LF)	DET	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	2 / 1	-0.64	0.372 elite	bad day
Dillon Dingler (C)	DET	4 / 4	1 / 1	1 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	0 / 1	-0.62	0.338 above average	bad day
Kevin McGonigle (3B)	DET	4 / 4	2 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	1 / 0	0.47	0.357 elite	—
Zach McKinstry (RF)	DET	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	-0.47	0.285 below average	—
Spencer Torkelson (1B)	DET	3 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	0.47	0.312 average	—
Javier Báez (2B)	DET	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	3 / 1	-0.47	0.255 weak	—
Brenton Doyle (CF)	CHW	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0.47	0.236 weak	—
Miguel Vargas (3B)	CHW	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	1 / 1	-0.47	0.379 elite	—
Randal Grichuk (LF)	CHW	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 0	1 / 0	0 / 1	-0.45	0.336 above average	—
Jake Rogers (C)	CHW	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	2 / 1	0.33	0.269 weak	—
Chase Meidroth (2B)	CHW	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0 / 1	-0.33	0.297 below average	—
Tommy Pham (DH)	CHW	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 0	2 / 1	-0.14	0.21 weak	—
Gleyber Torres (DH)	DET	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	1 / 1	1 / 1	0	0.325 above average	—
Max Clark (CF)	DET	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	1 / 1	0	0.311 average	—
John Peck (SS)	DET	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	0	0.332 above average	—
Luisangel Acuña (SS)	CHW	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0	0.265 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.442	positive favors home; negative favors away
Recent form	-0.77	positive favors home; negative favors away
Starting-pitcher matchup	0.084	positive favors home; negative favors away
Composite	-0.431	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.54-4.38	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	0.53	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	5-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	5-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	5-4	n/a	Capped at max 12 realistic runs
simulation_result	5-4	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.54-4.38	5-4	0.844	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	5-4	-0	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	5-4	5-4	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	5-4	5-4	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	5-4	5-4	-0	Capped at max 12 realistic runs
simulation_result	5-4	5-4	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

These are pregame model adjustments, not claims about actual postgame player error.

Team	Player	Role	Net effect	Largest change
DET	Javier Báez	batter / 2B	better	predicted_RBI 3.66
DET	Framber Valdez	pitcher / Starter	better	predicted_SO 3
CHW	Randal Grichuk	batter / LF	worse	predicted_AB 2.5
CHW	Chase Meidroth	batter / 2B	worse	predicted_SO 2.069
CHW	Miguel Vargas	batter / 3B	worse	predicted_H 1.89
DET	Max Clark	batter / CF	worse	predicted_AB -1.73
CHW	Munetaka Murakami	batter / 1B	worse	predicted_AB -1.578
CHW	Tommy Pham	batter / DH	better	predicted_AB 1.5

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	CHW pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	CHW
info	CURRENT_PITCHER_DUPLICATION	DET pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	DET
info	CURRENT_FINALIZED_MATCH_BOX	CHW model-prediction batting totals differ from final simulation box score (AB: current=31, sim=33; H: current=7, sim=8; 2B: current=1, sim=0; HR: current=1, sim=0; BB: current=4, sim=5; SO: current=8, sim=12; SF: current=0, sim=2).	CHW
info	CURRENT_PLAYER_MATCH_BOX	CHW current finalized hitter lines are model predictions and differ from the simulation box score: Tommy Pham: AB current=3 box=4, R current=0 box=1, BB current=0 box=1, SO current=1 box=2; Munetaka Murakami: AB current=4 box=2, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Randal Grichuk: AB current=4 box=5, 2B current=1 box=0, RBI current=1 box=0, SO current=1 box=2; Brenton Doyle: RBI current=0 box=1, R current=0 box=1, SF current=0 box=1; Jake Rogers: BB current=0 box=1, SO current=1 box=2; Braden Montgomery: AB current=3 box=4, RBI current=0 box=1, BB current=1 box=0; Chase Meidroth: H current=1 box=0, RBI current=0 box=1, BB current=1 box=0, SO current=1 box=3; Miguel Vargas: AB current=4 box=5, H current=1 box=3, RBI current=1 box=0, R current=1 box=0	CHW
info	CURRENT_FINALIZED_MATCH_BOX	DET model-prediction batting totals differ from final simulation box score (AB: current=31, sim=32; H: current=8, sim=5; 2B: current=1, sim=2; HR: current=1, sim=2; RBI: current=4, sim=5; R: current=6, sim=5; BB: current=4, sim=3; SO: current=7, sim=6).	DET
info	CURRENT_PLAYER_MATCH_BOX	DET current finalized hitter lines are model predictions and differ from the simulation box score: Gleyber Torres: AB current=4 box=3, H current=1 box=0; Kevin McGonigle: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Zach McKinstry: AB current=3 box=4, R current=0 box=1; Riley Greene: RBI current=1 box=0, BB current=1 box=0; Spencer Torkelson: AB current=3 box=4, R current=1 box=0, BB current=1 box=0, SO current=1 box=0; Javier Báez: AB current=3 box=2, H current=1 box=2, HR current=0 box=2, RBI current=0 box=4; Dillon Dingler: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Max Clark: H current=1 box=0, R current=1 box=0, BB current=0 box=1, SO current=1 box=0	DET
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

KC@HOUHIT

Predicted: HOU 4-5 (conf +0.06, lean)Actual: HOU 2-6Brier 0.221total bias +1

Root cause: Solid correct call

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 3; total bias +1
pregame home-minus-away evidence: season +0.118, recent -0.234, pitcher +0.011, composite +0.211
confidence pipeline reached 0.561 before later publication/recency controls; published home probability was 0.530
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

Seth Lugo (KC) allowed 3 more ER than projected (4 vs 1)
Miguel Ullola (HOU) allowed 3 fewer ER than projected (1 vs 4)

Team	Starter	Projected	Actual	ER err
KC	Seth Lugo	1 ER / 15 outs	4 ER / 15 outs	3
HOU	Miguel Ullola	4 ER / 9 outs	1 ER / 2 outs	-3

Decision internals: run factors home×0.996 / away×0.985; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

Every value is actual / projected. Run impact approximates how far each hitter's real line moved run expectancy from its projection; it orders the movers and is not a published stat.

Top deviations that drove the game:

LaMonte Wade Jr. (HOU 1B): 2-for-4, 1 HR, 1 RBI (projected 0.0 H, 0.0 HR) [run impact 1.85]
Taylor Trammell (HOU LF): 1-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.26]
Jeremy Peña (HOU SS): 2-for-5, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 1.09]
Daulton Varsho (HOU CF): 1-for-4, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Jac Caglianone (KC DH): 0-for-2, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.78]

Starter miss: Seth Lugo (KC) posted 7.20 ERA vs projected 1.80 (4 ER in 15 outs); he is materially worse pitching away (6.03 ERA there vs 4.09 on the other side), a venue-role weakness the model under-weighted.

What to do: Enable/raise the venue-split weight (context.venue_adjustment_weight) and validate in shadow; the split is credible for this starter.

Pitcher versus batter: what actually broke this game

AWAY pitching

Seth Lugo (KC) threw 15 outs, 4 ER = 7.2 ERA against a projected 1.8 ERA (5.4 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 6 runs on 11 hits and 5 HR, against a projected 4 runs / 8 hits (2 runs vs projection).

Who did the damage to him:

LaMonte Wade Jr.: 2-for-4, 1 HR, 1 RBI [+1.85 runs vs projection] season xwOBA 0.272, weak - a good day, not a model error
Taylor Trammell: 1-for-3, 1 HR, 1 RBI [+1.26 runs vs projection] season xwOBA 0.289, below average - a good day, not a model error
Jeremy Peña: 2-for-5, 1 HR, 2 RBI [+1.09 runs vs projection] season xwOBA 0.327, above average - real projection error

Verdict: both: the starter was hit hard AND the lineup underproduced

HOME pitching

Miguel Ullola (HOU) threw 2 outs, 1 ER = 13.5 ERA against a projected 12 ERA (1.5 ERA error) only 6 prior appearances in the covered season - the projection had little of his own record to stand on

Bullpen behind him: 0 ER over 4 outs.

The AWAY lineup he faced produced 2 runs on 3 hits and 0 HR, against a projected 4 runs / 7 hits (-2 runs vs projection).

Who did the damage to him:

John Rave: 1-for-3, 0 HR, 0 RBI [+0.31 runs vs projection] season xwOBA 0.318, average - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
KC	Seth Lugo (Starter)	15 / 15	7.2 / 1.8 (+5.40)	4 / 1	8 / 5	4 / 0	1 / 1	1.41	5 / 5	3.56	venue away: 6.03 vs 4.09 ERA (worse here) · opp run factor 1 (average) · 30 prior appearances this season
HOU	Miguel Ullola (Starter)	2 / 9	13.5 / 12	1 / 4	0 / 5	0 / 2	2 / 1	0	0 / 5	0	opp run factor 0.99 (weak) · 6 prior appearances this season
HOU	Bryan Abreu (Setup Man)	1 / 6	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0.53	1 / 1	1.34
HOU	Josh Hader (Closer)	3 / 6	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0.2	0 / 0	1.22

Hitters: actual / projected (sorted by run impact vs projection)

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
LaMonte Wade Jr. (1B)	HOU	4 / 3	2 / 0	1 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 1	1 / 1	1.85	0.272 weak	good day
Taylor Trammell (LF)	HOU	3 / 3	1 / 1	0 / 0	0 / 0	1 / 0	1 / 0	2 / 0	1 / 0	1 / 1	1.26	0.289 below average	good day
Jeremy Peña (SS)	HOU	5 / 4	2 / 1	0 / 1	0 / 0	1 / 0	2 / 1	1 / 1	0 / 0	1 / 1	1.09	0.327 above average	projection too low
Daulton Varsho (CF)	HOU	4 / 4	1 / 1	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 0	1 / 1	0.93	0.285 below average	good day
Jac Caglianone (DH)	KC	2 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	-0.78	0.353 elite	bad day
Michael Massey (2B)	KC	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.309 average	—
Salvador Perez (1B)	KC	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 1	-0.47	0.31 average	—
Bobby Witt Jr. (SS)	KC	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	2 / 1	-0.47	0.379 elite	—
Maikel Garcia (3B)	KC	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	0 / 1	-0.47	0.317 average	—
Cam Smith (RF)	HOU	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.319 average	—
Carter Jensen (C)	KC	3 / 4	2 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	1 / 1	0 / 1	-0.46	0.314 average	—
Isaac Collins (LF)	KC	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	-0.33	0.298 below average	—
John Rave (RF)	KC	3 / 3	1 / 1	1 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	1 / 1	0.31	0.318 average	—
Jose Altuve (2B)	HOU	4 / 4	1 / 1	1 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	0.31	0.271 weak	—
Yordan Alvarez (DH)	HOU	4 / 4	2 / 1	0 / 0	0 / 0	1 / 1	1 / 1	1 / 1	0 / 1	0 / 1	0.14	0.443 elite	—
Kyle Isbel (CF)	KC	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.26 weak	—
Isaac Paredes (3B)	HOU	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0 / 0	0	0.319 average	—
Yainer Diaz (C)	HOU	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 0	0	0.304 average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.118	positive favors home; negative favors away
Recent form	-0.234	positive favors home; negative favors away
Starting-pitcher matchup	0.011	positive favors home; negative favors away
Composite	0.211	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.5-5.67	n/a	Base ML historical model prediction
variance_applied	4-6	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.53	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	n/a	Capped at max 12 realistic runs
simulation_result	4-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.5-5.67	4-6	-0.168	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-6	4-5	-1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	4-5	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	4-5	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	4-5	0	Capped at max 12 realistic runs
simulation_result	4-5	4-5	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

These are pregame model adjustments, not claims about actual postgame player error.

Team	Player	Role	Net effect	Largest change
HOU	Cam Smith	batter / RF	better	predicted_RBI 4.67
KC	Bobby Witt Jr.	batter / SS	worse	predicted_AB -1.59
HOU	Yordan Alvarez	batter / DH	worse	predicted_SO 1.231
KC	Maikel Garcia	batter / 3B	worse	predicted_H -1.19
HOU	Jose Altuve	batter / 2B	worse	predicted_SO 1.161
KC	John Rave	batter / RF	worse	predicted_AB 1.077
KC	Michael Massey	batter / 2B	worse	predicted_AB -1.06
HOU	Miguel Ullola	pitcher / Starter	worse	predicted_IP -1

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	HOU pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	HOU
info	CURRENT_PITCHER_DUPLICATION	KC pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	KC
info	CURRENT_FINALIZED_MATCH_BOX	HOU model-prediction batting totals differ from final simulation box score (AB: current=33, sim=34; H: current=8, sim=10; 2B: current=1, sim=2; 3B: current=0, sim=1; RBI: current=4, sim=5; R: current=4, sim=5; BB: current=3, sim=2; SO: current=7, sim=9).	HOU
info	CURRENT_PLAYER_MATCH_BOX	HOU current finalized hitter lines are model predictions and differ from the simulation box score: Yordan Alvarez: AB current=4 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Jose Altuve: H current=1 box=2, SO current=1 box=2; Isaac Paredes: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Jeremy Peña: AB current=4 box=5, H current=1 box=2, RBI current=1 box=0, SO current=1 box=2; Yainer Diaz: RBI current=1 box=0, R current=0 box=1, SO current=0 box=1; LaMonte Wade Jr.: BB current=1 box=0, HBP current=0 box=1; Taylor Trammell: R current=0 box=1, BB current=0 box=1; Cam Smith: AB current=3 box=4, H current=1 box=3, 2B current=0 box=1, HR current=0 box=1	HOU
info	CURRENT_FINALIZED_MATCH_BOX	KC model-prediction batting totals differ from final simulation box score (AB: current=33, sim=32; H: current=7, sim=5; HR: current=1, sim=2; BB: current=3, sim=2; SO: current=9, sim=6).	KC
info	CURRENT_PLAYER_MATCH_BOX	KC current finalized hitter lines are model predictions and differ from the simulation box score: Jac Caglianone: 2B current=1 box=0, HR current=0 box=1, R current=0 box=1; Michael Massey: AB current=4 box=3, BB current=0 box=1, SO current=1 box=0; Salvador Perez: R current=0 box=1; Bobby Witt Jr.: AB current=4 box=3, HR current=0 box=1, RBI current=0 box=1; Maikel Garcia: H current=1 box=0, RBI current=1 box=0, R current=1 box=0; John Rave: AB current=3 box=4, H current=1 box=0, R current=1 box=0, SO current=1 box=0; Carter Jensen: 2B current=0 box=1, HR current=1 box=0, BB current=1 box=0; Kyle Isbel: SO current=1 box=0	KC
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

LAD@CINHIT

Predicted: LAD 9-5 (conf -0.12, moderate)Actual: LAD 8-2Brier 0.194total bias +4

Root cause: Solid correct call

Evidence-backed diagnosis

published system edge 0.12 (moderate); this is not the half-scale distance from 50%
team-run absolute error 4; total bias +4
pregame home-minus-away evidence: season -1.929, recent -1.414, pitcher -0.059, composite -2.264
confidence pipeline reached 0.711 before later publication/recency controls; published home probability was 0.440
score-volatility actions: supported_blowout_margin_expansion
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 4 (14 vs 10)
CIN offense over-projected by 3
Justin Wrobleski (LAD) allowed 3 fewer ER than projected (0 vs 3)
Brady Singer (CIN) allowed 3 more ER than projected (7 vs 4)

Team	Starter	Projected	Actual	ER err
LAD	Justin Wrobleski	3 ER / 15 outs	0 ER / 9 outs	-3
CIN	Brady Singer	4 ER / 14 outs	7 ER / 10 outs	3

Decision internals: run factors home×1.002 / away×1.06; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

Every value is actual / projected. Run impact approximates how far each hitter's real line moved run expectancy from its projection; it orders the movers and is not a published stat.

Top deviations that drove the game:

Kyle Tucker (LAD RF): 4-for-5, 1 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 2.63]
Miguel Rojas (LAD SS): 2-for-5, 1 HR, 1 RBI (projected 0.0 H, 0.0 HR) [run impact 1.87]
Elly De La Cruz (CIN DH): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.73]
Sal Stewart (CIN 1B): 2-for-4, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Dane Myers (CIN CF): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -1.11]

Starter miss: Brady Singer (CIN) posted 18.90 ERA vs projected 7.71 (7 ER in 10 outs); he faced a strong offense (opponent run factor 1.06, bucket strong).

What to do: Consider raising context.opponent_strength_weight; validate across the rolling window before promotion.

Pitcher versus batter: what actually broke this game

AWAY pitching

Justin Wrobleski (LAD) threw 9 outs, 0 ER = 0 ERA against a projected 5.4 ERA (-5.4 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 2 runs on 6 hits and 1 HR, against a projected 3 runs / 7 hits (-1 runs vs projection).

Who did the damage to him:

Sal Stewart: 2-for-4, 1 HR, 1 RBI [+1.4 runs vs projection] season xwOBA 0.346, elite - real projection error
TJ Friedl: 2-for-3, 0 HR, 0 RBI [+0.94 runs vs projection] season xwOBA 0.249, weak - a good day, not a model error
Héctor Rodríguez: 1-for-3, 0 HR, 0 RBI [+0.62 runs vs projection] season xwOBA 0.259, weak - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

HOME pitching

Brady Singer (CIN) threw 10 outs, 7 ER = 18.9 ERA against a projected 7.71 ERA (11.19 ERA error)

Bullpen behind him: 0 ER over 8 outs.

The AWAY lineup he faced produced 8 runs on 11 hits and 2 HR, against a projected 4 runs / 7 hits (4 runs vs projection).

Who did the damage to him:

Kyle Tucker: 4-for-5, 1 HR, 3 RBI [+2.63 runs vs projection] season xwOBA 0.33, above average - real projection error
Miguel Rojas: 2-for-5, 1 HR, 1 RBI [+1.87 runs vs projection] season xwOBA 0.304, average - a good day, not a model error
Teoscar Hernández: 1-for-4, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.335, above average - a good day, not a model error

Verdict: pitching: the starter gave up runs at a far worse rate than projected

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
LAD	Justin Wrobleski (Starter)	9 / 15	0 / 5.4 (-5.40)	0 / 3	0 / 6	0 / 0	0 / 1	1.65	3 / 4	5.27	venue away: 3.86 vs 3.43 ERA (ok) · opp run factor 1 (weak) · 24 prior appearances this season
CIN	Brady Singer (Starter)	10 / 14	18.9 / 7.71 (+11.19)	7 / 4	9 / 7	1 / 0	4 / 4	2.7	0 / 8	4.46	venue home: 4.5 vs 6.37 ERA (ok) · opp run factor 1.06 (strong) · 29 prior appearances this season
CIN	Luis Mey (Setup Man)	5 / 3	0 / 0	0 / 0	1 / 1	0 / 0	0 / 1	0.72	2 / 1	1.59
CIN	Tony Santillan (Middle Reliever (High Leverage))	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	0.82	0 / 0	1.04

Hitters: actual / projected (sorted by run impact vs projection)

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Kyle Tucker (RF)	LAD	5 / 4	4 / 1	2 / 0	0 / 0	1 / 0	3 / 1	2 / 1	0 / 1	0 / 1	2.63	0.33 above average	projection too low
Miguel Rojas (SS)	LAD	5 / 3	2 / 0	0 / 0	0 / 0	1 / 0	1 / 0	2 / 0	0 / 0	0 / 0	1.87	0.304 average	no signal
Elly De La Cruz (DH)	CIN	4 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	3 / 1	-1.73	0.378 elite	bad day
Sal Stewart (1B)	CIN	4 / 4	2 / 1	0 / 0	0 / 0	1 / 0	1 / 1	2 / 1	0 / 0	1 / 1	1.4	0.346 elite	projection too low
Dane Myers (CF)	CIN	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	-1.11	0.313 average	no signal
TJ Friedl (LF)	CIN	3 / 3	2 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0.94	0.249 weak	good day
Héctor Rodríguez (RF)	CIN	3 / 4	1 / 1	0 / 0	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0.62	0.259 weak	good day
Max Muncy (3B)	LAD	4 / 4	2 / 1	1 / 0	0 / 0	0 / 1	2 / 1	0 / 1	0 / 1	0 / 1	-0.48	0.345 elite	—
Tommy Edman (CF)	LAD	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 1	-0.47	0.306 average	—
Juan Brito (2B)	CIN	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	0 / 0	0 / 1	-0.47	0.292 below average	—
Matt McLain (SS)	CIN	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	-0.47	0.302 below average	—
Teoscar Hernández (LF)	LAD	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	1 / 0	1 / 1	0.33	0.335 above average	—
Tyler Stephenson (C)	CIN	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	0 / 1	-0.33	0.33 above average	—
Hunter Feduccia (C)	LAD	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	1 / 1	0.19	0.319 average	—
Freddie Freeman (1B)	LAD	3 / 4	2 / 1	0 / 1	0 / 0	0 / 0	1 / 0	2 / 0	1 / 1	0 / 1	0.16	0.365 elite	—
Alex Freeland (2B)	LAD	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	0 / 1	-0.14	0.294 below average	—
Ke'Bryan Hayes (3B)	CIN	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0	0.295 below average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-1.929	positive favors home; negative favors away
Recent form	-1.414	positive favors home; negative favors away
Starting-pitcher matchup	-0.059	positive favors home; negative favors away
Composite	-2.264	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.84-4.01	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	8-5	0.56	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	9-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	9-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	9-5	n/a	Capped at max 12 realistic runs
simulation_result	9-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.84-4.01	5-4	0.171	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	8-5	2	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	8-5	9-5	1	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	9-5	9-5	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	9-5	9-5	-0	Capped at max 12 realistic runs
simulation_result	9-5	9-5	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

These are pregame model adjustments, not claims about actual postgame player error.

Team	Player	Role	Net effect	Largest change
LAD	Kyle Tucker	batter / RF	better	predicted_RBI 5.46
LAD	Josue De Paula	batter / DH	boosted	predicted_H 2.6
LAD	Tommy Edman	batter / CF	worse	predicted_AB 2.394
CIN	Dane Myers	batter / CF	better	predicted_H 1.97
CIN	Elly De La Cruz	batter / DH	worse	predicted_H 1.862
LAD	Alex Freeland	batter / 2B	worse	predicted_BB 1.675
LAD	Miguel Rojas	batter / SS	better	predicted_AB 1.5
LAD	Hunter Feduccia	batter / C	worse	predicted_AB 1.462

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	CIN pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	CIN
info	CURRENT_PITCHER_DUPLICATION	LAD pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	LAD
info	CURRENT_FINALIZED_MATCH_BOX	CIN model pitcher projections differ from final simulation pitching box score (ER: current=8, sim=9).	CIN
info	CURRENT_FINALIZED_MATCH_BOX	CIN model-prediction batting totals differ from final simulation box score (AB: current=34, sim=38; H: current=7, sim=11; RBI: current=4, sim=5; R: current=3, sim=5; BB: current=3, sim=1; HBP: current=0, sim=1).	CIN
info	CURRENT_PLAYER_MATCH_BOX	CIN current finalized hitter lines are model predictions and differ from the simulation box score: Elly De La Cruz: AB current=4 box=5, H current=1 box=3, HR current=1 box=0, RBI current=1 box=0; Juan Brito: R current=0 box=1; Matt McLain: RBI current=0 box=1, SO current=1 box=0; Héctor Rodríguez: RBI current=0 box=1, R current=0 box=1; Ke'Bryan Hayes: AB current=3 box=4, SO current=1 box=2; TJ Friedl: AB current=3 box=4, H current=0 box=1, SO current=1 box=0; Sal Stewart: 2B current=0 box=1, BB current=0 box=1, SO current=1 box=2; Tyler Stephenson: H current=1 box=0, RBI current=1 box=0, BB current=1 box=0, SO current=1 box=2	CIN
info	CURRENT_FINALIZED_MATCH_BOX	LAD model-prediction batting totals differ from final simulation box score (AB: current=32, sim=38; H: current=8, sim=11; 2B: current=1, sim=2; 3B: current=0, sim=1; RBI: current=4, sim=9; R: current=4, sim=9; BB: current=4, sim=8; SO: current=7, sim=9).	LAD
info	CURRENT_PLAYER_MATCH_BOX	LAD current finalized hitter lines are model predictions and differ from the simulation box score: Tommy Edman: AB current=4 box=6; Max Muncy: 2B current=0 box=1, HR current=1 box=0, RBI current=1 box=0; Josue De Paula: AB current=3 box=5, H current=1 box=3, R current=0 box=1, BB current=1 box=0; Kyle Tucker: H current=1 box=2, 3B current=0 box=1, HR current=0 box=1, RBI current=1 box=6; Freddie Freeman: H current=1 box=2, 2B current=1 box=0, RBI current=0 box=1, R current=0 box=1; Miguel Rojas: AB current=3 box=4, H current=0 box=1, 2B current=0 box=1, R current=0 box=1; Teoscar Hernández: BB current=0 box=1, SO current=1 box=2; Hunter Feduccia: AB current=3 box=4, H current=1 box=0, R current=0 box=1, BB current=0 box=1	LAD
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

MIL@PITMISS

Predicted: MIL 6-4 (conf -0.12, moderate)Actual: PIT 4-7Brier 0.314total bias -1

Root cause: Backed the wrong side with real conviction

Evidence-backed diagnosis

published system edge 0.12 (moderate); this is not the half-scale distance from 50%
team-run absolute error 5; total bias -1
pregame home-minus-away evidence: season -1.170, recent -1.138, pitcher -0.061, composite -1.423
confidence pipeline reached 0.656 before later publication/recency controls; published home probability was 0.440
score-volatility actions: supported_blowout_margin_expansion
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

PIT offense under-projected by 3
committed to MIL at 0.12 confidence and it lost

Team	Starter	Projected	Actual	ER err
MIL	Kyle Harrison	3 ER / 15 outs	4 ER / 5 outs	1
PIT	Wilber Dotel	1 ER / 15 outs	1 ER / 6 outs	0

Decision internals: run factors home×0.95 / away×1.011; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

Every value is actual / projected. Run impact approximates how far each hitter's real line moved run expectancy from its projection; it orders the movers and is not a published stat.

Top deviations that drove the game:

Bo Naylor (MIL C): 1-for-3, 1 HR, 2 RBI (projected 0.0 H, 0.0 HR) [run impact 1.73]
Jackson Chourio (MIL LF): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Rafael Flores Jr. (PIT 1B): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Henry Davis (PIT C): 1-for-2, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 1.26]
Luis Lara (MIL CF): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.78]

Starter miss: Kyle Harrison (MIL) posted 21.60 ERA vs projected 5.40 (4 ER in 5 outs); he is materially worse pitching away (6.21 ERA there vs 1.26 on the other side), a venue-role weakness the model under-weighted.

What to do: Enable/raise the venue-split weight (context.venue_adjustment_weight) and validate in shadow; the split is credible for this starter.

Pitcher versus batter: what actually broke this game

AWAY pitching

Kyle Harrison (MIL) threw 5 outs, 4 ER = 21.6 ERA against a projected 5.4 ERA (16.2 ERA error)

Bullpen behind him: 2 ER over 3 outs.

The HOME lineup he faced produced 7 runs on 7 hits and 1 HR, against a projected 5 runs / 9 hits (2 runs vs projection).

Who did the damage to him:

Henry Davis: 1-for-2, 1 HR, 2 RBI [+1.26 runs vs projection] season xwOBA 0.285, below average - a good day, not a model error
Brandon Lowe: 1-for-2, 0 HR, 1 RBI [+0.64 runs vs projection] season xwOBA 0.323, above average - real projection error
Konnor Griffin: 1-for-4, 0 HR, 2 RBI [+0.33 runs vs projection] season xwOBA 0.3, below average - a good day, not a model error

Verdict: pitching: the starter gave up runs at a far worse rate than projected

HOME pitching

Wilber Dotel (PIT) threw 6 outs, 1 ER = 4.5 ERA against a projected 1.8 ERA (2.7 ERA error)

Bullpen behind him: 2 ER over 8 outs.

The AWAY lineup he faced produced 3 runs on 7 hits and 1 HR, against a projected 4 runs / 8 hits (-1 runs vs projection).

Who did the damage to him:

Bo Naylor: 1-for-3, 1 HR, 2 RBI [+1.73 runs vs projection] season xwOBA 0.296, below average - a good day, not a model error
Sal Frelick: 2-for-3, 0 HR, 1 RBI [+0.47 runs vs projection] season xwOBA 0.281, weak - a good day, not a model error
Joey Ortiz: 1-for-4, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.3, below average - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
MIL	Kyle Harrison (Starter)	5 / 15	21.6 / 5.4 (+16.20)	4 / 3	4 / 4	0 / 1	2 / 4	2.1	2 / 5	4.93	venue away: 6.21 vs 1.26 ERA (worse here) · opp run factor 0.95 (average) · 24 prior appearances this season
MIL	Aaron Ashby (Middle Reliever (High Leverage))	3 / 3	18 / 9 (+9.00)	2 / 1	2 / 1	1 / 0	1 / 0	0.48	0 / 0	1.12
PIT	Wilber Dotel (Starter)	6 / 15	4.5 / 1.8	1 / 1	3 / 4	0 / 0	1 / 3	0	2 / 6	2.44	opp run factor 1.01 (strong) · 19 prior appearances this season
PIT	Gregory Soto (Middle Reliever (High Leverage))	2 / 3	0 / 36 (-36.00)	0 / 4	1 / 5	0 / 1	0 / 0	0.59	0 / 1	1.32
PIT	Luke Weaver (Setup Man)	2 / 3	27 / 0 (+27.00)	2 / 0	3 / 0	1 / 0	0 / 2	0.28	0 / 1	1.06
PIT	Mason Montgomery (Closer)	4 / 3	0 / 9 (-9.00)	0 / 1	0 / 1	0 / 1	2 / 0	0.34	2 / 3	1.46

Hitters: actual / projected (sorted by run impact vs projection)

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Bo Naylor (C)	MIL	3 / 3	1 / 0	0 / 0	0 / 0	1 / 0	2 / 0	1 / 0	1 / 0	1 / 1	1.73	0.296 below average	good day
Jackson Chourio (LF)	MIL	4 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	2 / 1	-1.4	0.339 elite	bad day
Rafael Flores Jr. (1B)	PIT	3 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	2 / 1	-1.4	0.344 elite	bad day
Henry Davis (C)	PIT	2 / 3	1 / 1	0 / 0	0 / 0	1 / 0	2 / 0	2 / 0	1 / 0	1 / 1	1.26	0.285 below average	good day
Luis Lara (CF)	MIL	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	1 / 1	-0.78	0.3 below average	projection too high
Brandon Lowe (2B)	PIT	2 / 4	1 / 1	1 / 0	0 / 0	0 / 0	1 / 1	0 / 1	1 / 0	1 / 1	0.64	0.323 above average	projection too low
Sal Frelick (RF)	MIL	3 / 4	2 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 0	0.47	0.281 weak	—
Nick Gonzales (DH)	PIT	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	3 / 1	-0.47	0.327 above average	—
Gary Sánchez (DH)	MIL	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	1 / 1	-0.33	0.341 elite	—
Joey Ortiz (SS)	MIL	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	1 / 0	0 / 1	0.33	0.3 below average	—
Konnor Griffin (SS)	PIT	4 / 4	1 / 1	1 / 1	0 / 0	0 / 0	2 / 1	1 / 1	1 / 0	1 / 1	0.33	0.3 below average	—
Ronny Simon (RF)	PIT	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	1 / 0	0 / 1	0.33	0.317 average	—
Oneil Cruz (CF)	PIT	5 / 4	1 / 1	0 / 0	0 / 0	0 / 0	2 / 1	1 / 1	0 / 1	2 / 2	-0.33	0.355 elite	—
Bryan Reynolds (LF)	PIT	4 / 4	1 / 1	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	-0.33	0.355 elite	—
Brice Turang (2B)	MIL	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	2 / 1	-0.31	0.347 elite	—
Andrew Vaughn (1B)	MIL	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 0	1 / 1	-0.14	0.334 above average	—
David Hamilton (3B)	MIL	2 / 3	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0 / 0	0 / 1	0	0.272 weak	—
Jared Triolo (3B)	PIT	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	0	0.287 below average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-1.17	positive favors home; negative favors away
Recent form	-1.138	positive favors home; negative favors away
Starting-pitcher matchup	-0.061	positive favors home; negative favors away
Composite	-1.423	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	5.14-4.2	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-4	0.56	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	6-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-4	n/a	Capped at max 12 realistic runs
simulation_result	6-4	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	5.14-4.2	5-4	0.061	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	6-4	1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	6-4	6-4	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-4	6-4	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-4	6-4	-0	Capped at max 12 realistic runs
simulation_result	6-4	6-4	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

These are pregame model adjustments, not claims about actual postgame player error.

Team	Player	Role	Net effect	Largest change
PIT	Brandon Lowe	batter / 2B	better	predicted_H 2.69
MIL	Joey Ortiz	batter / SS	better	predicted_AB 1.85
MIL	Sal Frelick	batter / RF	better	predicted_RBI 1.82
MIL	Brice Turang	batter / 2B	better	predicted_AB -1.82
MIL	Bo Naylor	batter / C	worse	predicted_RBI 1.75
PIT	Oneil Cruz	batter / CF	better	predicted_SO -1.5
MIL	Jackson Chourio	batter / LF	worse	predicted_H -1.36
PIT	Jared Triolo	batter / 3B	worse	predicted_SO 1.359

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	MIL pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	MIL
info	CURRENT_PITCHER_DUPLICATION	PIT pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	PIT
info	CURRENT_FINALIZED_MATCH_BOX	MIL model pitcher projections differ from final simulation pitching box score (H: current=9, sim=10; HR: current=1, sim=2).	MIL
info	CURRENT_FINALIZED_MATCH_BOX	MIL model-prediction batting totals differ from final simulation box score (AB: current=32, sim=36; H: current=8, sim=10; 2B: current=2, sim=3; HR: current=1, sim=2; RBI: current=5, sim=6; R: current=4, sim=6; BB: current=4, sim=6; SO: current=8, sim=11).	MIL
info	CURRENT_PLAYER_MATCH_BOX	MIL current finalized hitter lines are model predictions and differ from the simulation box score: Luis Lara: AB current=4 box=5, H current=1 box=2, BB current=1 box=0, SO current=1 box=2; Gary Sánchez: AB current=3 box=4, RBI current=1 box=0, R current=0 box=1, SO current=1 box=2; Joey Ortiz: AB current=4 box=5, H current=1 box=2, R current=0 box=1, SO current=1 box=2; Sal Frelick: HR current=0 box=1, RBI current=0 box=2, R current=0 box=1, SF current=0 box=1; Andrew Vaughn: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0, BB current=0 box=1; Brice Turang: AB current=4 box=3, H current=1 box=2, BB current=1 box=2; Jackson Chourio: AB current=4 box=5, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Bo Naylor: H current=0 box=1, HR current=0 box=1, RBI current=0 box=2, R current=0 box=1	MIL
info	CURRENT_FINALIZED_MATCH_BOX	PIT model-prediction batting totals differ from final simulation box score (AB: current=34, sim=37; H: current=9, sim=10; 2B: current=2, sim=0; HR: current=1, sim=2; RBI: current=5, sim=4; R: current=5, sim=4; BB: current=3, sim=5; SO: current=10, sim=9).	PIT
info	CURRENT_PLAYER_MATCH_BOX	PIT current finalized hitter lines are model predictions and differ from the simulation box score: Rafael Flores Jr.: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Konnor Griffin: AB current=4 box=5, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Nick Gonzales: BB current=0 box=1; Ronny Simon: AB current=4 box=3, H current=1 box=0, BB current=0 box=1, SO current=1 box=0; Jared Triolo: BB current=0 box=1, SO current=1 box=2; Oneil Cruz: AB current=4 box=5, H current=1 box=2, HR current=0 box=1, BB current=1 box=0; Bryan Reynolds: 2B current=1 box=0, RBI current=1 box=0; Brandon Lowe: AB current=4 box=5, H current=1 box=4, HR current=0 box=1, RBI current=1 box=2	PIT
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

MIN@LAAMISS

Predicted: MIN 4-3 (conf -0.03, toss_up)Actual: LAA 4-5Brier 0.267total bias -2

Root cause: Genuine toss-up that flipped

Evidence-backed diagnosis

published system edge 0.03 (toss_up); this is not the half-scale distance from 50%
team-run absolute error 2; total bias -2
pregame home-minus-away evidence: season +0.079, recent +0.407, pitcher -0.034, composite -0.110
confidence pipeline reached 0.523 before later publication/recency controls; published home probability was 0.484
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

No large single-factor errors.

Team	Starter	Projected	Actual	ER err
MIN	Taj Bradley	3 ER / 18 outs	3 ER / 21 outs	0
LAA	Walbert Ureña	1 ER / 17 outs	0 ER / 21 outs	-1

Decision internals: run factors home×0.925 / away×0.959; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

Every value is actual / projected. Run impact approximates how far each hitter's real line moved run expectancy from its projection; it orders the movers and is not a published stat.

Top deviations that drove the game:

Walker Jenkins (MIN CF): 3-for-4, 1 HR, 2 RBI (projected 0.0 H, 0.0 HR) [run impact 2.01]
Zach Neto (LAA SS): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Oswald Peraza (LAA 2B): 2-for-3, 0 HR, 1 RBI (projected 0.0 H, 0.0 HR) [run impact 0.94]
Ben Ross (MIN SS): 1-for-4, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 0.78]
Josh Bell (MIN DH): 1-for-5, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.64]

Pitcher versus batter: what actually broke this game

AWAY pitching

Taj Bradley (MIN) threw 21 outs, 3 ER = 3.86 ERA against a projected 4.5 ERA (-0.64 ERA error)

Bullpen behind him: 1 ER over 4 outs.

The HOME lineup he faced produced 5 runs on 8 hits and 0 HR, against a projected 4 runs / 7 hits (1 runs vs projection).

Who did the damage to him:

Oswald Peraza: 2-for-3, 0 HR, 1 RBI [+0.94 runs vs projection] season xwOBA 0.274, weak - a good day, not a model error
Josh Lowe: 1-for-4, 0 HR, 0 RBI [+0.31 runs vs projection] season xwOBA 0.288, below average - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

HOME pitching

Walbert Ureña (LAA) threw 21 outs, 0 ER = 0 ERA against a projected 1.59 ERA (-1.59 ERA error)

Bullpen behind him: 2 ER over 5 outs.

The AWAY lineup he faced produced 4 runs on 10 hits and 1 HR, against a projected 4 runs / 7 hits (0 runs vs projection).

Who did the damage to him:

Walker Jenkins: 3-for-4, 1 HR, 2 RBI [+2.01 runs vs projection] season xwOBA 0.244, weak - a good day, not a model error
Ben Ross: 1-for-4, 0 HR, 0 RBI [+0.78 runs vs projection] season xwOBA 0.16, weak - a good day, not a model error
Brooks Lee: 2-for-4, 0 HR, 2 RBI [+0.47 runs vs projection] season xwOBA 0.275, weak - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
MIN	Taj Bradley (Starter)	21 / 18	3.86 / 4.5	3 / 3	5 / 5	0 / 1	1 / 1	2.46	9 / 6	7.21	venue away: 4.52 vs 3.05 ERA (worse here) · opp run factor 0.92 (weak) · 29 prior appearances this season
MIN	A.J. Minter (Middle Reliever (High Leverage))	2 / 3	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0.22	2 / 1	1.01
MIN	Jeff Hoffman (Setup Man)	1 / 3	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0.35	1 / 0	1.52
MIN	Yoendrys Gómez (Closer)	1 / 3	27 / 0 (+27.00)	1 / 0	1 / 0	0 / 0	1 / 0	0.57	0 / 2	1.04
LAA	Walbert Ureña (Starter)	21 / 17	0 / 1.59	0 / 1	4 / 3	0 / 0	0 / 2	2.07	9 / 5	6.32	venue home: 2.42 vs 3.59 ERA (ok) · opp run factor 0.96 (average) · 27 prior appearances this season
LAA	Ben Joyce (Closer)	3 / 3	18 / 18	2 / 2	2 / 4	1 / 0	0 / 0	0.66	3 / 1	1.17
LAA	Tayler Saucedo (Middle Reliever (High Leverage))	2 / 3	0 / 9 (-9.00)	0 / 1	0 / 1	0 / 1	1 / 0	0.14	1 / 0	0.85

Hitters: actual / projected (sorted by run impact vs projection)

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Walker Jenkins (CF)	MIN	4 / 3	3 / 0	0 / 0	0 / 0	1 / 0	2 / 0	2 / 0	0 / 1	1 / 1	2.01	0.244 weak	good day
Zach Neto (SS)	LAA	4 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	1 / 1	3 / 2	-1.4	0.311 average	no signal
Oswald Peraza (2B)	LAA	3 / 3	2 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 1	0.94	0.274 weak	good day
Ben Ross (SS)	MIN	4 / 2	1 / 0	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	2 / 1	0.78	0.16 weak	good day
Josh Bell (DH)	MIN	5 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	-0.64	0.333 above average	bad day
Kody Clemens (2B)	MIN	3 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 0	1 / 1	-0.6	0.336 above average	bad day
Emmanuel Rodriguez (LF)	MIN	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	3 / 1	-0.47	0.48 elite	—
Brooks Lee (3B)	MIN	4 / 4	2 / 1	0 / 0	0 / 0	0 / 0	2 / 1	0 / 1	0 / 0	0 / 1	0.47	0.275 weak	—
Royce Lewis (1B)	MIN	5 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	4 / 1	-0.47	0.311 average	—
Jose Siri (CF)	LAA	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	0 / 0	3 / 2	-0.47	0.305 average	—
Luke Keaschall (RF)	MIN	5 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	2 / 0	-0.33	0.319 average	—
Josh Lowe (RF)	LAA	4 / 3	1 / 1	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0.31	0.288 below average	—
Wade Meckler (LF)	LAA	5 / 4	2 / 1	0 / 1	0 / 0	0 / 0	1 / 0	1 / 1	0 / 1	1 / 1	-0.17	0.313 average	—
Victor Caratini (C)	MIN	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	0	0.315 average	—
Mike Trout (DH)	LAA	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	1 / 1	1 / 1	0	0.378 elite	—
Moisés Ballesteros (1B)	LAA	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	0 / 0	1 / 1	0	0.3 below average	—
Denzer Guzman (3B)	LAA	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0 / 0	1 / 1	0	0.264 weak	—
Tyler Heineman (C)	LAA	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0	0.24 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.079	positive favors home; negative favors away
Recent form	0.407	positive favors home; negative favors away
Starting-pitcher matchup	-0.034	positive favors home; negative favors away
Composite	-0.11	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.52-3.89	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-3	0.516	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-3	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-3	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-3	n/a	Capped at max 12 realistic runs
simulation_result	4-3	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.52-3.89	5-4	0.369	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	4-3	-0	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-3	4-3	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-3	4-3	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-3	4-3	-0	Capped at max 12 realistic runs
simulation_result	4-3	4-3	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

These are pregame model adjustments, not claims about actual postgame player error.

Team	Player	Role	Net effect	Largest change
MIN	Kody Clemens	batter / 2B	worse	predicted_AB -2.58
LAA	Jose Siri	batter / CF	better	predicted_H 2.425
LAA	Tyler Heineman	batter / C	worse	predicted_SO 1.565
MIN	Luke Keaschall	batter / RF	better	predicted_AB 1.521
MIN	Ben Ross	batter / SS	worse	predicted_AB 1.5
LAA	Mike Trout	batter / DH	worse	predicted_SO -1.165
MIN	Royce Lewis	batter / 1B	worse	predicted_H 1.14
LAA	Josh Lowe	batter / RF	worse	predicted_SO 1.061

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	LAA pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	LAA
info	CURRENT_PITCHER_DUPLICATION	MIN pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	MIN
info	CURRENT_FINALIZED_MATCH_BOX	LAA model-prediction batting totals differ from final simulation box score (AB: current=32, sim=33; H: current=7, sim=6; R: current=4, sim=3; BB: current=3, sim=1; SO: current=11, sim=9; HBP: current=0, sim=2).	LAA
info	CURRENT_PLAYER_MATCH_BOX	LAA current finalized hitter lines are model predictions and differ from the simulation box score: Mike Trout: H current=1 box=2, BB current=1 box=0, SO current=1 box=0; Wade Meckler: H current=1 box=0, 2B current=1 box=0, R current=1 box=0, BB current=1 box=0; Josh Lowe: H current=1 box=0, SO current=1 box=2, HBP current=0 box=1; Moisés Ballesteros: AB current=4 box=3, H current=1 box=0, BB current=0 box=1; Jose Siri: AB current=3 box=4, H current=1 box=3, HR current=0 box=1, RBI current=1 box=2; Denzer Guzman: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, SO current=1 box=2; Tyler Heineman: AB current=3 box=4, SO current=1 box=2; Zach Neto: 2B current=0 box=1, HR current=1 box=0, BB current=1 box=0, SO current=2 box=1	LAA
info	CURRENT_FINALIZED_MATCH_BOX	MIN model pitcher projections differ from final simulation pitching box score (H: current=5, sim=6).	MIN
info	CURRENT_FINALIZED_MATCH_BOX	MIN model-prediction batting totals differ from final simulation box score (AB: current=32, sim=35; H: current=7, sim=8; 2B: current=1, sim=2; BB: current=3, sim=2; SO: current=8, sim=6).	MIN
info	CURRENT_PLAYER_MATCH_BOX	MIN current finalized hitter lines are model predictions and differ from the simulation box score: Luke Keaschall: AB current=4 box=5, H current=1 box=2, 2B current=0 box=1, RBI current=0 box=1; Emmanuel Rodriguez: H current=1 box=0, RBI current=1 box=0, SO current=1 box=0; Kody Clemens: AB current=4 box=2, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Brooks Lee: 2B current=0 box=1, SO current=1 box=0; Josh Bell: 2B current=1 box=0, BB current=1 box=0, SO current=1 box=0; Royce Lewis: H current=1 box=2, SO current=1 box=0; Victor Caratini: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=1, R current=0 box=1; Walker Jenkins: AB current=3 box=4, BB current=1 box=0, SO current=1 box=2	MIN
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

OAK@TBHIT

Predicted: TB 4-7 (conf +0.16, moderate)Actual: TB 1-10Brier 0.176total bias +0

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.16 (moderate); this is not the half-scale distance from 50%
team-run absolute error 6; total bias +0
pregame home-minus-away evidence: season +1.683, recent +1.202, pitcher +0.051, composite +1.880
confidence pipeline reached 0.650 before later publication/recency controls; published home probability was 0.580
score-volatility actions: supported_blowout_margin_expansion
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

OAK offense over-projected by 3
TB offense under-projected by 3

Team	Starter	Projected	Actual	ER err
OAK	Jeffrey Springs	4 ER / 14 outs	6 ER / 10 outs	2
TB	Drew Rasmussen	2 ER / 18 outs	0 ER / 15 outs	-2

Decision internals: run factors home×1.079 / away×1.027; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

Every value is actual / projected. Run impact approximates how far each hitter's real line moved run expectancy from its projection; it orders the movers and is not a published stat.

Top deviations that drove the game:

Lawrence Butler (OAK RF): 2-for-4, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.07]
Junior Caminero (TB 3B): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.07]
Richie Palacios (TB 2B): 1-for-2, 0 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 0.97]
Zack Gelof (OAK 3B): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.93]
Denzel Clarke (OAK CF): 1-for-2, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 0.8]

Starter miss: Jeffrey Springs (OAK) posted 16.20 ERA vs projected 7.71 (6 ER in 10 outs); he faced a strong offense (opponent run factor 1.08, bucket average).

What to do: Consider raising context.opponent_strength_weight; validate across the rolling window before promotion.

Pitcher versus batter: what actually broke this game

AWAY pitching

Jeffrey Springs (OAK) threw 10 outs, 6 ER = 16.2 ERA against a projected 7.71 ERA (8.49 ERA error)

Bullpen behind him: 0 ER over 3 outs.

The HOME lineup he faced produced 9 runs on 7 hits and 0 HR, against a projected 5 runs / 9 hits (4 runs vs projection).

Who did the damage to him:

Richie Palacios: 1-for-2, 0 HR, 2 RBI [+0.97 runs vs projection] season xwOBA 0.309, average - a good day, not a model error
Liam Hicks: 2-for-4, 0 HR, 2 RBI [+0.8 runs vs projection] season xwOBA 0.316, average - a good day, not a model error
Chandler Simpson: 2-for-5, 0 HR, 2 RBI [+0.47 runs vs projection] season xwOBA 0.265, weak - a good day, not a model error

Verdict: both: the starter was hit hard AND the lineup underproduced

HOME pitching

Drew Rasmussen (TB) threw 15 outs, 0 ER = 0 ERA against a projected 3 ERA (-3 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The AWAY lineup he faced produced 1 runs on 5 hits and 1 HR, against a projected 3 runs / 8 hits (-2 runs vs projection).

Who did the damage to him:

Lawrence Butler: 2-for-4, 1 HR, 1 RBI [+1.07 runs vs projection] season xwOBA 0.321, above average - real projection error
Denzel Clarke: 1-for-2, 0 HR, 0 RBI [+0.8 runs vs projection] season xwOBA 0.202, weak - a good day, not a model error
Henry Bolte: 1-for-3, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.318, average - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
OAK	Jeffrey Springs (Starter)	10 / 14	16.2 / 7.71 (+8.49)	6 / 4	5 / 6	0 / 1	2 / 3	1.85	4 / 6	3.59	venue away: 4.76 vs 6.94 ERA (ok) · opp run factor 1.08 (average) · 27 prior appearances this season
OAK	Chris Roycroft (Reliever)	3 / 3	0 / 9 (-9.00)	0 / 1	0 / 2	0 / 1	2 / 0	n/a	2 / 0	n/a
TB	Drew Rasmussen (Starter)	15 / 18	0 / 3 (-3.00)	0 / 2	4 / 3	0 / 1	2 / 1	0.82	3 / 8	4.62	venue home: 3.05 vs 2.57 ERA (ok) · opp run factor 1.03 (weak) · 28 prior appearances this season

Hitters: actual / projected (sorted by run impact vs projection)

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Lawrence Butler (RF)	OAK	4 / 4	2 / 1	0 / 0	0 / 0	1 / 0	1 / 1	1 / 1	0 / 1	1 / 1	1.07	0.321 above average	projection too low
Junior Caminero (3B)	TB	3 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	2 / 1	1 / 1	-1.07	0.363 elite	bad day
Richie Palacios (2B)	TB	2 / 3	1 / 1	1 / 0	0 / 0	0 / 0	2 / 0	1 / 0	2 / 0	0 / 1	0.97	0.309 average	no signal
Zack Gelof (3B)	OAK	4 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	0 / 1	-0.93	0.311 average	no signal
Denzel Clarke (CF)	OAK	2 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	0.8	0.202 weak	good day
Donovan Walton (2B)	OAK	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	-0.8	0.332 above average	bad day
Liam Hicks (C)	TB	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	2 / 1	1 / 0	1 / 0	1 / 0	0.8	0.316 average	no signal
Yandy Díaz (DH)	TB	5 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	2 / 0	-0.8	0.35 elite	bad day
Jeff McNeil (DH)	OAK	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 0	-0.78	0.303 below average	projection too high
Jonny DeLuca (CF)	TB	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	-0.78	0.266 weak	projection too high
Tommy White (1B)	OAK	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.296 below average	—
Jonah Heim (C)	OAK	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 1	-0.47	0.287 below average	—
Chandler Simpson (LF)	TB	5 / 4	2 / 1	0 / 0	0 / 0	0 / 0	2 / 0	2 / 1	0 / 0	0 / 0	0.47	0.265 weak	—
Alika Williams (SS)	OAK	3 / 3	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	-0.45	0.243 weak	—
Henry Bolte (LF)	OAK	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	1 / 1	0.33	0.318 average	—
Taylor Walls (SS)	TB	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	2 / 0	1 / 1	1 / 0	1 / 1	0.33	0.249 weak	—
Jonathan Aranda (1B)	TB	4 / 4	1 / 1	1 / 0	0 / 0	0 / 0	1 / 1	1 / 1	1 / 1	2 / 1	0.31	0.347 elite	—
Ryan Vilade (RF)	TB	1 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	1 / 1	-0.14	0.331 above average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	1.683	positive favors home; negative favors away
Recent form	1.202	positive favors home; negative favors away
Starting-pitcher matchup	0.051	positive favors home; negative favors away
Composite	1.88	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.02-5.69	n/a	Base ML historical model prediction
variance_applied	4-6	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-7	0.58	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	4-7	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-7	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-7	n/a	Capped at max 12 realistic runs
simulation_result	4-7	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.02-5.69	4-6	0.336	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-6	4-7	1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	4-7	4-7	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-7	4-7	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-7	4-7	0	Capped at max 12 realistic runs
simulation_result	4-7	4-7	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

These are pregame model adjustments, not claims about actual postgame player error.

Team	Player	Role	Net effect	Largest change
OAK	Jeffrey Springs	pitcher / Starter	worse	predicted_IP -3.8
OAK	Chris Roycroft	pitcher / Reliever	worse	predicted_H 2
OAK	Tommy White	batter / 1B	boosted	predicted_RBI 1.8
TB	Taylor Walls	batter / SS	better	predicted_H 1.42
TB	Yandy Díaz	batter / DH	worse	predicted_RBI 1.41
TB	Liam Hicks	batter / C	worse	predicted_AB -1.366
TB	Ryan Vilade	batter / RF	worse	predicted_AB 1.25
OAK	Alika Williams	batter / SS	worse	predicted_AB 1.159

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	OAK pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	OAK
info	CURRENT_PITCHER_DUPLICATION	TB pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	TB
info	CURRENT_FINALIZED_MATCH_BOX	OAK model pitcher projections differ from final simulation pitching box score (ER: current=6, sim=7).	OAK
info	CURRENT_FINALIZED_MATCH_BOX	OAK model-prediction batting totals differ from final simulation box score (AB: current=32, sim=31; H: current=8, sim=5; 2B: current=2, sim=1; HR: current=1, sim=2; R: current=3, sim=4; BB: current=2, sim=3; SO: current=8, sim=10; SF: current=0, sim=1).	OAK
info	CURRENT_PLAYER_MATCH_BOX	OAK current finalized hitter lines are model predictions and differ from the simulation box score: Jeff McNeil: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, SO current=0 box=1; Henry Bolte: AB current=4 box=3, H current=1 box=0, R current=1 box=0, BB current=0 box=1; Tommy White: HR current=0 box=1, RBI current=0 box=2, R current=0 box=1, SO current=1 box=2; Alika Williams: AB current=3 box=4, 2B current=1 box=0; Zack Gelof: H current=1 box=2, 2B current=0 box=1, HR current=1 box=0, RBI current=1 box=0; Jonah Heim: AB current=3 box=4, HR current=0 box=1, R current=0 box=1, SO current=1 box=0; Donovan Walton: AB current=3 box=2, H current=1 box=0, RBI current=0 box=1, R current=0 box=1; Lawrence Butler: AB current=4 box=3, H current=1 box=0, RBI current=1 box=0, R current=1 box=0	OAK
info	CURRENT_FINALIZED_MATCH_BOX	TB model-prediction batting totals differ from final simulation box score (AB: current=32, sim=35; H: current=9, sim=10; HR: current=1, sim=2; RBI: current=4, sim=7; R: current=5, sim=7; SF: current=0, sim=1; HBP: current=0, sim=1).	TB
info	CURRENT_PLAYER_MATCH_BOX	TB current finalized hitter lines are model predictions and differ from the simulation box score: Liam Hicks: AB current=3 box=2, H current=1 box=0, R current=0 box=1, BB current=0 box=1; Jonny DeLuca: RBI current=0 box=1, R current=0 box=1, SO current=1 box=0; Jonathan Aranda: AB current=4 box=5, HR current=0 box=1, BB current=1 box=0; Junior Caminero: H current=1 box=2, HR current=1 box=0, RBI current=1 box=0; Yandy Díaz: RBI current=1 box=2, SO current=0 box=2; Ryan Vilade: AB current=3 box=4, H current=1 box=0; Chandler Simpson: H current=1 box=2, RBI current=0 box=1; Richie Palacios: AB current=3 box=4	TB
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

PHI@NYMMISS

Predicted: NYM 3-4 (conf +0.09, lean)Actual: PHI 3-0Brier 0.296total bias +4

Root cause: Starting-pitcher projection was wrong

Evidence-backed diagnosis

published system edge 0.09 (lean); this is not the half-scale distance from 50%
team-run absolute error 4; total bias +4
pregame home-minus-away evidence: season -0.312, recent +0.358, pitcher +0.068, composite +0.246
confidence pipeline reached 0.561 before later publication/recency controls; published home probability was 0.544
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 4 (7 vs 3)
NYM offense over-projected by 4
Aaron Nola (PHI) allowed 4 fewer ER than projected (0 vs 4)

Team	Starter	Projected	Actual	ER err
PHI	Aaron Nola	4 ER / 18 outs	0 ER / 21 outs	-4
NYM	Nolan McLean	2 ER / 18 outs	1 ER / 18 outs	-1

Decision internals: run factors home×0.992 / away×0.925; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

Every value is actual / projected. Run impact approximates how far each hitter's real line moved run expectancy from its projection; it orders the movers and is not a published stat.

Top deviations that drove the game:

Kyle Schwarber (PHI DH): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
A.J. Ewing (NYM CF): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -1.11]
Juan Soto (NYM DH): 0-for-2, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.07]
Bryce Harper (PHI RF): 3-for-4, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.94]
Brett Baty (NYM 2B): 0-for-1, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 0.66]

Starter miss: Aaron Nola (PHI) posted 0.00 ERA vs projected 6.00 (0 ER in 21 outs); no credible venue, opponent-strength or experience signal explains it — an outlier start (variance).

What to do: Variance — no model change is warranted for this start.

Pitcher versus batter: what actually broke this game

AWAY pitching

Aaron Nola (PHI) threw 21 outs, 0 ER = 0 ERA against a projected 6 ERA (-6 ERA error)

Bullpen behind him: 0 ER over 5 outs.

The HOME lineup he faced produced 0 runs on 3 hits and 0 HR, against a projected 4 runs / 8 hits (-4 runs vs projection).

Who did the damage to him:

Brett Baty: 0-for-1, 0 HR, 0 RBI [+0.66 runs vs projection] season xwOBA 0.31, average - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

HOME pitching

Nolan McLean (NYM) threw 18 outs, 1 ER = 1.5 ERA against a projected 3 ERA (-1.5 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The AWAY lineup he faced produced 3 runs on 8 hits and 0 HR, against a projected 4 runs / 7 hits (-1 runs vs projection).

Who did the damage to him:

Bryce Harper: 3-for-4, 0 HR, 1 RBI [+0.94 runs vs projection] season xwOBA 0.385, elite - real projection error
Justin Crawford: 1-for-2, 0 HR, 1 RBI [+0.47 runs vs projection] season xwOBA 0.289, below average - a good day, not a model error
Bryson Stott: 2-for-3, 0 HR, 0 RBI [+0.45 runs vs projection] season xwOBA 0.327, above average - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
PHI	Aaron Nola (Starter)	21 / 18	0 / 6 (-6.00)	0 / 4	3 / 6	0 / 1	3 / 1	1.86	4 / 3	4.88	venue away: 4.15 vs 5.3 ERA (ok) · opp run factor 0.99 (average) · 30 prior appearances this season
PHI	Jhoan Duran (Closer)	3 / 3	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0.13	2 / 0	1.27
PHI	Jonathan Bowlan (Middle Reliever (High Leverage))	2 / 3	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	0.2	2 / 0	1.17
NYM	Nolan McLean (Starter)	18 / 18	1.5 / 3	1 / 2	7 / 4	0 / 0	3 / 3	3.04	5 / 6	5.54	venue home: 3.57 vs 2.89 ERA (worse here) · opp run factor 0.92 (average) · 29 prior appearances this season

Hitters: actual / projected (sorted by run impact vs projection)

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Kyle Schwarber (DH)	PHI	4 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	1 / 1	-1.4	0.365 elite	bad day
A.J. Ewing (CF)	NYM	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	2 / 1	-1.11	0.322 above average	bad day
Juan Soto (DH)	NYM	2 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	2 / 1	0 / 1	-1.07	0.412 elite	bad day
Bryce Harper (RF)	PHI	4 / 4	3 / 1	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	1 / 1	1 / 1	0.94	0.385 elite	projection too low
Brett Baty (2B)	NYM	1 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	1 / 1	0.66	0.31 average	no signal
Justin Crawford (CF)	PHI	2 / 3	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 1	0.47	0.289 below average	—
Francisco Lindor (SS)	NYM	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0 / 1	-0.47	0.329 above average	—
Francisco Alvarez (C)	NYM	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.322 above average	—
Jared Young (1B)	NYM	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	0 / 1	-0.47	0.322 above average	—
Bryson Stott (3B)	PHI	3 / 4	2 / 1	1 / 0	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	0 / 1	0.45	0.327 above average	—
Luis Arraez (2B)	PHI	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 0	1 / 0	-0.14	0.306 average	—
Trea Turner (SS)	PHI	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	1 / 1	2 / 0	0 / 1	-0.12	0.291 below average	—
Alec Bohm (1B)	PHI	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 1	0	0.328 above average	—
Brandon Marsh (LF)	PHI	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	0	0.302 below average	—
Garrett Stubbs (C)	PHI	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0	0.212 weak	—
Carson Benge (RF)	NYM	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	0	0.338 above average	—
Nick Morabito (LF)	NYM	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.324 above average	—
Bo Bichette (3B)	NYM	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	0	0.326 above average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.312	positive favors home; negative favors away
Recent form	0.358	positive favors home; negative favors away
Starting-pitcher matchup	0.068	positive favors home; negative favors away
Composite	0.246	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.27-4.72	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	3-4	0.544	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	3-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-4	n/a	Capped at max 12 realistic runs
simulation_result	3-4	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.27-4.72	4-5	0.554	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	3-4	0	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	3-4	3-4	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-4	3-4	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-4	3-4	0	Capped at max 12 realistic runs
simulation_result	3-4	3-4	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

These are pregame model adjustments, not claims about actual postgame player error.

Team	Player	Role	Net effect	Largest change
PHI	Trea Turner	batter / SS	worse	predicted_AB -2.94
NYM	Jared Young	batter / 1B	better	predicted_2B 1.823
PHI	Kyle Schwarber	batter / DH	worse	predicted_AB -1.69
NYM	Francisco Alvarez	batter / C	better	predicted_RBI 1.59
PHI	Bryson Stott	batter / 3B	worse	predicted_RBI 1.562
PHI	Garrett Stubbs	batter / C	worse	predicted_SO 1.368
PHI	Luis Arraez	batter / 2B	worse	predicted_H -1.31
PHI	Brandon Marsh	batter / LF	worse	predicted_SO 1.085

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	NYM pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	NYM
info	CURRENT_PITCHER_DUPLICATION	PHI pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	PHI
info	CURRENT_FINALIZED_MATCH_BOX	NYM model-prediction batting totals differ from final simulation box score (AB: current=34, sim=32; 2B: current=1, sim=4; SO: current=9, sim=3).	NYM
info	CURRENT_PLAYER_MATCH_BOX	NYM current finalized hitter lines are model predictions and differ from the simulation box score: Juan Soto: AB current=4 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Carson Benge: H current=1 box=0, R current=1 box=0, SO current=1 box=0; A.J. Ewing: AB current=4 box=3, H current=1 box=2, 2B current=1 box=0, R current=0 box=1; Francisco Lindor: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Brett Baty: AB current=3 box=4, H current=0 box=1, 2B current=0 box=1, RBI current=0 box=1; Nick Morabito: AB current=3 box=2, 2B current=0 box=1, RBI current=0 box=1, R current=0 box=1; Francisco Alvarez: HR current=0 box=1, RBI current=0 box=2, R current=0 box=1, SO current=1 box=0; Jared Young: H current=1 box=2, 2B current=0 box=2, RBI current=1 box=0, R current=1 box=0	NYM
info	CURRENT_FINALIZED_MATCH_BOX	PHI model pitcher projections differ from final simulation pitching box score (H: current=7, sim=8; BB: current=2, sim=3).	PHI
info	CURRENT_FINALIZED_MATCH_BOX	PHI model-prediction batting totals differ from final simulation box score (AB: current=33, sim=31; H: current=7, sim=4; 3B: current=0, sim=1; HR: current=1, sim=0; RBI: current=4, sim=3; R: current=4, sim=3; BB: current=3, sim=5; SO: current=8, sim=9).	PHI
info	CURRENT_PLAYER_MATCH_BOX	PHI current finalized hitter lines are model predictions and differ from the simulation box score: Bryce Harper: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Alec Bohm: H current=1 box=2, R current=0 box=1; Bryson Stott: 3B current=0 box=1, RBI current=1 box=2, BB current=1 box=0, SO current=1 box=2; Luis Arraez: AB current=4 box=3, H current=1 box=0, BB current=0 box=1, SO current=0 box=1; Justin Crawford: BB current=0 box=1; Brandon Marsh: AB current=3 box=4, H current=1 box=0, SO current=1 box=2; Trea Turner: AB current=4 box=2, H current=1 box=0, 2B current=1 box=0, R current=1 box=0; Kyle Schwarber: AB current=4 box=3, 2B current=0 box=1, HR current=1 box=0, RBI current=1 box=0	PHI
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

SD@COLHIT

Predicted: SD 8-5 (conf -0.08, lean)Actual: SD 9-2Brier 0.212total bias +2

Root cause: Solid correct call

Evidence-backed diagnosis

published system edge 0.08 (lean); this is not the half-scale distance from 50%
team-run absolute error 4; total bias +2
pregame home-minus-away evidence: season -1.504, recent -1.654, pitcher -0.064, composite -1.922
confidence pipeline reached 0.697 before later publication/recency controls; published home probability was 0.460
score-volatility actions: supported_blowout_margin_expansion
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

COL offense over-projected by 3

Team	Starter	Projected	Actual	ER err
SD	Michael King	3 ER / 19 outs	2 ER / 18 outs	-1
COL	Tanner Gordon	6 ER / 15 outs	7 ER / 9 outs	1

Decision internals: run factors home×1.006 / away×1.07; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

Every value is actual / projected. Run impact approximates how far each hitter's real line moved run expectancy from its projection; it orders the movers and is not a published stat.

Top deviations that drove the game:

Xander Bogaerts (SD SS): 3-for-5, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 2.47]
Manny Machado (SD 3B): 0-for-3, 0 HR, 1 RBI (projected 1.0 H, 1.0 HR) [run impact -1.73]
Samad Taylor (SD LF): 2-for-4, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.11]
Jake Cronenworth (SD 2B): 2-for-5, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 1.09]
Dustin Harris (SD CF): 3-for-5, 0 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 0.94]

Starter miss: Tanner Gordon (COL) posted 21.00 ERA vs projected 10.80 (7 ER in 9 outs); he is materially worse pitching home (8.04 ERA there vs 3.00 on the other side), a venue-role weakness the model under-weighted.

What to do: Enable/raise the venue-split weight (context.venue_adjustment_weight) and validate in shadow; the split is credible for this starter.

Pitcher versus batter: what actually broke this game

AWAY pitching

Michael King (SD) threw 18 outs, 2 ER = 3 ERA against a projected 4.26 ERA (-1.26 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 2 runs on 9 hits and 0 HR, against a projected 4 runs / 9 hits (-2 runs vs projection).

Who did the damage to him:

Brett Sullivan: 2-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.283, weak - a good day, not a model error
Jake McCarthy: 1-for-3, 0 HR, 0 RBI [+0.31 runs vs projection] season xwOBA 0.326, above average - a good day, not a model error

Verdict: neither side broke materially - the result is closer to variance

HOME pitching

Tanner Gordon (COL) threw 9 outs, 7 ER = 21 ERA against a projected 10.8 ERA (10.2 ERA error)

Bullpen behind him: 0 ER over 9 outs.

The AWAY lineup he faced produced 9 runs on 15 hits and 1 HR, against a projected 5 runs / 10 hits (4 runs vs projection).

Who did the damage to him:

Xander Bogaerts: 3-for-5, 1 HR, 2 RBI [+2.47 runs vs projection] season xwOBA 0.32, average - a good day, not a model error
Samad Taylor: 2-for-4, 0 HR, 1 RBI [+1.11 runs vs projection] season xwOBA 0.307, average - a good day, not a model error
Jake Cronenworth: 2-for-5, 0 HR, 0 RBI [+1.09 runs vs projection] season xwOBA 0.269, weak - a good day, not a model error

Verdict: both: the starter was hit hard AND the lineup underproduced

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
SD	Michael King (Starter)	18 / 19	3 / 4.26	2 / 3	7 / 6	0 / 0	2 / 1	2.12	6 / 6	4.75	venue away: 3.34 vs 2.74 ERA (worse here) · opp run factor 1.01 (average) · 30 prior appearances this season
COL	Tanner Gordon (Starter)	9 / 15	21 / 10.8 (+10.20)	7 / 6	10 / 5	1 / 2	2 / 1	0.99	3 / 5	3.52	venue home: 8.04 vs 3 ERA (worse here) · opp run factor 1.07 (average) · 22 prior appearances this season
COL	Brennan Bernardino (Setup Man)	3 / 3	0 / 9 (-9.00)	0 / 1	0 / 1	0 / 0	0 / 2	0.17	2 / 2	0.65
COL	Jaden Hill (Middle Reliever (High Leverage))	3 / 3	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0.62	0 / 2	0.99
COL	Jordan Romano (Closer)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0.54	2 / 2	1.06

Hitters: actual / projected (sorted by run impact vs projection)

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Xander Bogaerts (SS)	SD	5 / 3	3 / 1	1 / 0	1 / 0	1 / 0	2 / 0	1 / 1	0 / 1	2 / 0	2.47	0.32 average	no signal
Manny Machado (3B)	SD	3 / 4	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	0 / 1	0 / 1	-1.73	0.334 above average	bad day
Samad Taylor (LF)	SD	4 / 3	2 / 1	1 / 0	0 / 0	0 / 0	1 / 0	2 / 1	1 / 0	1 / 1	1.11	0.307 average	no signal
Jake Cronenworth (2B)	SD	5 / 4	2 / 1	0 / 0	1 / 0	0 / 0	0 / 0	2 / 0	0 / 0	0 / 1	1.09	0.269 weak	good day
Dustin Harris (CF)	SD	5 / 4	3 / 1	0 / 0	0 / 0	0 / 0	3 / 1	1 / 0	0 / 0	1 / 0	0.94	0.336 above average	projection too low
Ty France (1B)	SD	4 / 4	3 / 1	1 / 1	0 / 0	0 / 0	0 / 1	1 / 0	0 / 0	0 / 1	0.94	0.333 above average	projection too low
Jackson Merrill (DH)	SD	4 / 4	1 / 1	0 / 0	1 / 0	0 / 0	2 / 1	1 / 1	0 / 0	1 / 1	0.62	0.349 elite	projection too low
Luis Campusano (C)	SD	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.333 above average	—
Troy Johnston (DH)	COL	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.325 above average	—
Zac Veen (RF)	COL	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	3 / 1	-0.47	0.297 below average	—
Brett Sullivan (C)	COL	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 0	0.47	0.283 weak	—
Connor Norby (2B)	COL	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 1	1 / 1	-0.33	0.307 average	—
Kyle Karros (3B)	COL	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0 / 1	1 / 1	-0.33	0.342 elite	—
Jake McCarthy (CF)	COL	3 / 4	1 / 1	0 / 1	1 / 0	0 / 0	0 / 1	1 / 1	0 / 0	0 / 1	0.31	0.326 above average	—
Cole Carrigg (SS)	COL	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 0	1 / 1	-0.31	0.313 average	—
Fernando Tatis Jr. (RF)	SD	4 / 4	1 / 2	1 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 0	1 / 1	-0.14	0.376 elite	—
Mickey Moniak (LF)	COL	3 / 4	2 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 0	0 / 1	-0.13	0.28 weak	—
TJ Rumfield (1B)	COL	2 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	1 / 0	0	0.336 above average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-1.504	positive favors home; negative favors away
Recent form	-1.654	positive favors home; negative favors away
Starting-pitcher matchup	-0.064	positive favors home; negative favors away
Composite	-1.922	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	5.53-3.88	n/a	Base ML historical model prediction
variance_applied	6-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	8-5	0.54	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	8-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	8-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	8-5	n/a	Capped at max 12 realistic runs
simulation_result	8-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	5.53-3.88	6-4	0.351	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-4	8-5	1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	8-5	8-5	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	8-5	8-5	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	8-5	8-5	-0	Capped at max 12 realistic runs
simulation_result	8-5	8-5	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

These are pregame model adjustments, not claims about actual postgame player error.

Team	Player	Role	Net effect	Largest change
SD	Jackson Merrill	batter / DH	better	predicted_RBI 4.439
COL	Zac Veen	batter / RF	better	predicted_RBI 3.423
SD	Dustin Harris	batter / CF	better	predicted_SO 2.567
SD	Fernando Tatis Jr.	batter / RF	better	predicted_SO 2.176
SD	Xander Bogaerts	batter / SS	worse	predicted_SO 1.29
COL	Kyle Karros	batter / 3B	worse	predicted_SO 1.271
COL	Mickey Moniak	batter / LF	worse	predicted_H 1.16
COL	Jake McCarthy	batter / CF	worse	predicted_AB 1.122

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	COL pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	COL
info	CURRENT_PITCHER_DUPLICATION	SD pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	SD
info	CURRENT_FINALIZED_MATCH_BOX	COL model pitcher projections differ from final simulation pitching box score (H: current=8, sim=11; ER: current=7, sim=8; HR: current=2, sim=3).	COL
info	CURRENT_FINALIZED_MATCH_BOX	COL model-prediction batting totals differ from final simulation box score (AB: current=33, sim=36; 2B: current=2, sim=3; R: current=4, sim=5; BB: current=3, sim=1).	COL
info	CURRENT_PLAYER_MATCH_BOX	COL current finalized hitter lines are model predictions and differ from the simulation box score: Jake McCarthy: AB current=4 box=5, 2B current=1 box=0; Cole Carrigg: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Connor Norby: R current=0 box=1, BB current=1 box=0; Troy Johnston: AB current=3 box=4, 2B current=0 box=1, SO current=1 box=0; Kyle Karros: H current=1 box=0, R current=1 box=0, BB current=1 box=0, SO current=1 box=2; TJ Rumfield: 2B current=0 box=1, RBI current=1 box=0, BB current=1 box=0; Mickey Moniak: H current=1 box=2, HR current=1 box=0, RBI current=1 box=0; Zac Veen: AB current=3 box=4, H current=1 box=2, 2B current=0 box=1, HR current=0 box=1	COL
info	CURRENT_FINALIZED_MATCH_BOX	SD model-prediction batting totals differ from final simulation box score (AB: current=33, sim=38; H: current=10, sim=11; 2B: current=2, sim=1; HR: current=1, sim=3; RBI: current=5, sim=8; R: current=5, sim=8; BB: current=2, sim=3; SO: current=7, sim=12).	SD
info	CURRENT_PLAYER_MATCH_BOX	SD current finalized hitter lines are model predictions and differ from the simulation box score: Dustin Harris: AB current=4 box=5, 2B current=0 box=1, RBI current=1 box=0, R current=0 box=1; Jackson Merrill: H current=1 box=2, HR current=0 box=2, RBI current=1 box=5, BB current=0 box=1; Xander Bogaerts: AB current=3 box=4, H current=1 box=0, BB current=1 box=0, SO current=0 box=2; Fernando Tatis Jr.: AB current=4 box=5, 2B current=1 box=0, HR current=0 box=1, RBI current=1 box=2; Manny Machado: AB current=4 box=5, H current=1 box=2, HR current=1 box=0, RBI current=1 box=0; Ty France: H current=1 box=2, 2B current=1 box=0, R current=0 box=1, BB current=0 box=1; Luis Campusano: R current=0 box=1, BB current=0 box=1; Jake Cronenworth: R current=0 box=1	SD
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

Cross-game lever tension

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

Day over day: did it improve or worsen

Compared against packet 20260916. Last recorded change: v4.30.0 (2026-09-08) Hitter shrinkage is derived from the between-hitter spread rather than fitted to slate error, so an established extreme profile is no longer flattened toward the league.

The adjustment the system made in response

Luna deep post-mortem (full)

1. Evidence receipt and gaps

Sources received

All seven source_manifest entries were checked against the supplied evidence:

Source	Status	Purpose
`prediction_review_20260917.json` — 229,581 bytes	Loaded	Official outcomes and scoring errors
`PREDICTION_SYSTEM_DEEP_AUDIT_20260917.json` — 150,866 bytes	Loaded	Probability, provenance, starter, and decision contracts
`continuous_learning_daily_20260917.json` — 1,555,702 bytes	Loaded	Pipeline stages, calibration, lineage, score distribution, starter comparisons
`detailed_20260917.html` — 214,325 bytes	Loaded	Full game and simulation rendering
`20260917_ultra_detailed_player_audit.html` — 20,306,685 bytes	Loaded	Player-level prediction trace
`audit_report_20260917.html` — 1,143,113 bytes	Loaded	Structural and invariant findings
`audit_report_20260917_integrity_revision.html`	Missing	Post-Luna remediation proof for immutable revision artifacts

The missing integrity-revision audit is an evidence gap. It does not invalidate the settled-game results, because the structural audits, prediction artifacts, and game-level audit receipts are present, but it prevents independent confirmation of post-remediation immutable-revision status.

Slate coverage

All nine expected games were reviewed:

BOS@TEX
DET@CHW
KC@HOU
LAD@CIN
MIL@PIT
MIN@LAA
OAK@TB
PHI@NYM
SD@COL

Coverage was complete:

Expected games: 9
Games structurally audited: 9
Forensic deep-audit games: 9
Continuous-learning games: 9
Starter comparisons: 9
Canonical artifacts loaded: 9
Learning eligibility: eligible; all critical/error structural gates passed

All nine game-level structural audits reported zero critical findings, zero errors, and zero warnings. They reported 70 informational findings in aggregate.

Internal contradictions and evidence limitations

BOS@TEX probability-lineage wording is internally awkward.

The row says published_minus_calibrated = -0.009473, the transformation says “no recorded transformation explains the publication delta,” but unexplained_material_delta is false. The delta is small and not material, so this is not a material bypass, but the wording should be normalized.

The missing integrity-revision artifact remains unresolved.

It is the only missing manifest entry.

Published probability and confidence edge use different representations.

For example, LAD@CIN has a governed/calibrated probability of 0.7106, but its published away-side probability is 0.56 and its displayed home-minus-away edge is -0.12. This is explainable by the named score-coherence and recency transformations, but reports should make the selected side and side-relative probability explicit.

The pipeline-stage table contains a meaningful winner-direction regression that is not obvious from aggregate final accuracy.

BOS@TEX moves historical ML TEX → TIE, then enhanced prediction moves TIE → TEX. The later recovery does not erase the intermediate regression.

No market cross-check was available.

There were zero source-labelled moneyline covers, so market agreement/disagreement cannot be evaluated.

Player-level results are noisy.

Only 13 of 53 material player deviations were supported by the hitter’s own season contact quality, a learnable share of 24.5%. The other deviations should not be treated as feature evidence.

---

2. Executive diagnosis

Overall result

Winner accuracy: 4/9 = 44.44%
Misses: 5
Reported walk-forward accuracy: 56.47%
Reported walk-forward Brier: 0.243312
Slate winner Brier from game records: average approximately 0.2568
Mean team-run error: approximately 3.67 runs
Mean total-run bias across the slate: approximately +1.11 runs
High-conviction games: 4
High-conviction misses: 2
High-conviction miss rate: 50%
Published confidence was not anti-predictive by the formal flag, but it was materially overconfident in the settled calibration buckets.

This was not a good prediction slate merely because four winners were correct. Winner accuracy was below the stated walk-forward baseline and below coin-flip performance on this slate. Exact-score performance was substantially weaker than winner performance.

Separate diagnoses

Winner accuracy

Poor on this slate:

4/9 correct
44.44% versus a 56.47% walk-forward baseline
Self-assessment: beat_coinflip: false, beat_validated_baseline: false
Trend: winner_accuracy is regressing; current slate is the lowest value in the displayed recent series at 44.44%

The five misses were BOS@TEX, DET@CHW, MIL@PIT, MIN@LAA, and PHI@NYM.

Calibration quality

Calibration was weak and overconfident:

0.50–0.55 bucket: average confidence 0.532, observed hit rate 0.400, calibration error +0.132
0.55–0.60 bucket: average confidence 0.5635, observed hit rate 0.500, calibration error +0.0635
Settled games in calibration monitor: 9
Weighted absolute calibration error: 0.1016
High-confidence cluster: 2/2 hits, but the cluster’s empirical hit rates were only 0.5536 and 0.5846 despite raw probabilities above 0.65

The model did reduce high raw probabilities through score coherence and matchup-recency guardrails. That is good lineage behavior, but the residual published confidence still overstated reliability in the moderate and lean buckets.

Score accuracy

Score accuracy was materially weak:

Holdout R²:
Away runs: 0.02324
Home runs: 0.002745
Total runs: 0.005629
The causal register correctly labels this a high-severity weak-run-model issue.
The model predicted margins of 2 runs or fewer in 6/9 games, while actual margins of at least 4 occurred in 4/9.
Actual margins: 1, 2, 4, 6, 3, 1, 9, 3, 7
Predicted margins: 1, 1, 1, 4, 2, 1, 3, 1, 3
Signed margin bias: -0.778
Cross-slate margin bias trend: regressing

The system systematically compressed the margin distribution, particularly in OAK@TB and SD@COL.

Starter accuracy

Starter performance was mixed and often decisive:

Mean starter ERA error: 2.16
Five of 17 listed starts exceeded projected ERA by at least 3.00:
Kyle Harrison: +16.2 ERA error
Brady Singer: +11.19
Tanner Gordon: +10.2
Jeffrey Springs: +8.49
Seth Lugo: +5.4
One thin-history starter was identified: Miguel Ullola, six prior appearances in the feed.
The starter lever had balanced cross-game evidence: 4 games helped, 4 games hurt. No broad starter-weight adjustment is justified from this slate alone.

Venue-split evidence was specifically credible for several blown starters, especially Seth Lugo, Kyle Harrison, and Tanner Gordon. This supports a bounded shadow test of context.venue_adjustment_weight, but not an unvalidated production change.

Artifact integrity

Artifact integrity was strong for the settled game artifacts:

All nine game-level structural audits passed.
All reported zero critical, error, and warning findings.
Probability reversions:
Artifacts checked: 9
Blend flips: 0
Side survived to publication: 9
Overwritten: 0
Six material calibrated-to-published movements were explained by score-coherence and/or recency transformations.
Simulation produced zero winner flips on this slate.

However, the missing integrity-revision audit prevents a fully closed artifact-revision receipt.

Causal issue register reconciliation

`weak_run_model_signal` — High

Severity: high
Evidence: home/away/total R² = 0.0027/0.0232/0.0056
Affected games: all nine
Action: treat exact scores as low-confidence; improve features or target before tuning score tails
Guardrail: a small positive R² only means the model beats guessing the average; it does not make exact scores reliable

This is a material model-validation risk, not a contract failure. The appropriate action is an isolated feature/target backtest, not a nightly score tweak.

`complete_slate_audit_coverage` — Info

Severity: info
Evidence: 9/9 games structurally audited
Affected games: all nine
Action: retain per-game audit receipts
Guardrail: full audit coverage does not cancel within-game findings

Correctly classified as informational.

`probability_publication_delta_explained` — Info

Severity: info
Evidence: six material movements, all named as score-coherence and/or recency transformations
Affected games: DET@CHW, KC@HOU, LAD@CIN, MIL@PIT, OAK@TB, SD@COL
Action: continue displaying every probability layer and transformation
Guardrail: a published probability above the calibrated probability is not a governor bypass when the transformation is recorded

This is supported. BOS@TEX has a small wording inconsistency but no material unexplained delta.

`simulation_inert_on_slate` — Info for this slate, escalated persistent failure across slates

Severity in the register: info
Evidence on this slate: zero flip candidates and zero published flips
Affected games: all nine
Action on the slate: measure simulation contribution across rolling slates
Guardrail: zero winner flips is not itself a defect and does not prove that no score reconciliation occurred

Cross-slate evidence changes the operational diagnosis: this has persisted for 11 consecutive slates and is now an established code/diagnostic failure pattern.

---

3. Every game

The game records use confidence_home_minus_away; negative values indicate an away-side edge. The published pick probability below is the probability of the selected side.

BOS@TEX — miss; correct-winner/wrong-score: no

Predicted: TEX 4, BOS 3; predicted winner TEX
Actual: BOS 4, TEX 3; winner BOS
Published selected-side probability: TEX 0.553827
System confidence edge: home-minus-away +0.1077, moderate
Brier: 0.306724
Team-run absolute error: 2
Total-run bias: 0
Starter lines:
BOS Sonny Gray: projected 2 ER/18 outs, actual 3 ER/19 outs; ERA 3.00 projected vs 4.26 actual; ER error +1
TEX Tyler Alexander: projected 1 ER/15 outs, actual 0 ER/3 outs; ERA 1.80 vs 0.00; ER error -1
Edges:
Season home-minus-away: -0.484
Recent: +0.262
Pitcher: -0.061
Composite: -0.154
Controls:
Governor applied
Confidence pipeline reached 0.5633 before publication
No recency guardrail
No score-volatility change
Starter-experience shrinkage not applied
Score path: 3.915–4.294 historical ML → 4–4 variance baseline → 3–4 enhanced → 3–4 through Savant, target review, cap, and simulation.
Player adjustments:
Sonny Gray predicted strikeouts increased by 2
Danny Jansen RBI projection worsened by 1.67
Jake Burger RBI projection improved by 1.45
Corey Seager AB projection worsened by 1.27
Trevor Story strikeout projection worsened by 1.34
Material player changes: 22
Audit: PASS; zero critical/error/warning findings; eight informational notes. Saved artifact was not the cause.
Causes:
Primary: wrong conviction side
Secondary: none
Actual game-level explanation: BOS lineup scored below projection by 2, TEX lineup below projection by 1; result was close, but the directional side was wrong.
Important caution: this was a moderate 0.11 edge, not a near-coin-flip classification and not a high-conviction pick.

DET@CHW — miss; correct-winner/wrong-score: no

Predicted: DET 5, CHW 4; predicted winner DET
Actual: DET 1, CHW 3; winner CHW
Published selected-side probability: DET approximately 0.530
System confidence edge: home-minus-away -0.060, lean
Brier: 0.280900
Team-run absolute error: 5
Total-run bias: +5
Starter lines:
DET Framber Valdez: projected 3 ER/18 outs, actual 3 ER/15 outs; projected ERA 4.50 vs actual 5.40; ER error 0
CHW Erick Fedde: projected 5 ER/14 outs, actual 1 ER/14 outs; projected ERA 9.64 vs actual 1.93; ER error -4
Edges:
Season: -0.442
Recent: -0.770
Pitcher: +0.084
Composite: -0.431
Controls:
Governor and recency guardrail applied
Confidence pipeline reached 0.5837 before later controls
Starter-experience shrinkage not applied
supported_blowout_margin_expansion applied to the away side
Score path: 4.539–4.383 historical → 5–4 variance baseline → 5–4 enhanced → 5–4 through final simulation.
Player adjustments:
Javier Báez RBI projection improved by 3.66
Randal Grichuk AB projection worsened by 2.50
Chase Meidroth strikeout projection worsened by 2.07
Miguel Vargas hit projection worsened by 1.89
Material player changes: 25
Audit: PASS; zero critical/error/warning findings; seven informational notes.
Causes:
Primary: starter projection
Secondary: run environment over
Fedde allowed four fewer ER than projected. DET’s offense was over-projected by four. This was a large score miss and a wrong side, not merely a one-run variance outcome.

KC@HOU — hit; correct-winner/wrong-score: yes

Predicted: HOU 5, KC 4; predicted winner HOU
Actual: HOU 6, KC 2; winner HOU
Published selected-side probability: HOU 0.529531
System confidence edge: home-minus-away +0.0591, lean
Brier: 0.221341
Team-run absolute error: 3
Total-run bias: +1
Starter lines:
KC Seth Lugo: projected 1 ER/15 outs, actual 4 ER/15 outs; ERA 1.80 vs 7.20; ER error +3
HOU Miguel Ullola: projected 4 ER/9 outs, actual 1 ER/2 outs; ERA 12.00 vs 13.50; ER error -3
Edges:
Season: +0.118
Recent: -0.234
Pitcher: +0.011
Composite: +0.211
Controls:
Governor and recency guardrail applied
Thin-history starter shrinkage applied to Ullola: six prior appearances, shrink factor 0.675; probability changed from 0.5606 to 0.540874 without changing the winner
No score-volatility change
Large Savant run-environment swing flag
Score path: 4.539–4.383 historical → 5–4 variance → 5–4 enhanced → 5–4 final.
Player adjustments:
LaMonte Wade Jr. exceeded projection by 1.85 runs
Taylor Trammell by 1.26
Jeremy Peña by 1.09
Daulton Varsho by 0.93
Jac Caglianone underperformed by 0.78
Audit: PASS; zero critical/error/warning findings; seven informational notes.
Causes: no registered primary/secondary driver. The winner was correct, but the score was wrong.
Forensic interpretation: Seth Lugo’s credible away split was underweighted. The recommended shadow lever is context.venue_adjustment_weight, not an immediate production change.

LAD@CIN — hit; correct-winner/wrong-score: yes

Predicted: LAD 9, CIN 5; predicted winner LAD
Actual: LAD 8, CIN 2; winner LAD
Published selected-side probability: LAD 0.56
System confidence edge: home-minus-away -0.12, moderate
Brier: 0.193600
Team-run absolute error: 4
Total-run bias: +4
Starter lines:
LAD Justin Wrobleski: projected 3 ER/15 outs, actual 0 ER/9 outs; ERA 5.40 vs 0.00; ER error -3
CIN Brady Singer: projected 4 ER/14 outs, actual 7 ER/10 outs; ERA 7.71 vs 18.90; ER error +3
Edges:
Season: -1.929
Recent: -1.414
Pitcher: -0.059
Composite: -2.264
Controls:
Governor, score coherence, and matchup-recency guardrail applied
Confidence pipeline reached 0.7106 before publication
supported_blowout_margin_expansion
High raw cluster member; published probability was reduced to 0.56
Score path: approximately 5–4 historical → 6–4 variance → 8–5 enhanced → 9–5 final.
Player adjustments:
Kyle Tucker exceeded projection by 2.63 runs, supported by above-average xwOBA 0.330
Miguel Rojas by 1.87
Sal Stewart by 1.40, supported by elite xwOBA 0.346
Elly De La Cruz underperformed by 1.73
Audit: PASS; zero critical/error/warning findings; eight informational notes.
Causes: none registered; this was a solid winner call with score overstatement.
Important classification: correct winner, wrong score. Brady Singer’s strong-offense matchup and large ERA miss were directionally useful, but the exact 9–5 forecast was too high.

MIL@PIT — miss; correct-winner/wrong-score: no

Predicted: MIL 6, PIT 4; predicted winner MIL
Actual: MIL 4, PIT 7; winner PIT
Published selected-side probability: MIL 0.56
System confidence edge: home-minus-away -0.12, moderate
Brier: 0.313600
Team-run absolute error: 5
Total-run bias: -1
Starter lines:
MIL Kyle Harrison: projected 3 ER/15 outs, actual 4 ER/5 outs; ERA 5.40 vs 21.60; ER error +1, ERA error +16.2
PIT Wilber Dotel: projected 1 ER/15 outs, actual 1 ER/6 outs; ERA 1.80 vs 4.50; ER error 0
Edges:
Season: -1.170
Recent: -1.138
Pitcher: -0.061
Composite: -1.423
Controls:
Governor and recency guardrail applied
supported_blowout_margin_expansion toward MIL
Confidence pipeline reached 0.6562 before publication
Score path: 5–4 historical → 5–4 variance → 6–4 enhanced → 6–4 final.
Player adjustments:
Brandon Lowe improved by 2.69 projected hits
Joey Ortiz AB improved by 1.85
Sal Frelick RBI improved by 1.82
Jackson Chourio underperformed by 1.36
Bo Naylor exceeded RBI projection by 1.75
Audit: PASS; zero critical/error/warning findings; eight informational notes.
Causes:
Primary: wrong conviction side
Secondary: lean-side miss in the registered labels, despite the displayed moderate 0.12 edge
Kyle Harrison’s credible away split was underweighted; PIT offense was under-projected by 3.
A 0.12 edge is moderate, not a near coin flip. The problem was not that the edge was “almost zero”; it was that a moderate directional signal was treated too confidently relative to the calibration evidence.

MIN@LAA — miss; correct-winner/wrong-score: no

Predicted: MIN 4, LAA 3; predicted winner MIN
Actual: MIN 4, LAA 5; winner LAA
Published selected-side probability: MIN 0.516481
System confidence edge: home-minus-away -0.033, toss-up
Brier: 0.266753
Team-run absolute error: 2
Total-run bias: -2
Starter lines:
MIN Taj Bradley: projected 3 ER/18 outs, actual 3 ER/21 outs; ERA 4.50 vs 3.86; ER error 0
LAA Walbert Ureña: projected 1 ER/17 outs, actual 0 ER/21 outs; ERA 1.59 vs 0.00; ER error -1
Edges:
Season: +0.079
Recent: +0.407
Pitcher: -0.034
Composite: -0.110
Controls:
Governor and recency guardrail applied
Confidence pipeline reached 0.5228
No score-volatility change
No experience shrinkage
Score path: 4.523–3.893 historical → 5–4 variance → 4–3 enhanced → 4–3 final.
Player adjustments:
Walker Jenkins exceeded projection by 2.01 runs
Oswald Peraza by 0.94
Zach Neto underperformed by 1.40
Mike Trout strikeout projection worsened by 1.17
Audit: PASS; zero critical/error/warning findings; eight informational notes.
Causes:
Primary: toss-up variance
Secondary: none
This is not evidence for a model change. Both starters were close to the projected ER lines, and the actual one-run result was consistent with ordinary variance.

OAK@TB — hit; correct-winner/wrong-score: yes

Predicted: TB 7, OAK 4; predicted winner TB
Actual: TB 10, OAK 1; winner TB
Published selected-side probability: TB 0.58
System confidence edge: home-minus-away +0.16, moderate
Brier: 0.176400
Team-run absolute error: 6
Total-run bias: 0
Starter lines:
OAK Jeffrey Springs: projected 4 ER/14 outs, actual 6 ER/10 outs; ERA 7.71 vs 16.20; ER error +2
TB Drew Rasmussen: projected 2 ER/18 outs, actual 0 ER/15 outs; ERA 3.00 vs 0.00; ER error -2
Edges:
Season: +1.683
Recent: +1.202
Pitcher: +0.052
Composite: +1.880
Controls:
Governor and recency guardrail applied
Confidence pipeline reached 0.6502
supported_blowout_margin_expansion
Large Savant run-environment swing flag
Score path: roughly 5–4 historical → 5–4 variance → 6–4 enhanced → 7–4 final.
Player adjustments:
Lawrence Butler exceeded projection by 1.07 runs, supported by above-average xwOBA 0.321
Richie Palacios by 0.97
Junior Caminero underperformed by 1.07
Zack Gelof underperformed by 0.93
Audit: PASS; zero critical/error/warning findings; eight informational notes.
Causes: none registered; correct winner, wrong score.
Important classification: correct winner, wrong score. The model identified the side but materially understated the margin.

PHI@NYM — miss; correct-winner/wrong-score: no

Predicted: NYM 4, PHI 3; predicted winner NYM
Actual: PHI 3, NYM 0; winner PHI
Published selected-side probability: NYM 0.544183
System confidence edge: home-minus-away +0.0884, lean
Brier: 0.296135
Team-run absolute error: 4
Total-run bias: +4
Starter lines:
PHI Aaron Nola: projected 4 ER/18 outs, actual 0 ER/21 outs; ERA 6.00 vs 0.00; ER error -4
NYM Nolan McLean: projected 2 ER/18 outs, actual 1 ER/18 outs; ERA 3.00 vs 1.50; ER error -1
Edges:
Season: -0.312
Recent: +0.358
Pitcher: +0.068
Composite: +0.247
Controls:
Governor and recency guardrail applied
Confidence pipeline reached 0.5612
No score-volatility change
Large Savant run-environment swing flag
Score path: 4–5 historical → 4–5 variance → 3–4 enhanced → 3–4 final.
Player adjustments:
Bryce Harper exceeded projection by 0.94 runs, supported by elite xwOBA 0.385
Kyle Schwarber underperformed by 1.40
Trea Turner AB projection worsened by 2.94
Francisco Alvarez RBI projection improved by 1.59
Audit: PASS; zero critical/error/warning findings; eight informational notes.
Causes:
Primary: starter projection
Secondary: lean-side miss
Nola’s four-ER projection error was decisive. The forensic attribution calls the start an outlier with no credible venue, opponent, or experience explanation. No model change is warranted for Nola’s individual result.

SD@COL — hit; correct-winner/wrong-score: yes

Predicted: SD 8, COL 5; predicted winner SD
Actual: SD 9, COL 2; winner SD
Published selected-side probability: SD 0.54
System confidence edge: home-minus-away -0.08, lean
Brier: 0.211600
Team-run absolute error: 4
Total-run bias: +2
Starter lines:
SD Michael King: projected 3 ER/19 outs, actual 2 ER/18 outs; ERA 4.26 vs 3.00; ER error -1
COL Tanner Gordon: projected 6 ER/15 outs, actual 7 ER/9 outs; ERA 10.80 vs 21.00; ER error +1
Edges:
Season: -1.504
Recent: -1.654
Pitcher: -0.064
Composite: -1.922
Controls:
Governor and recency guardrail applied
Confidence pipeline reached 0.6972
supported_blowout_margin_expansion
High raw cluster member; published probability reduced to 0.54
Score path: roughly 5–4 historical → 5–4 variance → 7–4 enhanced → 8–5 final.
Player adjustments:
Xander Bogaerts exceeded projection by 2.47
Dustin Harris by 0.94, supported by above-average xwOBA 0.336
Manny Machado underperformed by 1.73
Tanner Gordon’s credible home weakness was underweighted
Audit: PASS; zero critical/error/warning findings; eight informational notes.
Causes: none registered; solid correct winner call with wrong score.
Important classification: correct winner, wrong score.

Game-level summary

Correct winners:

KC@HOU
LAD@CIN
OAK@TB
SD@COL

Correct winner but wrong score:

KC@HOU
LAD@CIN
OAK@TB
SD@COL

Misses:

BOS@TEX — wrong side
DET@CHW — starter and run-environment miss
MIL@PIT — wrong side, underweighted venue split
MIN@LAA — toss-up variance
PHI@NYM — starter projection miss and lean-side miss

There were no exact-score hits reported.

---

4. Miss reconstruction

Earliest wrong-direction stages

BOS@TEX

Earliest adverse movement: historical ML projected TEX ahead, but variance application moved the direction to a tie; enhanced prediction restored TEX.
The decisive problem was not a late artifact or simulation flip. The fundamental evidence was contradictory:
Season edge favored BOS by 0.484
Recent form favored TEX by 0.262
Pitcher edge favored BOS by 0.061
Composite favored BOS by 0.154
Diagnosis: bad weighting / wrong conviction side, with irreducible close-game variance also present.
The 0.11 edge was moderate, not near-coin-flip confidence.

DET@CHW

Historical ML already favored DET, consistent with the eventual wrong direction.
Enhanced, Savant, target review, and simulation preserved DET.
The earliest wrong evidence was the run/starter input structure that projected DET for 5 and CHW for 4, especially the CHW starter line of 5 ER/9.64 ERA.
Actual Fedde allowed 1 ER, four fewer than projected.
Diagnosis: bad starter input/projection, compounded by run-environment overprojection and overconfidence in the away side.

MIL@PIT

Historical ML favored MIL; variance preserved that direction.
Enhanced prediction increased MIL’s expected runs from 5 to 6 and maintained the wrong side.
Savant adjustments slightly reduced both teams but did not correct direction.
Kyle Harrison’s away split was credible and materially adverse, yet the venue-role weakness was underweighted.
Diagnosis: bad weighting, specifically venue-split weighting, plus a large starter realization error and PIT offense underprojection.

MIN@LAA

Historical ML favored MIN; variance, enhanced prediction, and final simulation retained MIN.
The evidence was nearly balanced:
Composite home edge only -0.110
Published away probability about 0.516
Displayed edge only 0.033
Both starters were broadly accurate by ER, and the actual one-run result flipped the predicted side.
Diagnosis: irreducible variance, not a validated model defect.

PHI@NYM

Historical ML favored NYM and the enhanced score path retained NYM.
The strongest adverse realization was Aaron Nola: projected 4 ER, actual 0 ER.
No credible contextual signal explains the outlier.
Diagnosis: starter projection miss, specifically irreducible/outlier variance, plus a lean-side miss.

Loss-cause classification

Across five losses:

Starter projection: 2 registered primary drivers — DET@CHW and PHI@NYM
Toss-up variance: 1 — MIN@LAA
Wrong conviction side: 2 — BOS@TEX and MIL@PIT
Bad input: most clearly CHW starter projection line; no structural identity error was found
Bad weighting: venue-role weighting for MIL@PIT, and possibly BOS@TEX conflicting season/recent weighting
Calibration: confidence was too high relative to observed performance, but calibration did not independently create each wrong side
Offense/bullpen miss:
PIT offense under-projected by 3 in MIL@PIT
DET offense over-projected by 4 in DET@CHW
NYM offense underperformed by 4 in PHI@NYM
Run-environment miss: strongest in DET@CHW and PHI@NYM; total runs were over-projected by 5 and 4 respectively
Irreducible variance: strongest in MIN@LAA and PHI@NYM’s Nola outlier

A 0.12 confidence edge is moderate. It should not be described as a near coin flip, but neither should it be treated as high-confidence certainty. The appropriate response is calibration and distribution control, not removal of all directional signal.

Pipeline-stage quality across all stages

Stage	Winner accuracy	Team-run MAE	Total-run MAE	Margin MAE
historical_ml	4/9 = 44.44%	2.0863	2.1738	3.5824
variance_applied	4/9	2.0556	2.1111	3.6667
enhanced_prediction	4/9	1.8889	2.0000	3.5556
embedded_savant_run_environment	4/9	1.9444	2.1111	3.4444
fundamental_edge_target_review	4/9	1.9444	2.1111	3.4444
capped_target	4/9	1.9444	2.1111	3.4444
simulation_result	4/9	1.9444	2.1111	3.4444
published_final	4/9	1.9444	2.1111	3.4444

Transitions:

Historical ML → variance applied:

Margin MAE regressed +0.0843
Team-run MAE improved -0.0307
Winner accuracy unchanged
BOS moved TEX → TIE: an intermediate winner-direction regression

Variance applied → enhanced prediction:

Margin MAE improved -0.1111
Team-run MAE improved -0.1667
BOS moved TIE → TEX: later recovery
This recovery does not erase the intermediate regression

Enhanced → embedded Savant:

Margin MAE improved -0.1112
Team-run MAE worsened +0.0555
Winner accuracy unchanged

Embedded Savant → fundamental review:

No change

Fundamental review → capped:

No change

Capped → simulation:

No change

Simulation → published final:

No change

The system’s later stages improved run-error metrics relative to historical ML, but none improved winner accuracy, and simulation made no measurable winner-direction intervention.

---

5. Cross-game counterfactuals

Lever tension: widen run distributions

Lever: score.variance_multiplier / score.variance_floor

Targets: Brier score, calibration error

Helps

BOS@TEX: lost a 0.11-confidence pick; wider tails would reduce overconfidence in a close matchup.
MIL@PIT: lost a 0.12-confidence pick; wider tails would better represent the possibility of a moderate edge failing.

Hurts

No harmed games were recorded in the lever-tension entry.

Net evidence

2 games helped, 0 hurt on this slate.
Cross-slate evidence shows lineup runs were below projection on 8 of 12 slates, but the registered bias is not statistically separable from noise.
A bounded shadow experiment is reasonable, but widening variance cannot repair the near-zero run-model signal or guarantee better exact scores.

Lever tension: pitcher form weights / prior games

Lever: player.pitcher.v4 form weights / player.pitcher_prior_games

Targets: starter ER MAE, starter outs MAE

Helps

DET@CHW: Fedde was four ER better than projected.
KC@HOU: Lugo was three ER worse and Ullola three ER better than projected.
LAD@CIN: Singer was three ER worse than projected.
PHI@NYM: Nola was four ER better than projected.

Hurts

BOS@TEX: starter projection was relatively accurate; ER error 1.
MIL@PIT: starter ER error 1 for Harrison under the registered absolute-ER measure.
MIN@LAA: starter ER errors 0 and -1.
SD@COL: starter ER errors -1 and +1.

Net evidence

4 helped, 4 hurt.
No broad form-weight change is supported.
The venue-specific problem is narrower than the general pitcher-form lever. Use context.venue_adjustment_weight in a shadow test for credible split cases instead of changing global pitcher priors.

Lever tension: recent/season run-mean weighting

Lever: recent.* / season.* run-mean weighting

Targets: team-run MAE, total-run MAE, accuracy

Helps

DET@CHW: reduced reliance on an incorrect high run environment would help.
LAD@CIN: score was directionally correct; better weighting may reduce the 4-run total overprojection.
MIL@PIT: PIT offense was under-projected.
OAK@TB: winner was right, but margin was understated.
PHI@NYM: lower run environment would help.
SD@COL: margin was understated.

Hurts

BOS@TEX: team-run error was only 2; more reweighting could disturb a close score.
MIN@LAA: team-run error was only 2; no evidence supports a broad adjustment from this game.

Net evidence

6 helped, 2 hurt.
This is the strongest slate-level bounded model lever among the listed tensions, but cross-slate evidence does not support an unvalidated production change because run-model R² is near zero. A shadow change is preferable.

Tempting fixes that should not be made

Do not globally increase pitcher form weighting from these nine games; the evidence is 4–4.
Do not tune to individual outlier starts such as Aaron Nola, Fedde, or Singer.
Do not treat all 53 player deviations as learnable; only 13 had supporting season contact-quality evidence.
Do not increase confidence simply because LAD, OAK, or SD won; those were correct winners but wrong scores.
Do not remove moderate edges entirely because BOS and MIL lost. A 0.12 edge is moderate, not a coin-flip label.
Do not force a parameter change for the persistent simulation failure; the registered action is a code/diagnostic fix.
Do not tune score tails as a substitute for fixing the near-zero run-model features and target.

---

6. Data and contract defects

Critical findings

No critical audit findings were reported in the nine game audits or the structural audit summary.

The missing integrity-revision artifact is a manifest gap, but it was not reported as a critical settled-game contract failure.

Error findings

No error audit findings were reported.

Warning findings

No warning audit findings were reported.

Informational findings

The aggregate structural audit reported 70 informational findings. The principal informational findings were:

All nine games structurally audited.
All nine game artifacts passed.
Six material calibrated-to-published probability movements had named transformations.
Simulation produced zero winner flips.
Numerous game-level informational notes documented confidence controls, player changes, score-volatility decisions, and recency handling.

Model defects versus artifact bugs

Prediction-model defects

Weak run-model signal — high severity

Total-run R² 0.005629
Home-run R² 0.002745
Away-run R² 0.02324
Requires feature/target investigation and isolated retraining backtest.

Simulation inertness — persistent code/diagnostic defect

11 consecutive slates
Zero flip candidates and zero published flips on this slate
The self-assessment identifies the likely cause: simulation uses the same target-realization draws, early exits after finding a play-by-play realization, and then treats roughly two selected draws as a probability distribution.
This is a code/diagnostic problem, not a parameter-tuning problem.

Margin compression

Actual margins ≥4 in 4/9 games, predicted margins ≤2 in 6/9.
Cross-slate margin-bias trend is regressing.

Calibration overconfidence

Weighted absolute calibration error 0.1016
50% high-conviction miss rate
Moderate and lean buckets both overconfident.

Player projection separation

Hit and strikeout spread ratios only 0.402 and 0.423.
RBI per-player model lost to a flat slate mean: MAE 0.642 versus 0.597.
HR model also lost to the flat mean: 0.189 versus 0.171.

Presentation/artifact defects

BOS@TEX lineage wording says no transformation explains a delta while marking it non-material and not unexplained. This should be clarified.
The missing integrity-revision audit is a presentation/receipt gap unless independently shown to reflect an actual missing immutable-revision check.
Probability layers should consistently identify:

raw model probability
calibrated probability
governor input
governor output
score-coherence probability
recency adjustment
published selected-side probability

Coverage reconciliation

The coverage matrix is clean for the nine settled games:

all_games_structurally_audited: true
canonical_artifacts_loaded: 9
continuous_learning_games: 9
expected_games: 9
forensic_deep_audit_games: 9
games_with_starter_comparisons: 9
structural_audits_completed: 9

The slate should not be called fully artifact-clean because the integrity-revision audit is missing, but it is structurally complete and learning-eligible.

Probability-lineage audit

Materially relevant rows:

DET@CHW: calibrated 0.5837 → published 0.53; named score-coherence and recency guardrail
KC@HOU: 0.5606 → 0.529531; named transformations
LAD@CIN: 0.7106 → 0.56; score coherence plus matchup-recency guardrail
MIL@PIT: 0.6562 → 0.56; named transformations
OAK@TB: 0.6502 → 0.58; named transformations
SD@COL: 0.6972 → 0.54; named transformations

These six material movements have recorded explanations; no material governor bypass is established.

BOS@TEX moved from calibrated 0.5633 to published 0.553827, a delta of only -0.009473. The row’s wording is inconsistent, but this is not a material unexplained delta.

MIN@LAA and PHI@NYM had small deltas and named recency/score-coherence handling.

No probability reversion or overwritten side was identified.

---

7. Cross-slate persistence

A single slate is not enough to establish most model patterns, but several findings have already escalated and must be treated as established failures.

Escalated findings

`simulation_inert_on_slate`

Consecutive-slate count: 11
Incidence in window: 11/11
Present now: yes
Target lever listed: simulation.allow_score
Parameter actionable: no
Metric group: run model
Metrics harmed: team-run MAE and total-run MAE through lack of independent simulation signal
Direction of fix: confirm the simulation module is running and not silently no-oping
Required action: code/diagnostic repair, not a forced parameter move

The self-assessment gives the concrete suspected defect: separate target-realization search from independent win-probability simulation; prevent early exit; use a registered batch size such as 200–500 draws or an independent closed-form distribution.

`weak_run_model_signal`

Consecutive-slate count: 11
Incidence in window: 11/11
Present now: yes
Target levers listed: training.xgb.n_estimators, training.xgb.max_depth, training.xgb.learning_rate
Parameter actionable: no, despite listed training levers
Metrics harmed: team-run MAE and total-run MAE
Direction of fix: investigate features and target with an isolated training backtest
Required action: add or validate market no-vig probability where source-labelled, park/weather environment, bullpen rest/fatigue, and platoon-adjusted lineup quality; compare same-fingerprint holdout R²; bump model schema if semantics change

This is an established failure, not single-slate noise.

Non-escalated finding streaks

Finding	Consecutive count	Present now	Interpretation
high_raw_probability_cluster	0	No	Occurred once in the 11-slate window; this slate had two high-raw games, but the registered streak is not current
pipeline_winner_instability	0	No	Incidence 2 in window, but no current streak
score_margin_compression	0	No	Incidence 1 in window under this finding label; current margin compression is visible but not escalated under this streak record
simulation_inert_on_slate	11	Yes	Established persistent failure
single_slate_calibration_instability	0	No	Seven occurrences in window but no consecutive current streak
weak_run_model_signal	11	Yes	Established persistent failure

Metric trends

Each direction is explicit:

margin_bias: regressing, mean -1.5357
mean_team_run_error_on_losses: improving, mean 5.5165, with current slate at 3.6
weighted_calibration_error: improving, mean 0.1781, current 0.1016
winner_accuracy: regressing, mean 0.5961, current 0.4444

“Improving” calibration and loss-run error do not mean they are good; they mean the recent series is moving in a favorable direction. Current calibration remains materially imperfect.

Top-confidence-pick loss streak

Reported loss streak: 0
Per-slate sequence from 20260907 through 20260917:
miss, hit, hit, hit, miss, miss, hit, miss, hit, hit, hit
Current top-confidence pick: hit
Therefore, no persistent top-confidence loss streak is present.

Other structural persistence

Margin compression appears repeatedly in the raw margin distribution and is reflected in the regressing margin-bias trend, but the formal finding_streaks entry reports only zero consecutive slates. It should be monitored rather than falsely labeled an escalated streak.
The run model’s weak holdout R² and simulation inertness are the strongest persistent structural signals.
The player-level spread problem is a current-slate diagnostic, not an escalated cross-slate finding in the supplied streak structure.

---

8. Adjustments for today

Recommendations are ranked by persistence, scope, expected metric impact, affected games, and regression risk.

1. Repair and independently measure simulation — highest evidence strength

Pattern: simulation_inert_on_slate, persistent for 11 slates
Exact registered target: simulation.allow_score
Parameter direction: do not force a parameter value; repair the code path
Bounded action:
Confirm simulation.allow_score is active.
Separate target-realization search from win-probability simulation.
Remove early exits from the probability batch.
Register an independent simulation batch size, initially 200–500 draws.
Expected metric:
Increase nonzero simulation contribution and flip-candidate rate
Improve winner-direction diagnostics
Potentially improve Brier and winner accuracy, but causal effect is not guaranteed
Helps:
All nine games in principle, especially close/moderate-edge games BOS@TEX, MIL@PIT, and MIN@LAA
Hurts:
No specific harmed game is established; risk is compute cost and introducing unvalidated probability noise
Gate:
Shadow only until independent simulation contribution is measurable.
Roll back if calibration error or winner accuracy worsens across a predefined rolling sample.

2. Run an isolated run-model feature/target backtest — persistent high-severity issue

Pattern: weak_run_model_signal, persistent for 11 slates
Exact registered levers: training.xgb.n_estimators, training.xgb.max_depth, training.xgb.learning_rate
Parameter direction: abstain from direct nightly parameter movement
Bounded action:
Add/test the registered feature-contract candidates: source-labelled no-vig market probability, park/weather environment, bullpen rest/fatigue, and platoon-adjusted lineup quality.
Compare same-fingerprint holdout R² and MAE.
Do not promote without improvement over the current R² baseline.
Expected metric:
Increase home, away, and total-run R²
Reduce team-run MAE, total-run MAE, and margin compression
Helps:
Broadly all games, with greatest relevance to DET@CHW, MIL@PIT, OAK@TB, PHI@NYM, and SD@COL
Hurts:
Potential overfit risk in BOS@TEX and MIN@LAA, where current team-run errors were only 2
Gate:
Isolated backtest, schema-version update, then shadow deployment. No direct production tuning from this slate.

3. Bounded confidence-cap reduction for the strongest single-slate driver

Strongest single-slate driver: confidence overstatement relative to outcome, evidenced by 44.44% winner accuracy, 0.1016 weighted calibration error, and a 50% high-conviction miss rate.
Exact registered parameter: confidence.moderate_cap
Direction: decrease
Small bounded magnitude: reduce by 0.02 from its current registered value
Expected metric:
Lower weighted calibration error and Brier score
Reduce high-conviction miss severity
No expected direct improvement in raw winner accuracy
Helps:
BOS@TEX and MIL@PIT, the moderate-edge losses
Potentially DET@CHW and PHI@NYM, the lean-side losses
Hurts:
Could reduce useful confidence in correct moderate calls such as LAD@CIN and OAK@TB
Could understate SD@COL, which was a correct lean call
Gate:
Apply only as a shadow calibration experiment alongside existing score-variance controls.
Roll back if rolling Brier or calibration error worsens.

4. Optional shadow test of run-mean weighting

Exact lever: recent.* / season.* run-mean weighting
Direction: modestly increase adaptive recent/season run-environment weighting only within a shadow branch; do not change production from this slate alone.
Expected metric:
Reduce team-run MAE, total-run MAE, and margin compression
Helps:
DET@CHW, LAD@CIN, MIL@PIT, OAK@TB, PHI@NYM, SD@COL
Hurts:
BOS@TEX and MIN@LAA, where scores were comparatively close
This is ranked below the persistent code and run-model actions because the lever is not tied to an escalated finding and the underlying run model has near-zero predictive signal.

Changes to abstain from today

Do not directly alter global pitcher form weights; the lever tension is 4–4.
Do not promote context.venue_adjustment_weight globally yet. It is a sensible shadow experiment for Lugo, Harrison, and Gordon, but it is not a registered escalated pattern.
Do not tune to the individual Nola, Fedde, Singer, or Springs outcomes.
Do not force a simulation parameter change in place of the required code repair.

---

9. Bottom line

Known

The slate went 4–9, or 44.44%, below the walk-forward baseline and below coin flip.
Four correct winners were also wrong on exact score.
The score model has near-zero holdout R² and materially compresses winning margins.
Confidence was overestimated in the 0.50–0.60 buckets; weighted calibration error was 0.1016.
Five losses were attributable to two wrong-conviction-side calls, two starter-projection failures, and one genuine toss-up variance result.
All nine game artifacts passed structural QA with no critical, error, or warning findings.
Six material probability movements were documented through named transformations; no material probability bypass is established.
Simulation inertness and weak run-model signal are persistent across 11 consecutive slates and are established failures.
The top-confidence pick did not have a loss streak; it was a hit on the current slate.

Uncertain

Whether venue-split weighting will improve starter forecasts across slates; this slate provides suggestive cases but balanced global starter evidence.
Whether widening score variance will improve Brier/calibration without worsening useful correct calls.
Whether recent/season run-mean weighting can overcome the underlying lack of run-model feature signal.
The causal impact of an independent simulation distribution until the code path is repaired and measured.
The missing integrity-revision artifact’s effect on immutable revision verification.

Single safest next action

Repair and instrument the simulation probability path before changing model parameters: separate independent win-probability draws from target-realization search, remove early exit, register the batch size, and run it in shadow with rollback monitoring. This directly addresses an 11-slate established failure, preserves the existing forecast path, and produces the most valuable evidence for the next slate without overfitting today’s nine games.

Metric	Prior	Current	Δ	Result
brier_score	0.2391	0.2425	+0.003405	regressed
accuracy	0.6167	0.5917	-0.025	regressed
team_run_mae	2.5458	2.5333	-0.0125	improved
total_run_mae	3.625	3.5333	-0.0917	improved
calibration_error	0.085	0.0728	-0.0122	improved
starter_er_mae	1.7333	1.7326	-0.0007	improved
starter_outs_mae	3.2292	3.2984	+0.0692	regressed

Deep slate report — 20260917

Process, end to end

Slate summary

What drove the losses

Source coverage receipt

High-level causal issue register

Cross-slate trends (persistent-pattern memory)

Escalated persistent patterns

Finding streaks across the window

Metric direction across the window

Pipeline stage regression

Probability lineage and calibration

Score distribution and margin compression

Per-game evidence coverage

Every game, in depth (9)

BOS@TEXMISS

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

DET@CHWMISS

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

KC@HOUHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

LAD@CINHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

MIL@PITMISS

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

MIN@LAAMISS

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

OAK@TBHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

PHI@NYMMISS

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

SD@COLHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

Cross-game lever tension

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

`weak_run_model_signal` — High

`complete_slate_audit_coverage` — Info

`probability_publication_delta_explained` — Info

`simulation_inert_on_slate` — Info for this slate, escalated persistent failure across slates

`simulation_inert_on_slate`

`weak_run_model_signal`