Deep slate report 20260906

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_20260906.json	n/a
forensic deep audit	loaded	probability, provenance, starter and decision contracts	/app/output/reports/PREDICTION_SYSTEM_DEEP_AUDIT_20260906.json	n/a
continuous learning	loaded	pipeline-stage quality, calibration, probability lineage	/app/output/learning/continuous_learning_daily_20260906.json	n/a
game artifacts	loaded	per-game model trace and structural audit source	/app/output/games	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 (15)

Severity	Domain	Finding and evidence	Affected	Action & guardrail
HIGH	model_validation	Validated run models explain very little holdout variance home/away/total R² = 0.0074/0.0182/0.0147	ARI@HOU, ATL@PHI, BOS@BAL, CHC@MIA, DET@CLE, LAA@PIT, MIL@CIN, MIN@CHW, NYY@SD, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC, WAS@LAD	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 15/15 games structurally audited	ARI@HOU, ATL@PHI, BOS@BAL, CHC@MIA, DET@CLE, LAA@PIT, MIL@CIN, MIN@CHW, NYY@SD, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC, WAS@LAD	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 8 material movement(s), all explained by named score-coherence and/or recency transformations	CHC@MIA, DET@CLE, MIL@CIN, MIN@CHW, SF@NYM, STL@COL, TOR@KC, WAS@LAD	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	ARI@HOU, ATL@PHI, BOS@BAL, CHC@MIA, DET@CLE, LAA@PIT, MIL@CIN, MIN@CHW, NYY@SD, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC, WAS@LAD	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	10 consecutive · 10 in window	simulation.allow_score	confirm the simulation module is running, not silently no-oping	simulation_inert_on_slate has persisted 10 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	10 consecutive · 10 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 10 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	3/10	no	param
pipeline_winner_instability	0	4/10	no	code/diagnostic
score_margin_compression	0	3/10	no	param
simulation_inert_on_slate	10	10/10	yes	code/diagnostic
single_slate_calibration_instability	0	5/10	no	param
weak_run_model_signal	10	10/10	yes	code/diagnostic

Metric	Direction	Series (old → new)
winner_accuracy	improving	[0.4667, 0.9412, 0.5, 0.6667, 0.4667, 0.3333, 0.5556, 0.625, 0.5333, 0.6]
weighted_calibration_error	improving	[0.2179, 0.3961, 0.0693, 0.1661, 0.2476, 0.21, 0.1474, 0.1316, 0.1, 0.0937]
margin_bias	improving	[-1.333, -1.882, -2.786, -3.5, -2.533, -3.067, -1.778, -0.688, -1.467, -1.467]
team_run_error_on_losses	improving	[5.0, 3.0, 7.0, 8.25, 7.125, 6.4, 6.75, 6.0, 5.286, 5.333]

Stage	Winner hits	Accuracy	Team-run MAE	Total-run MAE	Margin MAE
historical_ml	8/15	53.3%	2.676	4.645	2.446
variance_applied	8/15	53.3%	2.667	4.8	2.267
enhanced_prediction	9/15	60%	2.7	4.2	2.733
embedded_savant_run_environment	9/15	60%	2.667	4.133	2.667
fundamental_edge_target_review	9/15	60%	2.667	4.133	2.667
capped_target	9/15	60%	2.667	4.133	2.667
simulation_result	9/15	60%	2.667	4.133	2.667
published_final	9/15	60%	2.667	4.133	2.667

Transition	Δ accuracy	Δ hits	Δ team-run MAE	Δ margin MAE	Games whose winner/tie changed
historical_ml → variance_applied	0	0	-0.0095	-0.1788	ATL@PHI: PHI → TIE; CHC@MIA: CHC → TIE; NYY@SD: NYY → TIE; TOR@KC: TOR → TIE; WAS@LAD: WAS → TIE
variance_applied → enhanced_prediction	0.0667	1	0.0333	0.4666	ATL@PHI: TIE → PHI; CHC@MIA: TIE → CHC; NYY@SD: TIE → NYY; TOR@KC: TIE → TOR; WAS@LAD: TIE → LAD
enhanced_prediction → embedded_savant_run_environment	0	0	-0.0333	-0.0666
embedded_savant_run_environment → fundamental_edge_target_review	0	0	0	0
fundamental_edge_target_review → capped_target	0	0	0	0
capped_target → simulation_result	0	0	0	0
simulation_result → published_final	0	0	0	0

Game	Outcome	Raw model	Calibrated	Governor	Score coherence	Published	Explanation
ARI@HOU	hit	0.5431	0.5439	0.5439	0.5439	0.5373	no recorded transformation explains the publication delta
ATL@PHI	miss	0.5382	0.5596	0.5596	0.5596	0.5506	no recorded transformation explains the publication delta
BOS@BAL	hit	0.5869	0.5682	0.5682	0.5682	0.5493	score coherence followed by matchup-recency guardrail
CHC@MIA	miss	0.5979	0.6773	0.6773	0.6712	0.58	score coherence followed by matchup-recency guardrail
DET@CLE	hit	0.5774	0.5452	0.5452	0.5452	0.5203	score coherence followed by matchup-recency guardrail
LAA@PIT	hit	0.612	0.5858	0.5858	0.5851	0.5723	no recorded transformation explains the publication delta
MIL@CIN	miss	0.6356	0.6257	0.6257	0.6257	0.53	score coherence followed by matchup-recency guardrail
MIN@CHW	hit	0.652	0.5833	0.5833	0.5833	0.53	score coherence followed by matchup-recency guardrail
NYY@SD	miss	0.5007	0.5494	0.5494	0.5494	0.5357	score coherence followed by matchup-recency guardrail
OAK@SEA	hit	0.5843	0.5612	0.5612	0.5612	0.5442	no recorded transformation explains the publication delta
SF@NYM	hit	0.5825	0.5475	0.5475	0.5475	0.5163	score coherence followed by matchup-recency guardrail
STL@COL	hit	0.6129	0.5821	0.5821	0.5821	0.5594	score coherence followed by matchup-recency guardrail
TB@TEX	miss	0.5636	0.5379	0.5379	0.5379	0.5274	score coherence followed by matchup-recency guardrail
TOR@KC	miss	0.5669	0.5724	0.5724	0.5724	0.5371	score coherence followed by matchup-recency guardrail
WAS@LAD	hit	0.5293	0.5772	0.5772	0.5772	0.53	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
30	6	10	2	12	0.74	-0.3

Slate	Games	Winner acc	Mean starter ERA error	Mean lineup runs vs proj	Sides broke on pitching	Sides broke on offense
20260906	15	0.600	—	—
20260905	15	0.533	-1.17	-0.07	2	6
20260904	16	0.625	-0.23	-0.44	5	12
20260903	9	0.556	0.24	-0.83	4	6
20260902	15	0.333	0.72	0.30	7	6
20260901	15	0.467	1.48	1.07	9	3
20260831	12	0.667	0.83	0.08	2	7
20260830	14	0.500	-0.36	-0.29	3	7
20260829	17	0.941	—	—
20260828	15	0.467	—	—

Game	Pred margin	Actual margin	Pred − actual
ARI@HOU	1	1	0
ATL@PHI	1	1	0
BOS@BAL	1	2	-1
CHC@MIA	2	7	-5
DET@CLE	1	1	0
LAA@PIT	2	1	1
MIL@CIN	2	4	-2
MIN@CHW	2	9	-7
NYY@SD	1	1	0
OAK@SEA	1	2	-1
SF@NYM	2	2	0
STL@COL	1	2	-1
TB@TEX	1	2	-1
TOR@KC	1	5	-4
WAS@LAD	1	2	-1

ARI@HOUHIT

Predicted: ARI 4-3 (conf -0.07, lean)Actual: ARI 3-2Brier 0.214total bias +2

Root cause: Solid correct call

Evidence-backed diagnosis

published system edge 0.07 (lean); this is not the half-scale distance from 50%
team-run absolute error 2; total bias +2
pregame home-minus-away evidence: season -0.026, recent +0.420, pitcher +0.122, composite +0.180
confidence pipeline reached 0.544 before later publication/recency controls; published home probability was 0.463
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
ARI	Eduardo Rodriguez	2 ER / 18 outs	1 ER / 18 outs	-1
HOU	Peter Lambert	2 ER / 16 outs	3 ER / 21 outs	1

Decision internals: run factors home×1.105 / away×0.983; 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:

Jose Altuve (HOU 2B): 3-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 1.25]
Isaac Paredes (HOU 3B): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -1.11]
Ketel Marte (ARI 2B): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.93]
Nolan Arenado (ARI 3B): 1-for-4, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]

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

No settled pitching line for this club. The forecast named Eduardo Rodriguez (ARI) at a projected 3 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 HOME lineup he faced produced 2 runs on 6 hits and 1 HR, against a projected 4 runs / 7 hits (-2 runs vs projection).

Who did the damage to him:

Jose Altuve: 3-for-4, 0 HR, 0 RBI [+1.25 runs vs projection] season xwOBA 0.275, weak - a good day, not a model error
Christian Vázquez: 1-for-3, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.238, weak - a good day, not a model error
Christian Walker: 1-for-3, 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

HOME pitching

Peter Lambert (HOU) threw 21 outs, 3 ER = 3.86 ERA against a projected 3.38 ERA (0.48 ERA error)

Bullpen behind him: 0 ER over 3 outs.

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

Who did the damage to him:

Nolan Arenado: 1-for-4, 1 HR, 2 RBI [+0.93 runs vs projection] season xwOBA 0.283, weak - a good day, not a model error
Corbin Carroll: 2-for-4, 0 HR, 0 RBI [+0.45 runs vs projection] season xwOBA 0.344, elite - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
HOU	Peter Lambert (Starter)	21 / 16	3.86 / 3.38	3 / 2	7 / 2	1 / 0	0 / 1	6 / 6	venue home: 4.04 vs 3 ERA (worse here) · opp run factor 0.98 (average) · 24 prior appearances this season
HOU	Bryan Abreu (Setup Man)	3 / 3	0 / 0	0 / 0	1 / 1	0 / 0	0 / 0	1 / 2

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
HOU	Jose Altuve (2B)	4 / 4	3 / 1	1 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	0 / 1	1.25	0.275 weak	good day
HOU	Isaac Paredes (3B)	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	2 / 0	-1.11	0.319 average	no signal
ARI	Ketel Marte (2B)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0 / 0	1 / 0	-0.93	0.349 elite	bad day
ARI	Nolan Arenado (3B)	4 / 4	1 / 1	0 / 0	0 / 0	1 / 0	2 / 0	1 / 0	0 / 0	1 / 1	0.93	0.283 weak	good day
ARI	Jordan Lawlar (CF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.309 average	—
HOU	Jeremy Peña (SS)	4 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	-0.47	0.325 above average	—
HOU	Christian Vázquez (C)	3 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0.47	0.238 weak	—
HOU	Cam Smith (RF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	3 / 1	-0.47	0.317 average	—
HOU	Daulton Varsho (CF)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	1 / 1	-0.47	0.291 below average	—
ARI	Corbin Carroll (RF)	4 / 5	2 / 1	1 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0 / 1	0.45	0.344 elite	—
ARI	Gabriel Moreno (DH)	4 / 4	1 / 1	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	-0.33	0.365 elite	—
ARI	Geraldo Perdomo (SS)	1 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	-0.33	0.328 above average	—
HOU	Yordan Alvarez (DH)	5 / 4	1 / 1	0 / 0	0 / 0	1 / 1	1 / 1	1 / 1	0 / 1	1 / 1	-0.33	0.447 elite	—
HOU	Christian Walker (1B)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 0	1 / 1	0.33	0.3 below average	—
ARI	Ildemaro Vargas (1B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 0	0	0.3 below average	—
HOU	Nelson Velázquez (LF)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	2 / 1	0	0.338 above average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.026	positive favors home; negative favors away
Recent form	0.42	positive favors home; negative favors away
Starting-pitcher matchup	0.122	positive favors home; negative favors away
Composite	0.18	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.53-4.34	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-3	0.537	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run)
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	4.53-4.34	5-4	0.807	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)
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
ARI	Ildemaro Vargas	batter / 1B	better	predicted_H 2.28
HOU	Nelson Velázquez	batter / LF	better	predicted_AB 1.5
ARI	Geraldo Perdomo	batter / SS	worse	predicted_SO 1.464
HOU	Isaac Paredes	batter / 3B	worse	predicted_SO 1.452
HOU	Yordan Alvarez	batter / DH	worse	predicted_AB 1.429
ARI	Gabriel Moreno	batter / DH	worse	predicted_AB 1.402
ARI	Ketel Marte	batter / 2B	worse	predicted_RBI 1.38
ARI	Nolan Arenado	batter / 3B	worse	predicted_AB -1.29

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	ARI pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	ARI
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_FINALIZED_MATCH_BOX	ARI model pitcher projections differ from final simulation pitching box score (H: current=7, sim=10).	ARI
info	CURRENT_FINALIZED_MATCH_BOX	ARI model-prediction batting totals differ from final simulation box score (AB: current=36, sim=37; H: current=8, sim=7; 3B: current=0, sim=2; HR: current=1, sim=0; BB: current=3, sim=2; SO: current=6, sim=10).	ARI
info	CURRENT_PLAYER_MATCH_BOX	ARI current finalized hitter lines are model predictions and differ from the simulation box score: Gabriel Moreno: AB current=4 box=5, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Corbin Carroll: AB current=5 box=4, 2B current=0 box=1, RBI current=0 box=1, BB current=1 box=0; Tommy Troy: R current=1 box=0, BB current=0 box=1, SO current=1 box=2; Ketel Marte: HR current=1 box=0, RBI current=1 box=2, SO current=0 box=1; Nolan Arenado: AB current=4 box=3, 3B current=0 box=1, R current=0 box=1, BB current=0 box=1; Geraldo Perdomo: AB current=4 box=5, H current=1 box=0, BB current=1 box=0, SO current=0 box=2; James McCann: H current=1 box=0, RBI current=1 box=0; Ildemaro Vargas: H current=1 box=3, 3B current=0 box=1, R current=0 box=1	ARI
info	CURRENT_FINALIZED_MATCH_BOX	HOU model-prediction batting totals differ from final simulation box score (AB: current=34, sim=40; H: current=7, sim=10; 2B: current=1, sim=3; HR: current=1, sim=2; RBI: current=4, sim=3; R: current=4, sim=3; BB: current=3, sim=2; SO: current=8, sim=7).	HOU
info	CURRENT_PLAYER_MATCH_BOX	HOU current finalized hitter lines are model predictions and differ from the simulation box score: Jeremy Peña: AB current=5 box=4, H current=1 box=2, R current=1 box=0, BB current=0 box=1; Isaac Paredes: AB current=4 box=5, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Nelson Velázquez: AB current=3 box=4, H current=0 box=2, 2B current=0 box=1, R current=0 box=1; Jose Altuve: AB current=4 box=5, 2B current=0 box=1, R current=1 box=0, SO current=1 box=0; Yordan Alvarez: AB current=4 box=5, H current=1 box=2, 2B current=0 box=1, HR current=1 box=0; Christian Vázquez: AB current=3 box=4, SO current=1 box=0; Cam Smith: AB current=3 box=4; Daulton Varsho: AB current=4 box=5, HR current=0 box=1, RBI current=0 box=1, R current=0 box=1	HOU
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

ATL@PHIMISS

Predicted: PHI 3-4 (conf +0.10, moderate)Actual: ATL 5-4Brier 0.303total bias -2

Root cause: Backed the wrong side with real conviction

Evidence-backed diagnosis

published system edge 0.10 (moderate); this is not the half-scale distance from 50%
team-run absolute error 2; total bias -2
pregame home-minus-away evidence: season -0.391, recent -0.010, pitcher -0.101, composite -0.260
confidence pipeline reached 0.560 before later publication/recency controls; published home probability was 0.551
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 9 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

committed to PHI at 0.10 confidence and it lost

Team	Starter	Projected	Actual	ER err
ATL	Tyler Mahle	2 ER / 17 outs	0 ER / 18 outs	-2
PHI	Aaron Nola	3 ER / 16 outs	1 ER / 21 outs	-2

Decision internals: run factors home×0.996 / away×1.097; 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:

Matt Olson (ATL 1B): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -1.11]
Bryce Harper (PHI RF): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -1.11]
Michael Harris II (ATL CF): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.93]
Mauricio Dubón (ATL SS): 2-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 0.78]
Bryson Stott (PHI 3B): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.64]

Starter miss: Aaron Nola (PHI) posted 1.29 ERA vs projected 5.06 (1 ER in 21 outs); he is materially worse pitching home (5.72 ERA there vs 4.26 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

Tyler Mahle (ATL) threw 18 outs, 0 ER = 0 ERA against a projected 3.18 ERA (-3.18 ERA error)

Bullpen behind him: 0 ER over 3 outs.

The HOME lineup he faced produced 3 runs on 6 hits and 1 HR, against a projected 4 runs / 8 hits (-1 runs vs projection).

Who did the damage to him:

Garrett Stubbs: 1-for-3, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.212, weak - a good day, not a model error
Alec Bohm: 1-for-4, 0 HR, 1 RBI [+0.31 runs vs projection] season xwOBA 0.322, above average - a good day, not a model error
Trea Turner: 1-for-4, 0 HR, 0 RBI [+0.31 runs vs projection] season xwOBA 0.293, below average - a good day, not a model error

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

HOME pitching

Aaron Nola (PHI) threw 21 outs, 1 ER = 1.29 ERA against a projected 5.06 ERA (-3.77 ERA error)

Bullpen behind him: 1 ER over 5 outs.

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

Who did the damage to him:

Mauricio Dubón: 2-for-4, 0 HR, 0 RBI [+0.78 runs vs projection] season xwOBA 0.302, below average - a good day, not a model error
Austin Riley: 1-for-4, 0 HR, 1 RBI [+0.62 runs vs projection] season xwOBA 0.3, below average - a good day, not a model error
Ronald Acuña Jr.: 2-for-3, 1 HR, 4 RBI [+0.45 runs vs projection] season xwOBA 0.364, elite - 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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
ATL	Tyler Mahle (Starter)	18 / 17	0 / 3.18 (-3.18)	0 / 2	4 / 5	0 / 1	1 / 0	3 / 4	venue away: 6.16 vs 2.4 ERA (worse here) · opp run factor 1 (average) · 23 prior appearances this season
ATL	Dylan Lee (Setup Man)	3 / 3	0 / 9 (-9.00)	0 / 1	0 / 0	0 / 0	0 / 1	1 / 2
PHI	Aaron Nola (Starter)	21 / 16	1.29 / 5.06 (-3.77)	1 / 3	4 / 5	0 / 0	1 / 2	5 / 1	venue home: 5.72 vs 4.26 ERA (worse here) · opp run factor 1.1 (strong) · 28 prior appearances this season
PHI	Jhoan Duran (Closer)	3 / 3	9 / 0 (+9.00)	1 / 0	2 / 0	0 / 0	0 / 0	2 / 1
PHI	Brooks Raley (Middle Reliever (High Leverage))	2 / 2	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 0

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
ATL	Matt Olson (1B)	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	-1.11	0.353 elite	bad day
PHI	Bryce Harper (RF)	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	3 / 1	-1.11	0.392 elite	bad day
ATL	Michael Harris II (CF)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	0 / 1	-0.93	0.342 elite	bad day
ATL	Mauricio Dubón (SS)	4 / 4	2 / 1	1 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 0	0.78	0.302 below average	good day
PHI	Bryson Stott (3B)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	-0.64	0.327 above average	bad day
ATL	Austin Riley (3B)	4 / 4	1 / 1	0 / 0	1 / 0	0 / 0	1 / 1	0 / 0	0 / 0	0 / 1	0.62	0.3 below average	good day
ATL	Sean Murphy (C)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.263 weak	—
PHI	Brandon Marsh (LF)	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.306 average	—
PHI	Luis Arraez (2B)	3 / 5	1 / 2	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.306 average	—
PHI	Garrett Stubbs (C)	3 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0.47	0.212 weak	—
ATL	Ronald Acuña Jr. (RF)	3 / 5	2 / 1	1 / 0	0 / 0	1 / 1	4 / 1	1 / 1	0 / 1	0 / 1	0.45	0.364 elite	—
ATL	Mike Yastrzemski (LF)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	1 / 0	2 / 1	0.33	0.309 average	—
ATL	Ozzie Albies (2B)	4 / 4	1 / 1	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 1	0.31	0.284 below average	—
PHI	Alec Bohm (1B)	4 / 4	1 / 1	1 / 0	0 / 0	0 / 0	1 / 1	0 / 0	0 / 0	1 / 1	0.31	0.322 above average	—
PHI	Trea Turner (SS)	4 / 5	1 / 1	1 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	0 / 1	0.31	0.293 below average	—
ATL	Drake Baldwin (DH)	5 / 4	2 / 1	1 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	2 / 1	0.14	0.365 elite	—
PHI	Kyle Schwarber (DH)	3 / 4	1 / 1	0 / 0	0 / 0	1 / 1	1 / 1	2 / 1	1 / 1	0 / 1	0	0.367 elite	—
PHI	Justin Crawford (CF)	2 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0	0.282 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.391	positive favors home; negative favors away
Recent form	-0.01	positive favors home; negative favors away
Starting-pitcher matchup	-0.101	positive favors home; negative favors away
Composite	-0.26	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.44-4.49	n/a	Base ML historical model prediction
variance_applied	4-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	3-4	0.551	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	4.44-4.49	4-4	-0.055	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
PHI	Trea Turner	batter / SS	worse	predicted_RBI 2.55
ATL	Michael Harris II	batter / CF	better	predicted_AB -1.89
PHI	Brandon Marsh	batter / LF	worse	predicted_H 1.7
ATL	Mauricio Dubón	batter / SS	worse	predicted_SO 1.504
PHI	Justin Crawford	batter / CF	worse	predicted_SO 1.424
ATL	Mike Yastrzemski	batter / LF	boosted	predicted_AB 1.339
ATL	Austin Riley	batter / 3B	worse	predicted_SO -1.237
PHI	Garrett Stubbs	batter / C	worse	predicted_AB 1.167

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	ATL pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	ATL
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	ATL model pitcher projections differ from final simulation pitching box score (H: current=6, sim=9; HR: current=1, sim=2).	ATL
info	CURRENT_FINALIZED_MATCH_BOX	ATL model-prediction batting totals differ from final simulation box score (AB: current=36, sim=33; H: current=9, sim=6; HR: current=2, sim=0; RBI: current=5, sim=3; R: current=4, sim=3; BB: current=3, sim=4; SO: current=8, sim=5).	ATL
info	CURRENT_PLAYER_MATCH_BOX	ATL current finalized hitter lines are model predictions and differ from the simulation box score: Drake Baldwin: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Mauricio Dubón: H current=1 box=0, RBI current=1 box=0, SO current=0 box=2; Michael Harris II: AB current=5 box=3, H current=1 box=3, 2B current=0 box=1, HR current=1 box=0; Ozzie Albies: RBI current=0 box=1, R current=0 box=1, SO current=1 box=0; Matt Olson: AB current=4 box=3, H current=1 box=0, 2B current=1 box=0, R current=1 box=0; Ronald Acuña Jr.: AB current=5 box=4, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Sean Murphy: AB current=3 box=4, H current=1 box=0; Austin Riley: AB current=4 box=3, H current=1 box=0, RBI current=1 box=0, BB current=0 box=1	ATL
info	CURRENT_FINALIZED_MATCH_BOX	PHI model pitcher projections differ from final simulation pitching box score (BB: current=3, sim=4).	PHI
info	CURRENT_FINALIZED_MATCH_BOX	PHI model-prediction batting totals differ from final simulation box score (H: current=8, sim=9; 2B: current=2, sim=1; HR: current=1, sim=2; BB: current=3, sim=1; SO: current=8, sim=7; HBP: current=0, sim=1).	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=2, 2B current=1 box=0, RBI current=1 box=0, BB current=1 box=0; Alec Bohm: H current=1 box=0, RBI current=1 box=0, SO current=1 box=0; Bryson Stott: AB current=4 box=3, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Brandon Marsh: H current=1 box=2, 2B current=0 box=1, R current=0 box=1; Luis Arraez: AB current=5 box=4, H current=2 box=1; Kyle Schwarber: AB current=4 box=5, H current=1 box=2, BB current=1 box=0; Trea Turner: AB current=5 box=4, H current=1 box=2, HR current=0 box=1, RBI current=0 box=3; Garrett Stubbs: AB current=3 box=4, SO current=1 box=0	PHI
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

BOS@BALHIT

Predicted: BOS 4-3 (conf -0.10, lean)Actual: BOS 3-1Brier 0.203total bias +3

Root cause: Solid correct call

Evidence-backed diagnosis

published system edge 0.10 (lean); this is not the half-scale distance from 50%
team-run absolute error 3; total bias +3
pregame home-minus-away evidence: season -0.555, recent -0.178, pitcher -0.066, composite -0.675
confidence pipeline reached 0.568 before later publication/recency controls; published home probability was 0.451
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

total runs over-projected by 3 (7 vs 4)
Kyle Bradish (BAL) allowed 3 fewer ER than projected (0 vs 3)

Team	Starter	Projected	Actual	ER err
BOS	Payton Tolle	0 ER / 18 outs	1 ER / 18 outs	1
BAL	Kyle Bradish	3 ER / 16 outs	0 ER / 12 outs	-3

Decision internals: run factors home×0.967 / away×1.033; 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:

Eli White (BOS CF): 2-for-4, 1 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Nick Sogard (BOS 2B): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -1.11]
Gunnar Henderson (BAL SS): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Colton Cowser (BAL CF): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Pete Alonso (BAL 1B): 2-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.79]

Starter miss: Kyle Bradish (BAL) posted 0.00 ERA vs projected 5.06 (0 ER in 12 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

Payton Tolle (BOS) threw 18 outs, 1 ER = 1.5 ERA against a projected 0 ERA (1.5 ERA error)

Bullpen behind him: 0 ER over 6 outs.

The HOME lineup he faced produced 1 runs on 5 hits and 0 HR, against a projected 4 runs / 7 hits (-3 runs vs projection).

Who did the damage to him:

Rece Hinds: 1-for-3, 0 HR, 0 RBI [+0.78 runs vs projection] season xwOBA 0.158, weak - a good day, not a model error
Coby Mayo: 1-for-4, 0 HR, 0 RBI [+0.31 runs vs projection] season xwOBA 0.317, average - a good day, not a model error

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

HOME pitching

Kyle Bradish (BAL) threw 12 outs, 0 ER = 0 ERA against a projected 5.06 ERA (-5.06 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The AWAY 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:

Eli White: 2-for-4, 1 HR, 3 RBI [+1.4 runs vs projection] season xwOBA 0.275, weak - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
BOS	Payton Tolle (Starter)	18 / 18	1.5 / 0	1 / 0	4 / 4	0 / 0	0 / 0	12 / 7	venue away: 2.59 vs 3.88 ERA (ok) · opp run factor 0.97 (average) · 23 prior appearances this season
BOS	Aroldis Chapman (Closer)	3 / 3	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	2 / 0
BOS	Erik Miller (Setup Man)	3 / 3	0 / 9 (-9.00)	0 / 1	0 / 3	0 / 0	0 / 0	2 / 0
BAL	Kyle Bradish (Starter)	12 / 16	0 / 5.06 (-5.06)	0 / 3	2 / 4	1 / 0	4 / 2	3 / 5	venue home: 3.46 vs 4.45 ERA (ok) · opp run factor 1.03 (average) · 27 prior appearances this season

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
BOS	Eli White (CF)	4 / 3	2 / 1	0 / 0	0 / 0	1 / 0	3 / 1	1 / 0	0 / 0	0 / 1	1.4	0.275 weak	good day
BOS	Nick Sogard (2B)	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	3 / 1	-1.11	0.321 above average	bad day
BAL	Gunnar Henderson (SS)	3 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	-0.8	0.315 average	no signal
BAL	Colton Cowser (CF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	2 / 1	-0.8	0.319 average	no signal
BAL	Pete Alonso (1B)	4 / 5	2 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	2 / 2	-0.79	0.392 elite	bad day
BAL	Rece Hinds (LF)	3 / 3	1 / 0	1 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	2 / 2	0.78	0.158 weak	good day
BOS	Adley Rutschman (C)	4 / 5	1 / 1	1 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	1 / 1	-0.62	0.334 above average	bad day
BOS	Roman Anthony (RF)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	2 / 1	-0.47	0.356 elite	—
BOS	Jarren Duran (LF)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.296 below average	—
BAL	Blaze Alexander (2B)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	2 / 1	-0.47	0.352 elite	—
BAL	Leody Taveras (RF)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.272 weak	—
BOS	Trevor Story (SS)	4 / 5	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 1	-0.31	0.252 weak	—
BAL	Coby Mayo (DH)	4 / 3	1 / 1	1 / 0	0 / 0	0 / 0	0 / 1	1 / 1	0 / 0	1 / 1	0.31	0.317 average	—
BOS	Caleb Durbin (3B)	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	2 / 0	1 / 1	0.19	0.289 below average	—
BOS	Mickey Gasper (1B)	1 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	2 / 1	0 / 0	-0.14	0.377 elite	—
BOS	Andruw Monasterio (DH)	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	2 / 1	0	0.297 below average	—
BAL	Carlos Narváez (C)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.265 weak	—
BAL	Christian Encarnacion-Strand (3B)	4 / 5	1 / 1	1 / 1	0 / 0	0 / 0	1 / 1	0 / 0	0 / 0	3 / 1	0	0.345 elite	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.555	positive favors home; negative favors away
Recent form	-0.178	positive favors home; negative favors away
Starting-pitcher matchup	-0.066	positive favors home; negative favors away
Composite	-0.675	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.8-3.27	n/a	Base ML historical model prediction
variance_applied	5-3	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-3	0.549	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	4.8-3.27	5-3	0.471	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-3	4-3	-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	4-3	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
BOS	Payton Tolle	pitcher / Starter	better	predicted_SO 5
BAL	Pete Alonso	batter / 1B	worse	predicted_H 2.14
BOS	Adley Rutschman	batter / C	worse	predicted_AB -2.06
BOS	Aroldis Chapman	pitcher / Closer	worse	predicted_SO -2
BOS	Erik Miller	pitcher / Setup Man	worse	predicted_SO -2
BAL	Blaze Alexander	batter / 2B	deboosted	predicted_AB 1.846
BOS	Roman Anthony	batter / RF	better	predicted_RBI 1.811
BAL	Carlos Narváez	batter / C	better	predicted_H 1.65

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	BAL pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	BAL
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_FINALIZED_MATCH_BOX	BAL model-prediction batting totals differ from final simulation box score (AB: current=33, sim=37; H: current=7, sim=10; 2B: current=1, 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=0).	BAL
info	CURRENT_PLAYER_MATCH_BOX	BAL current finalized hitter lines are model predictions and differ from the simulation box score: Blaze Alexander: AB current=3 box=5, RBI current=0 box=1, SO current=1 box=0; Coby Mayo: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Rece Hinds: AB current=3 box=4, H current=0 box=1, 2B current=0 box=1, RBI current=1 box=0; Leody Taveras: AB current=3 box=4, 3B current=0 box=1, RBI current=0 box=1, R current=0 box=1; Carlos Narváez: AB current=3 box=4, H current=0 box=2, 2B current=0 box=1, RBI current=0 box=1; Christian Encarnacion-Strand: AB current=5 box=4, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Pete Alonso: AB current=5 box=4, H current=1 box=3, 2B current=0 box=1, HR current=1 box=0; Gunnar Henderson: AB current=5 box=4, 2B current=0 box=1, R current=1 box=0, BB current=1 box=0	BAL
info	CURRENT_FINALIZED_MATCH_BOX	BOS model-prediction batting totals differ from final simulation box score (AB: current=35, sim=33; H: current=8, sim=6; 2B: current=2, sim=4; HR: current=1, sim=0; BB: current=3, sim=2; SO: current=8, sim=5).	BOS
info	CURRENT_PLAYER_MATCH_BOX	BOS current finalized hitter lines are model predictions and differ from the simulation box score: Mickey Gasper: AB current=3 box=4, H current=1 box=0, R current=1 box=0, BB current=1 box=0; Roman Anthony: AB current=4 box=3, 2B current=0 box=1, RBI current=0 box=2; Nick Sogard: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Adley Rutschman: AB current=5 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Trevor Story: AB current=5 box=4, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Andruw Monasterio: AB current=3 box=4, SO current=1 box=2; Caleb Durbin: 2B current=0 box=1, R current=0 box=1, SO current=1 box=0; Eli White: H current=1 box=2, 2B current=0 box=1, R current=0 box=1, SO current=1 box=0	BOS
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

CHC@MIAMISS

Predicted: CHC 6-4 (conf -0.16, moderate)Actual: MIA 3-10Brier 0.336total bias -3

Root cause: Starting-pitcher projection was wrong

Evidence-backed diagnosis

published system edge 0.16 (moderate); this is not the half-scale distance from 50%
team-run absolute error 9; total bias -3
pregame home-minus-away evidence: season -0.841, recent -0.858, pitcher -0.079, composite -1.211
confidence pipeline reached 0.677 before later publication/recency controls; published home probability was 0.420
score-volatility actions: high_total_environment_lift, 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 under-projected by 3 (10 vs 13)
CHC offense over-projected by 3
MIA offense under-projected by 6
Clay Holmes (CHC) allowed 6 more ER than projected (8 vs 2)
committed to CHC at 0.16 confidence and it lost

Team	Starter	Projected	Actual	ER err
CHC	Clay Holmes	2 ER / 18 outs	8 ER / 13 outs	6
MIA	Tyler Phillips	4 ER / 15 outs	3 ER / 15 outs	-1

Decision internals: run factors home×0.976 / away×1.055; 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:

Javier Sanoja (MIA 2B): 2-for-4, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 1.09]
Michael Busch (CHC 1B): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Heriberto Hernández (MIA LF): 2-for-4, 1 HR, 2 RBI (projected 1.0 H, 1.0 HR) [run impact 0.8]
Graham Pauley (MIA 3B): 0-for-2, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 0.66]
Michael Conforto (CHC DH): 1-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.6]

Starter miss: Clay Holmes (CHC) posted 16.62 ERA vs projected 3.00 (8 ER in 13 outs); he is materially worse pitching away (2.42 ERA there vs 1.77 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

Clay Holmes (CHC) threw 13 outs, 8 ER = 16.62 ERA against a projected 3 ERA (13.62 ERA error)

Bullpen behind him: 2 ER over 8 outs.

The HOME lineup he faced produced 10 runs on 11 hits and 2 HR, against a projected 5 runs / 8 hits (5 runs vs projection).

Who did the damage to him:

Javier Sanoja: 2-for-4, 1 HR, 2 RBI [+1.09 runs vs projection] season xwOBA 0.288, below average - a good day, not a model error
Heriberto Hernández: 2-for-4, 1 HR, 2 RBI [+0.8 runs vs projection] season xwOBA 0.355, elite - real projection error
Graham Pauley: 0-for-2, 0 HR, 0 RBI [+0.66 runs vs projection] season xwOBA 0.231, weak - a good day, not a model error

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

HOME pitching

Tyler Phillips (MIA) threw 15 outs, 3 ER = 5.4 ERA against a projected 7.2 ERA (-1.8 ERA error)

Bullpen behind him: 0 ER over 3 outs.

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

No individual hitter beat his projection by a meaningful margin.

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
CHC	Clay Holmes (Starter)	13 / 18	16.62 / 3 (+13.62)	8 / 2	6 / 2	1 / 1	3 / 3	2 / 7	venue away: 2.42 vs 1.77 ERA (worse here) · opp run factor 0.98 (average) · 14 prior appearances this season
CHC	Colin Rea (Setup Man)	8 / 4	6.75 / 6.75	2 / 1	5 / 1	1 / 0	2 / 0	2 / 0
MIA	Tyler Phillips (Starter)	15 / 15	5.4 / 7.2	3 / 4	5 / 3	1 / 2	1 / 5	3 / 3	venue home: 2.72 vs 5.82 ERA (ok) · opp run factor 1.05 (strong) · 33 prior appearances this season
MIA	Victor Vodnik (Middle Reliever (High Leverage))	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 2

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
MIA	Javier Sanoja (2B)	4 / 3	2 / 1	0 / 1	0 / 0	1 / 0	2 / 0	2 / 0	0 / 0	0 / 0	1.09	0.288 below average	good day
CHC	Michael Busch (1B)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	-0.8	0.328 above average	bad day
MIA	Heriberto Hernández (LF)	4 / 4	2 / 1	0 / 0	0 / 0	1 / 1	2 / 1	2 / 1	1 / 0	0 / 1	0.8	0.355 elite	projection too low
MIA	Graham Pauley (3B)	2 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	2 / 0	0 / 1	0.66	0.231 weak	good day
CHC	Michael Conforto (DH)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 0	0 / 1	-0.6	0.335 above average	bad day
CHC	Pedro Ramirez (2B)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.289 below average	—
MIA	Jakob Marsee (CF)	4 / 4	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	1 / 1	0 / 1	0.47	0.307 average	—
MIA	Otto Lopez (SS)	4 / 4	2 / 1	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	0 / 0	0 / 1	0.47	0.324 above average	—
CHC	Alex Bregman (3B)	2 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 0	0 / 1	-0.45	0.318 average	—
MIA	Kyle Stowers (DH)	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	2 / 1	2 / 2	-0.45	0.336 above average	—
CHC	Pete Crow-Armstrong (CF)	3 / 6	1 / 1	0 / 0	0 / 0	1 / 1	2 / 1	1 / 1	0 / 1	0 / 1	-0.33	0.371 elite	—
MIA	Griffin Conine (1B)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	1 / 1	1 / 1	0 / 1	0 / 1	-0.33	0.37 elite	—
CHC	Seiya Suzuki (RF)	3 / 5	1 / 1	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	1 / 1	0	0.339 elite	—
CHC	Nico Hoerner (SS)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0	0.324 above average	—
CHC	Carson Kelly (C)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	0 / 0	2 / 1	0	0.31 average	—
MIA	Owen Caissie (RF)	5 / 3	1 / 1	0 / 0	0 / 0	0 / 0	2 / 0	1 / 0	0 / 0	3 / 1	0	0.294 below average	—
MIA	Joe Mack (C)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	2 / 0	0 / 0	0 / 0	0 / 1	0	0.29 below average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.841	positive favors home; negative favors away
Recent form	-0.858	positive favors home; negative favors away
Starting-pitcher matchup	-0.079	positive favors home; negative favors away
Composite	-1.211	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.17-3.69	n/a	Base ML historical model prediction
variance_applied	4-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-4	0.58	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	4.17-3.69	4-4	-0.476	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-4	6-4	2	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
CHC	Seiya Suzuki	batter / RF	better	predicted_AB -2.73
MIA	Owen Caissie	batter / RF	better	predicted_RBI 2.62
CHC	Michael Busch	batter / 1B	better	predicted_RBI 2.44
CHC	Pedro Ramirez	batter / 2B	worse	predicted_SO 2.417
CHC	Alex Bregman	batter / 3B	worse	predicted_AB -2.21
MIA	Kyle Stowers	batter / DH	worse	predicted_AB -1.695
CHC	Michael Conforto	batter / DH	worse	predicted_AB 1.5
CHC	Pete Crow-Armstrong	batter / CF	worse	predicted_H -1.39

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	CHC pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	CHC
info	CURRENT_PITCHER_DUPLICATION	MIA pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	MIA
info	CURRENT_FINALIZED_MATCH_BOX	CHC model-prediction batting totals differ from final simulation box score (AB: current=37, sim=30; H: current=9, sim=4; 2B: current=2, sim=0; HR: current=2, sim=3; BB: current=4, sim=6; SO: current=8, sim=9; SF: current=0, sim=1).	CHC
info	CURRENT_PLAYER_MATCH_BOX	CHC current finalized hitter lines are model predictions and differ from the simulation box score: Michael Busch: AB current=4 box=3, H current=1 box=2, HR current=0 box=2, RBI current=0 box=3; Seiya Suzuki: AB current=5 box=2, 2B current=1 box=0, HR current=0 box=1, RBI current=1 box=2; Nico Hoerner: H current=1 box=0, SO current=0 box=1; Alex Bregman: AB current=4 box=2, H current=1 box=0, 2B current=1 box=0, BB current=0 box=1; Pete Crow-Armstrong: AB current=6 box=5, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Ian Happ: AB current=4 box=3, H current=1 box=0, RBI current=1 box=0; Pedro Ramirez: AB current=3 box=4, H current=1 box=0, SO current=1 box=3; Carson Kelly: AB current=4 box=3, RBI current=1 box=0, BB current=0 box=1, SO current=1 box=0	CHC
info	CURRENT_FINALIZED_MATCH_BOX	MIA model pitcher projections differ from final simulation pitching box score (H: current=5, sim=4).	MIA
info	CURRENT_FINALIZED_MATCH_BOX	MIA model-prediction batting totals differ from final simulation box score (H: current=8, sim=4; 2B: current=2, sim=0; HR: current=1, sim=2; R: current=5, sim=4; SO: current=9, sim=7).	MIA
info	CURRENT_PLAYER_MATCH_BOX	MIA current finalized hitter lines are model predictions and differ from the simulation box score: Kyle Stowers: AB current=4 box=2, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Griffin Conine: AB current=3 box=4, RBI current=1 box=0, BB current=1 box=0, SO current=1 box=2; Owen Caissie: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=3, R current=0 box=1; Joe Mack: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=1, R current=0 box=1; Heriberto Hernández: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Javier Sanoja: H current=1 box=0, 2B current=1 box=0; Jakob Marsee: AB current=4 box=3, H current=1 box=0, R current=1 box=0; Graham Pauley: RBI current=1 box=0	MIA
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

DET@CLEHIT

Predicted: CLE 3-4 (conf +0.04, lean)Actual: CLE 2-3Brier 0.23total bias +2

Root cause: Solid correct call

Evidence-backed diagnosis

published system edge 0.04 (lean); this is not the half-scale distance from 50%
team-run absolute error 2; total bias +2
pregame home-minus-away evidence: season -0.130, recent +0.434, pitcher +0.005, composite +0.229
confidence pipeline reached 0.545 before later publication/recency controls; published home probability was 0.520
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

No large single-factor errors.

Team	Starter	Projected	Actual	ER err
DET	Jackson Jobe	1 ER / 15 outs	2 ER / 15 outs	1
CLE	Gavin Williams	2 ER / 16 outs	2 ER / 18 outs	0

Decision internals: run factors home×0.964 / 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:

Angel Martínez (CLE LF): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.07]
Brayan Rocchio (CLE SS): 1-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Max Clark (DET CF): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.78]
John Peck (DET SS): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.78]
Steven Kwan (CLE DH): 2-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 0.76]

Pitcher versus batter: what actually broke this game

AWAY pitching

Jackson Jobe (DET) threw 15 outs, 2 ER = 3.6 ERA against a projected 1.8 ERA (1.8 ERA error) only 5 prior appearances in the covered season - the projection had little of his own record to stand on

Bullpen behind him: 0 ER over 6 outs.

The HOME lineup he faced produced 3 runs on 6 hits and 1 HR, against a projected 4 runs / 7 hits (-1 runs vs projection).

Who did the damage to him:

Brayan Rocchio: 1-for-3, 1 HR, 1 RBI [+0.93 runs vs projection] season xwOBA 0.291, below average - a good day, not a model error
Steven Kwan: 2-for-4, 0 HR, 0 RBI [+0.76 runs vs projection] season xwOBA 0.322, above average - real projection error
Nathaniel Lowe: 0-for-2, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.335, above average - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

HOME pitching

Gavin Williams (CLE) threw 18 outs, 2 ER = 3 ERA against a projected 3.38 ERA (-0.38 ERA error)

Bullpen behind him: 0 ER over 12 outs.

The AWAY lineup he faced produced 2 runs on 4 hits and 0 HR, against a projected 4 runs / 8 hits (-2 runs vs projection).

No individual hitter beat his projection by a meaningful margin.

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
DET	Jackson Jobe (Starter)	15 / 15	3.6 / 1.8	2 / 1	4 / 2	1 / 1	2 / 1	7 / 3	venue away: 2.94 vs 7.88 ERA (ok) · opp run factor 0.96 (weak) · 5 prior appearances this season
DET	Tyler Holton (Setup Man)	3 / 4	0 / 20.25 (-20.25)	0 / 3	0 / 3	0 / 1	0 / 0	0 / 0
DET	Kenley Jansen (Closer)	3 / 5	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1
CLE	Gavin Williams (Starter)	18 / 16	3 / 3.38	2 / 2	4 / 2	0 / 0	3 / 1	11 / 9	venue home: 3.01 vs 4.79 ERA (ok) · opp run factor 0.96 (average) · 28 prior appearances this season
CLE	Cade Smith (Closer)	6 / 4	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	5 / 1
CLE	Erik Sabrowski (Middle Reliever (High Leverage))	3 / 4	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1
CLE	Hunter Gaddis (Setup Man)	3 / 3	0 / 9 (-9.00)	0 / 1	0 / 1	0 / 1	1 / 0	0 / 1

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
CLE	Angel Martínez (LF)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 0	1 / 1	-1.07	0.303 below average	projection too high
CLE	Brayan Rocchio (SS)	3 / 4	1 / 1	0 / 0	0 / 0	1 / 0	1 / 1	1 / 1	0 / 0	1 / 1	0.93	0.291 below average	good day
DET	Max Clark (CF)	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	3 / 1	-0.78	0.311 average	no signal
DET	John Peck (SS)	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	2 / 1	-0.78	0.325 above average	bad day
CLE	Steven Kwan (DH)	4 / 4	2 / 1	2 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	1 / 0	0.76	0.322 above average	projection too low
DET	Ben Malgeri (RF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	2 / 1	-0.47	0.322 above average	—
CLE	Jo Adell (RF)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	2 / 1	-0.47	0.324 above average	—
CLE	Angel Genao (3B)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	1 / 0	-0.47	0.29 below average	—
CLE	Petey Halpin (CF)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.283 weak	—
CLE	Nathaniel Lowe (1B)	2 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0.33	0.335 above average	—
DET	Riley Greene (DH)	4 / 4	2 / 1	2 / 0	0 / 0	0 / 1	2 / 1	0 / 1	0 / 1	1 / 1	-0.17	0.369 elite	—
DET	Kevin McGonigle (3B)	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	2 / 1	1 / 0	-0.14	0.35 elite	—
DET	Hao-Yu Lee (2B)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	2 / 1	-0.14	0.299 below average	—
CLE	Travis Bazzana (2B)	4 / 4	2 / 1	1 / 1	0 / 0	0 / 0	1 / 1	0 / 0	0 / 1	0 / 1	0.14	0.299 below average	—
CLE	Patrick Bailey (C)	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	0.14	0.308 average	—
DET	Colt Keith (1B)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	2 / 1	0	0.335 above average	—
DET	Brett Callahan (LF)	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	4 / 1	0	0.31 average	—
DET	Dillon Dingler (C)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	0 / 1	0	0.343 elite	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.13	positive favors home; negative favors away
Recent form	0.434	positive favors home; negative favors away
Starting-pitcher matchup	0.005	positive favors home; negative favors away
Composite	0.229	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.62-5.23	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	3-4	0.52	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	3.62-5.23	4-5	-0.609	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
DET	Brett Callahan	batter / LF	worse	predicted_SO 3.143
DET	Dillon Dingler	batter / C	worse	predicted_AB -1.99
CLE	Brayan Rocchio	batter / SS	worse	predicted_AB -1.68
CLE	Angel Martínez	batter / LF	worse	predicted_AB -1.598
DET	Max Clark	batter / CF	worse	predicted_RBI 1.44
CLE	Travis Bazzana	batter / 2B	better	predicted_RBI 1.42
DET	Colt Keith	batter / 1B	worse	predicted_SO 1.397
DET	Kevin McGonigle	batter / 3B	worse	predicted_SO 1.39

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	CLE pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	CLE
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	CLE model-prediction batting totals differ from final simulation box score (AB: current=33, sim=28; H: current=7, sim=5; 2B: current=1, sim=0; HR: current=1, sim=2; BB: current=3, sim=1; SO: current=7, sim=4; SF: current=0, sim=1).	CLE
info	CURRENT_PLAYER_MATCH_BOX	CLE current finalized hitter lines are model predictions and differ from the simulation box score: Steven Kwan: H current=1 box=0, R current=1 box=0, BB current=1 box=0; Angel Martínez: AB current=4 box=2, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Travis Bazzana: H current=1 box=2, 2B current=1 box=0, HR current=0 box=1, RBI current=1 box=2; Nathaniel Lowe: H current=0 box=1, HR current=0 box=1, RBI current=0 box=1, R current=0 box=1; Jo Adell: H current=1 box=0, RBI current=1 box=0; Patrick Bailey: H current=0 box=1, R current=0 box=1, BB current=1 box=0; Angel Genao: AB current=4 box=3, H current=1 box=0, R current=1 box=0, SO current=0 box=1; Petey Halpin: SO current=1 box=0	CLE
info	CURRENT_FINALIZED_MATCH_BOX	DET model-prediction batting totals differ from final simulation box score (AB: current=33, sim=31; H: current=8, sim=4; 2B: current=2, sim=1; 3B: current=0, sim=1; RBI: current=4, sim=3; R: current=4, sim=3; BB: current=3, sim=2; SO: current=8, sim=12).	DET
info	CURRENT_PLAYER_MATCH_BOX	DET current finalized hitter lines are model predictions and differ from the simulation box score: Colt Keith: AB current=3 box=4, H current=1 box=0, SO current=1 box=2; Riley Greene: 2B current=0 box=1, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Kevin McGonigle: HR current=0 box=1, RBI current=0 box=1, BB current=1 box=0, SO current=0 box=2; Ben Malgeri: AB current=3 box=2, H current=1 box=0, SO current=1 box=0; Brett Callahan: AB current=3 box=4, SO current=1 box=4; Max Clark: AB current=4 box=3, 2B current=1 box=0, 3B current=0 box=1, RBI current=1 box=2; Dillon Dingler: AB current=5 box=3, RBI current=1 box=0, BB current=0 box=1; Hao-Yu Lee: AB current=3 box=4, H current=1 box=0	DET
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

LAA@PITHIT

Predicted: PIT 3-5 (conf +0.14, moderate)Actual: PIT 0-1Brier 0.183total bias +7

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.14 (moderate); this is not the half-scale distance from 50%
team-run absolute error 7; total bias +7
pregame home-minus-away evidence: season +0.719, recent +0.432, pitcher -0.131, composite +0.575
confidence pipeline reached 0.586 before later publication/recency controls; published home probability was 0.572
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 7 (8 vs 1)
LAA offense over-projected by 3
PIT offense over-projected by 4
Walbert Ureña (LAA) allowed 3 fewer ER than projected (1 vs 4)

Team	Starter	Projected	Actual	ER err
LAA	Walbert Ureña	4 ER / 17 outs	1 ER / 24 outs	-3
PIT	Paul Skenes	0 ER / 17 outs	0 ER / 16 outs	0

Decision internals: run factors home×0.958 / away×1.089; 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:

Bryan Reynolds (PIT LF): 0-for-2, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Zach Neto (LAA SS): 1-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.26]
Henry Davis (PIT C): 1-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Oneil Cruz (PIT CF): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Esmerlyn Valdez (PIT RF): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]

Starter miss: Walbert Ureña (LAA) posted 1.12 ERA vs projected 6.35 (1 ER in 24 outs); he is materially worse pitching away (3.71 ERA there vs 2.42 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

Walbert Ureña (LAA) threw 24 outs, 1 ER = 1.12 ERA against a projected 6.35 ERA (-5.23 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 1 runs on 3 hits and 1 HR, against a projected 5 runs / 8 hits (-4 runs vs projection).

Who did the damage to him:

Henry Davis: 1-for-3, 1 HR, 1 RBI [+0.93 runs vs projection] season xwOBA 0.288, below average - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

HOME pitching

Paul Skenes (PIT) threw 16 outs, 0 ER = 0 ERA against a projected 0 ERA (0 ERA error)

Bullpen behind him: 0 ER over 5 outs.

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

Who did the damage to him:

Josh Lowe: 0-for-1, 0 HR, 0 RBI [+0.66 runs vs projection] season xwOBA 0.291, below average - a good day, not a model error
Mike Trout: 2-for-4, 0 HR, 0 RBI [+0.45 runs vs projection] season xwOBA 0.375, elite - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
LAA	Walbert Ureña (Starter)	24 / 17	1.12 / 6.35 (-5.23)	1 / 4	3 / 2	1 / 0	1 / 1	4 / 8	venue away: 3.71 vs 2.42 ERA (worse here) · opp run factor 0.96 (strong) · 25 prior appearances this season
PIT	Paul Skenes (Starter)	16 / 17	0 / 0	0 / 0	4 / 2	0 / 0	1 / 3	2 / 2	venue home: 3.94 vs 4.01 ERA (ok) · opp run factor 1.09 (weak) · 28 prior appearances this season
PIT	Gregory Soto (Middle Reliever (High Leverage))	2 / 3	0 / 9 (-9.00)	0 / 1	0 / 1	0 / 1	0 / 0	0 / 2
PIT	Mason Montgomery (Closer)	3 / 3	0 / 9 (-9.00)	0 / 1	0 / 1	0 / 1	1 / 0	1 / 2

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
PIT	Bryan Reynolds (LF)	2 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	1 / 1	1 / 1	-1.4	0.36 elite	bad day
LAA	Zach Neto (SS)	3 / 5	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	1 / 2	-1.26	0.305 average	no signal
PIT	Henry Davis (C)	3 / 3	1 / 1	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 0	0 / 1	0.93	0.288 below average	good day
PIT	Oneil Cruz (CF)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	-0.8	0.35 elite	bad day
PIT	Esmerlyn Valdez (RF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	-0.8	0.328 above average	bad day
LAA	Vaughn Grissom (3B)	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	0 / 1	-0.78	0.314 average	no signal
LAA	Josh Lowe (CF)	1 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	1 / 1	0.66	0.291 below average	good day
LAA	Moisés Ballesteros (1B)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.289 below average	—
LAA	Adam Frazier (LF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.257 weak	—
PIT	Nick Gonzales (DH)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 1	-0.47	0.327 above average	—
PIT	Konnor Griffin (SS)	3 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 1	-0.47	0.297 below average	—
PIT	Jacob Gonzalez (3B)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.309 average	—
LAA	Mike Trout (DH)	4 / 4	2 / 1	1 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	0.45	0.375 elite	—
LAA	Wade Meckler (RF)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	-0.33	0.311 average	—
PIT	Brandon Lowe (2B)	3 / 4	2 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	-0.17	0.323 above average	—
LAA	Tyler Heineman (C)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0	0.243 weak	—
LAA	Christian Moore (2B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 1	0	0.249 weak	—
PIT	Spencer Horwitz (1B)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0	0.311 average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.719	positive favors home; negative favors away
Recent form	0.432	positive favors home; negative favors away
Starting-pitcher matchup	-0.131	positive favors home; negative favors away
Composite	0.575	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.37-5.51	n/a	Base ML historical model prediction
variance_applied	4-6	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.572	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by season_quality_guardrail)
embedded_savant_run_environment	3-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-5	n/a	Capped at max 12 realistic runs
simulation_result	3-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.37-5.51	4-6	0.867	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 season_quality_guardrail)
embedded_savant_run_environment	4-5	3-5	1	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-5	3-5	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-5	3-5	0	Capped at max 12 realistic runs
simulation_result	3-5	3-5	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
PIT	Oneil Cruz	batter / CF	worse	predicted_AB -2.47
PIT	Spencer Horwitz	batter / 1B	worse	predicted_RBI 1.79
PIT	Bryan Reynolds	batter / LF	better	predicted_2B 1.764
PIT	Henry Davis	batter / C	worse	predicted_RBI 1.63
PIT	Esmerlyn Valdez	batter / RF	worse	predicted_SO -1.523
LAA	Adam Frazier	batter / LF	worse	predicted_AB 1.5
PIT	Jacob Gonzalez	batter / 3B	better	predicted_H 1.418
LAA	Zach Neto	batter / SS	better	predicted_SO -1.387

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	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	LAA model pitcher projections differ from final simulation pitching box score (H: current=3, sim=7; BB: current=3, sim=5).	LAA
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; HR: current=1, sim=2; RBI: current=4, sim=3; R: current=4, sim=3; BB: current=3, sim=5; SO: current=9, sim=8; HBP: current=0, sim=1).	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=0, RBI current=1 box=0, R current=1 box=0, SO current=1 box=2; Wade Meckler: AB current=3 box=4, H current=1 box=0, R current=1 box=0, BB current=1 box=0; Vaughn Grissom: AB current=4 box=3, 2B current=1 box=0, BB current=0 box=1; Moisés Ballesteros: H current=1 box=0, BB current=0 box=1, SO current=1 box=0; Josh Lowe: AB current=3 box=4, H current=0 box=1, HR current=0 box=1, RBI current=0 box=1; Adam Frazier: AB current=3 box=4; Zach Neto: AB current=5 box=4, H current=1 box=2, 2B current=0 box=1, SO current=2 box=0; Tyler Heineman: AB current=3 box=4, H current=0 box=1, SO current=0 box=1	LAA
info	CURRENT_FINALIZED_MATCH_BOX	PIT model-prediction batting totals differ from final simulation box score (AB: current=33, sim=31; H: current=8, sim=7; 2B: current=1, sim=3; HR: current=1, sim=0; RBI: current=4, sim=5; BB: current=4, sim=5; HBP: current=0, sim=1).	PIT
info	CURRENT_PLAYER_MATCH_BOX	PIT current finalized hitter lines are model predictions and differ from the simulation box score: Nick Gonzales: H current=1 box=0, RBI current=1 box=0, SO current=1 box=2; Konnor Griffin: AB current=5 box=4, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Oneil Cruz: AB current=4 box=2, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Esmerlyn Valdez: RBI current=1 box=0, SO current=1 box=0; Jacob Gonzalez: AB current=3 box=4, H current=1 box=2, 2B current=0 box=1, RBI current=0 box=1; Bryan Reynolds: H current=1 box=2, 2B current=0 box=2, HR current=1 box=0, BB current=1 box=0; Brandon Lowe: H current=1 box=0, 2B current=1 box=0, R current=1 box=0, BB current=1 box=0; Henry Davis: RBI current=0 box=2, R current=0 box=1, BB current=0 box=1	PIT
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

MIL@CINMISS

Predicted: MIL 7-5 (conf -0.06, lean)Actual: CIN 8-12Brier 0.281total bias -8

Root cause: Run environment far higher than projected (under-dispersed)

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 8; total bias -8
pregame home-minus-away evidence: season -1.480, recent -0.258, pitcher -0.023, composite -1.288
confidence pipeline reached 0.626 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 under-projected by 8 (12 vs 20)
CIN offense under-projected by 7

Team	Starter	Projected	Actual	ER err
MIL	Kyle Harrison	2 ER / 15 outs	4 ER / 11 outs	2
CIN	Brady Singer	5 ER / 15 outs	6 ER / 15 outs	1

Decision internals: run factors home×1.016 / away×1.039; 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:

Jake Bauers (MIL 1B): 1-for-2, 1 HR, 2 RBI (projected 0.0 H, 0.0 HR) [run impact 2.06]
Juan Brito (CIN 2B): 3-for-4, 0 HR, 1 RBI (projected 0.0 H, 0.0 HR) [run impact 2.05]
Elly De La Cruz (CIN DH): 2-for-4, 1 HR, 4 RBI (projected 1.0 H, 0.0 HR) [run impact 1.73]
Dane Myers (CIN CF): 2-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 1.44]
Brice Turang (MIL 2B): 2-for-3, 0 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 1.11]

Starter miss: Kyle Harrison (MIL) posted 9.82 ERA vs projected 3.60 (4 ER in 11 outs); he is materially worse pitching away (5.94 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 11 outs, 4 ER = 9.82 ERA against a projected 3.6 ERA (6.22 ERA error)

Bullpen behind him: 3 ER over 3 outs.

The HOME lineup he faced produced 12 runs on 14 hits and 1 HR, against a projected 5 runs / 8 hits (7 runs vs projection).

Who did the damage to him:

Juan Brito: 3-for-4, 0 HR, 1 RBI [+2.05 runs vs projection] season xwOBA 0.277, weak - a good day, not a model error
Elly De La Cruz: 2-for-4, 1 HR, 4 RBI [+1.73 runs vs projection] season xwOBA 0.368, elite - real projection error
Dane Myers: 2-for-4, 0 HR, 0 RBI [+1.44 runs vs projection] season xwOBA 0.314, average - a good day, not a model error

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

HOME pitching

Brady Singer (CIN) threw 15 outs, 6 ER = 10.8 ERA against a projected 9 ERA (1.8 ERA error)

Bullpen behind him: 0 ER over 7 outs.

The AWAY lineup he faced produced 8 runs on 7 hits and 1 HR, against a projected 5 runs / 8 hits (3 runs vs projection).

Who did the damage to him:

Jake Bauers: 1-for-2, 1 HR, 2 RBI [+2.06 runs vs projection] season xwOBA 0.377, elite - real projection error
Brice Turang: 2-for-3, 0 HR, 3 RBI [+1.11 runs vs projection] season xwOBA 0.336, above average - real projection error
Garrett Mitchell: 2-for-3, 0 HR, 1 RBI [+1.09 runs vs projection] season xwOBA 0.35, elite - real projection 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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
MIL	Kyle Harrison (Starter)	11 / 15	9.82 / 3.6 (+6.22)	4 / 2	9 / 3	1 / 1	4 / 3	4 / 5	venue away: 5.94 vs 1.26 ERA (worse here) · opp run factor 1.02 (weak) · 22 prior appearances this season
MIL	Abner Uribe (Setup Man)	1 / 3	0 / 18 (-18.00)	0 / 2	0 / 3	0 / 0	0 / 1	1 / 2
MIL	Antonio Senzatela (Middle Reliever (High Leverage))	2 / 3	40.5 / 9 (+31.50)	3 / 1	2 / 4	0 / 0	2 / 0	1 / 0
CIN	Brady Singer (Starter)	15 / 15	10.8 / 9	6 / 5	4 / 6	1 / 0	6 / 2	3 / 5	venue home: 4.08 vs 5.67 ERA (ok) · opp run factor 1.04 (strong) · 27 prior appearances this season
CIN	Luis Mey (Setup Man)	3 / 4	0 / 0	0 / 0	0 / 2	0 / 0	0 / 0	1 / 1
CIN	Brock Burke (Middle Reliever (High Leverage))	3 / 3	0 / 18 (-18.00)	0 / 2	1 / 1	0 / 1	0 / 1	0 / 2
CIN	Emilio Pagán (Closer)	1 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 2

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
MIL	Jake Bauers (1B)	2 / 3	1 / 0	0 / 0	0 / 0	1 / 0	2 / 0	3 / 1	3 / 1	0 / 1	2.06	0.377 elite	projection too low
CIN	Juan Brito (2B)	4 / 3	3 / 0	1 / 0	0 / 0	0 / 0	1 / 0	3 / 0	1 / 0	0 / 1	2.05	0.277 weak	good day
CIN	Elly De La Cruz (DH)	4 / 4	2 / 1	1 / 1	0 / 0	1 / 0	4 / 0	1 / 1	2 / 1	1 / 1	1.73	0.368 elite	projection too low
CIN	Dane Myers (CF)	4 / 3	2 / 1	1 / 0	0 / 0	0 / 0	0 / 0	3 / 0	2 / 0	0 / 1	1.44	0.314 average	no signal
MIL	Brice Turang (2B)	3 / 4	2 / 1	1 / 0	0 / 0	0 / 0	3 / 1	1 / 1	2 / 1	0 / 1	1.11	0.336 above average	projection too low
MIL	Garrett Mitchell (CF)	3 / 4	2 / 1	0 / 0	1 / 0	0 / 0	1 / 0	2 / 0	0 / 0	0 / 1	1.09	0.35 elite	projection too low
MIL	Jackson Chourio (LF)	4 / 5	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	2 / 1	1 / 0	1 / 1	-1.07	0.333 above average	bad day
CIN	JJ Bleday (LF)	5 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	-0.8	0.346 elite	bad day
MIL	Sal Frelick (RF)	5 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	-0.47	0.283 weak	—
CIN	Tyler Stephenson (C)	5 / 4	2 / 1	0 / 0	0 / 0	0 / 1	2 / 1	0 / 1	1 / 1	2 / 1	-0.46	0.335 above average	—
CIN	Matt McLain (SS)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	1 / 0	2 / 1	0.33	0.307 average	—
MIL	Christian Yelich (DH)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	-0.31	0.317 average	—
CIN	Sal Stewart (1B)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	1 / 1	0 / 1	1 / 1	0 / 1	-0.31	0.351 elite	—
CIN	Eugenio Suárez (3B)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	2 / 0	2 / 1	0.19	0.29 below average	—
MIL	Cooper Pratt (SS)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	1 / 0	0 / 1	-0.14	0.313 average	—
MIL	David Hamilton (3B)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	-0.14	0.276 weak	—
CIN	Héctor Rodríguez (RF)	5 / 4	3 / 1	0 / 0	0 / 0	0 / 1	1 / 0	2 / 1	0 / 0	1 / 2	0.01	0.277 weak	—
MIL	Bo Naylor (C)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	0 / 0	2 / 1	0	0.296 below average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-1.48	positive favors home; negative favors away
Recent form	-0.258	positive favors home; negative favors away
Starting-pitcher matchup	-0.023	positive favors home; negative favors away
Composite	-1.288	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.98-3.08	n/a	Base ML historical model prediction
variance_applied	5-3	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	7-5	0.53	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	7-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	7-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	7-5	n/a	Capped at max 12 realistic runs
simulation_result	7-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.98-3.08	5-3	0.106	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-3	7-5	-0	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	7-5	7-5	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	7-5	7-5	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	7-5	7-5	-0	Capped at max 12 realistic runs
simulation_result	7-5	7-5	-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
CIN	Dane Myers	batter / CF	better	predicted_H 2.4
CIN	Elly De La Cruz	batter / DH	better	predicted_H 1.892
MIL	Brice Turang	batter / 2B	worse	predicted_SO 1.86
MIL	Garrett Mitchell	batter / CF	better	predicted_RBI 1.69
MIL	Jake Bauers	batter / 1B	deboosted	predicted_H 1.56
MIL	Bo Naylor	batter / C	better	predicted_RBI 1.5
MIL	Cooper Pratt	batter / SS	worse	predicted_SO 1.302
MIL	David Hamilton	batter / 3B	better	predicted_SO 1.26

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	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_FINALIZED_MATCH_BOX	CIN model-prediction batting totals differ from final simulation box score (AB: current=32, sim=37; H: current=8, sim=10; 2B: current=2, sim=4; HR: current=2, sim=1; RBI: current=4, sim=5; BB: current=4, sim=5; SO: current=10, sim=8; 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: H current=1 box=3, 2B current=1 box=2, RBI current=0 box=1, SO current=1 box=0; Eugenio Suárez: AB current=3 box=4, RBI current=1 box=0, R current=0 box=1, BB current=0 box=1; Tyler Stephenson: H current=1 box=2, 2B current=0 box=1, HR current=1 box=0; Dane Myers: AB current=3 box=4, H current=1 box=3, 2B current=0 box=1, HR current=0 box=1; Matt McLain: AB current=3 box=4, H current=1 box=0, BB current=0 box=1; Juan Brito: AB current=3 box=4; Héctor Rodríguez: H current=1 box=0, HR current=1 box=0, R current=1 box=0, SO current=2 box=0; JJ Bleday: BB current=1 box=0, HBP current=0 box=1	CIN
info	CURRENT_FINALIZED_MATCH_BOX	MIL model-prediction batting totals differ from final simulation box score (AB: current=34, sim=36; H: current=8, sim=9; 2B: current=1, sim=2; 3B: current=0, sim=1; RBI: current=4, sim=7; R: current=5, sim=7; 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: Christian Yelich: AB current=4 box=5, 2B current=1 box=0, BB current=1 box=0; Jackson Chourio: H current=1 box=2, 2B current=0 box=1, HR current=1 box=0, SO current=1 box=2; Sal Frelick: RBI current=0 box=1, R current=0 box=1, BB current=0 box=1; Jake Bauers: AB current=3 box=4, H current=0 box=2, 3B current=0 box=1, RBI current=0 box=1; Garrett Mitchell: RBI current=0 box=2, R current=0 box=1, SO current=1 box=0; Brice Turang: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Bo Naylor: AB current=4 box=3, HR current=0 box=1, RBI current=1 box=2, R current=0 box=1; Cooper Pratt: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=0 box=1	MIL
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

MIN@CHWHIT

Predicted: CHW 4-6 (conf +0.06, lean)Actual: CHW 1-10Brier 0.221total bias -1

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 7; total bias -1
pregame home-minus-away evidence: season +0.372, recent -0.018, pitcher -0.011, composite +0.537
confidence pipeline reached 0.583 before later publication/recency controls; published home probability was 0.530
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

MIN offense over-projected by 3
CHW offense under-projected by 4
Bailey Ober (MIN) allowed 4 fewer ER than projected (1 vs 5)

Team	Starter	Projected	Actual	ER err
MIN	Bailey Ober	5 ER / 16 outs	1 ER / 13 outs	-4
CHW	Bryan Hudson	0 ER / 15 outs	0 ER / 5 outs	0

Decision internals: run factors home×0.989 / away×0.999; 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:

Ryan Jeffers (MIN C): 0-for-2, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Brooks Lee (MIN 3B): 2-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Luke Keaschall (MIN RF): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -1.11]

Starter miss: Bailey Ober (MIN) posted 2.08 ERA vs projected 8.44 (1 ER in 13 outs); he is materially worse pitching away (6.20 ERA there vs 3.16 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

Bailey Ober (MIN) threw 13 outs, 1 ER = 2.08 ERA against a projected 8.44 ERA (-6.36 ERA error)

Bullpen behind him: 0 ER over 0 outs.

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

No individual hitter beat his projection by a meaningful margin.

Verdict: offense: the lineup scored well below its projection

HOME pitching

No settled pitching line for this club. The forecast named Bryan Hudson (CHW) at a projected 0 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 3 hits and 1 HR, against a projected 5 runs / 8 hits (-4 runs vs projection).

Who did the damage to him:

Brooks Lee: 2-for-3, 1 HR, 1 RBI [+1.4 runs vs projection] season xwOBA 0.276, weak - 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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
MIN	Bailey Ober (Starter)	13 / 16	2.08 / 8.44 (-6.36)	1 / 5	4 / 4	0 / 2	3 / 3	6 / 4	venue away: 6.2 vs 3.16 ERA (worse here) · opp run factor 0.99 (average) · 21 prior appearances this season

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
MIN	Ryan Jeffers (C)	2 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	0 / 1	-1.4	0.357 elite	bad day
MIN	Brooks Lee (3B)	3 / 5	2 / 1	0 / 0	0 / 0	1 / 0	1 / 1	1 / 0	0 / 0	0 / 1	1.4	0.276 weak	good day
MIN	Luke Keaschall (RF)	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	2 / 0	-1.11	0.314 average	no signal
MIN	Josh Bell (DH)	3 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	2 / 1	-0.47	0.331 above average	—
MIN	Kody Clemens (2B)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 1	-0.47	0.339 elite	—
MIN	Trevor Larnach (LF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	3 / 0	-0.47	0.327 above average	—
MIN	Kaelen Culpepper (SS)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	-0.47	0.313 average	—
MIN	Walker Jenkins (CF)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	2 / 1	-0.33	0.24 weak	—
MIN	Royce Lewis (1B)	3 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	-0.31	0.315 average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.372	positive favors home; negative favors away
Recent form	-0.018	positive favors home; negative favors away
Starting-pitcher matchup	-0.011	positive favors home; negative favors away
Composite	0.537	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.3-6.14	n/a	Base ML historical model prediction
variance_applied	3-6	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 matchup_recency_guardrail)
embedded_savant_run_environment	4-6	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-6	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-6	n/a	Capped at max 12 realistic runs
simulation_result	4-6	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.3-6.14	3-6	0.154	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	3-6	4-5	-2	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	4-5	4-6	1	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-6	4-6	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-6	4-6	0	Capped at max 12 realistic runs
simulation_result	4-6	4-6	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	Walker Jenkins	batter / CF	better	predicted_RBI 2.79
CHW	Miguel Vargas	batter / 3B	better	predicted_HR 1.8
MIN	Trevor Larnach	batter / LF	better	predicted_H 1.66
CHW	Chase Meidroth	batter / 2B	better	predicted_RBI 1.64
CHW	Munetaka Murakami	batter / 1B	worse	predicted_SO -1.505
CHW	Kyle Teel	batter / C	worse	predicted_SO -1.333
MIN	Ryan Jeffers	batter / C	worse	predicted_RBI -1.27
MIN	Josh Bell	batter / DH	worse	predicted_SO 1.184

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	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	CHW model-prediction batting totals differ from final simulation box score (H: current=8, sim=7; HR: current=1, sim=3; RBI: current=4, sim=6; R: current=5, sim=6; BB: current=4, sim=3; SO: current=9, sim=7).	CHW
info	CURRENT_PLAYER_MATCH_BOX	CHW current finalized hitter lines are model predictions and differ from the simulation box score: Sam Antonacci: AB current=3 box=4, RBI current=0 box=1, BB current=1 box=0; Munetaka Murakami: AB current=4 box=3, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Braden Montgomery: 2B current=0 box=1, RBI current=1 box=0, R current=0 box=1, BB current=1 box=0; Tristan Peters: AB current=3 box=4, 2B current=1 box=0, R current=0 box=1, SO current=0 box=1; Colson Montgomery: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Miguel Vargas: H current=1 box=2, HR current=0 box=2, RBI current=1 box=2, BB current=1 box=0; Andrew Benintendi: AB current=3 box=4, 2B current=0 box=1, RBI current=0 box=1, R current=0 box=1; Kyle Teel: AB current=3 box=2, BB current=0 box=1, SO current=1 box=0	CHW
info	CURRENT_FINALIZED_MATCH_BOX	MIN model pitcher projections differ from final simulation pitching box score (H: current=6, sim=7; HR: current=2, sim=3).	MIN
info	CURRENT_FINALIZED_MATCH_BOX	MIN model-prediction batting totals differ from final simulation box score (AB: current=36, sim=35; 2B: current=2, sim=0; HR: current=1, sim=2; RBI: current=5, sim=4; R: current=5, sim=4; BB: current=3, sim=1; SO: current=7, sim=9).	MIN
info	CURRENT_PLAYER_MATCH_BOX	MIN current finalized hitter lines are model predictions and differ from the simulation box score: Luke Keaschall: 2B current=1 box=0, BB current=1 box=0, SO current=0 box=1; Josh Bell: AB current=5 box=4, RBI current=1 box=0, SO current=1 box=2; Ryan Jeffers: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Brooks Lee: AB current=5 box=4, RBI current=1 box=0, R current=0 box=1; Kody Clemens: AB current=4 box=3, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Royce Lewis: 2B current=1 box=0, RBI current=1 box=0, SO current=1 box=2; Trevor Larnach: AB current=3 box=4, H current=1 box=2, HR current=0 box=1, RBI current=0 box=1; Walker Jenkins: AB current=3 box=4, H current=0 box=1, HR current=0 box=1, RBI current=0 box=3	MIN
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

NYY@SDMISS

Predicted: NYY 4-3 (conf -0.07, lean)Actual: SD 3-4Brier 0.287total bias +0

Root cause: Several errors of similar size

Evidence-backed diagnosis

published system edge 0.07 (lean); this is not the half-scale distance from 50%
team-run absolute error 2; total bias +0
pregame home-minus-away evidence: season -0.776, recent -0.628, pitcher -0.026, composite -0.820
confidence pipeline reached 0.549 before later publication/recency controls; published home probability was 0.464
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

No large single-factor errors.

Team	Starter	Projected	Actual	ER err
NYY	Gerrit Cole	3 ER / 19 outs	4 ER / 21 outs	1
SD	Michael King	2 ER / 18 outs	0 ER / 18 outs	-2

Decision internals: run factors home×0.989 / away×1.014; 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:

Ben Rice (NYY DH): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Luis Campusano (SD C): 2-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Jake Cronenworth (SD 2B): 1-for-4, 1 HR, 1 RBI (projected 0.0 H, 0.0 HR) [run impact 1.07]
Manny Machado (SD 3B): 1-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Jackson Merrill (SD CF): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.78]

Starter miss: Michael King (SD) posted 0.00 ERA vs projected 3.00 (0 ER in 18 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

Gerrit Cole (NYY) threw 21 outs, 4 ER = 5.14 ERA against a projected 4.26 ERA (0.88 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 4 runs on 6 hits and 4 HR, against a projected 3 runs / 7 hits (1 runs vs projection).

Who did the damage to him:

Luis Campusano: 2-for-3, 1 HR, 1 RBI [+1.4 runs vs projection] season xwOBA 0.344, elite - real projection error
Jake Cronenworth: 1-for-4, 1 HR, 1 RBI [+1.07 runs vs projection] season xwOBA 0.271, weak - a good day, not a model error
Manny Machado: 1-for-3, 1 HR, 1 RBI [+0.93 runs vs projection] season xwOBA 0.336, above average - real projection error

Verdict: offense: the lineup scored well below its projection

HOME pitching

Michael King (SD) threw 18 outs, 0 ER = 0 ERA against a projected 3 ERA (-3 ERA error)

Bullpen behind him: 2 ER over 9 outs.

The AWAY lineup he faced produced 2 runs on 6 hits and 0 HR, against a projected 5 runs / 8 hits (-3 runs vs projection).

Who did the damage to him:

Spencer Jones: 1-for-3, 0 HR, 1 RBI [+0.64 runs vs projection] season xwOBA 0.327, above average - real projection error
George Lombard Jr.: 1-for-4, 0 HR, 0 RBI [+0.31 runs vs projection] season xwOBA 0.283, weak - 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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
NYY	Gerrit Cole (Starter)	21 / 19	5.14 / 4.26	4 / 3	6 / 4	4 / 0	0 / 1	3 / 5	venue away: 3.23 vs 3.35 ERA (ok) · opp run factor 0.99 (average) · 18 prior appearances this season
SD	Michael King (Starter)	18 / 18	0 / 3 (-3.00)	0 / 2	4 / 5	0 / 1	3 / 2	3 / 5	venue home: 2.94 vs 3.29 ERA (ok) · opp run factor 1.01 (average) · 28 prior appearances this season
SD	Adrian Morejon (Setup Man)	3 / 3	18 / 0 (+18.00)	2 / 0	3 / 0	0 / 0	1 / 0	1 / 1
SD	Bradgley Rodriguez (Middle Reliever (High Leverage))	3 / 3	0 / 18 (-18.00)	0 / 2	0 / 1	0 / 1	0 / 1	2 / 0
SD	Mason Miller (Closer)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	3 / 1

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
NYY	Ben Rice (DH)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	2 / 1	-1.4	0.376 elite	bad day
SD	Luis Campusano (C)	3 / 3	2 / 1	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 0	0 / 1	1.4	0.344 elite	projection too low
SD	Jake Cronenworth (2B)	4 / 4	1 / 0	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 1	0 / 1	1.07	0.271 weak	good day
SD	Manny Machado (3B)	3 / 4	1 / 1	0 / 0	0 / 0	1 / 0	1 / 1	1 / 1	1 / 1	0 / 1	0.93	0.336 above average	projection too low
SD	Jackson Merrill (CF)	3 / 5	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	0 / 1	-0.78	0.338 above average	bad day
NYY	Spencer Jones (CF)	3 / 4	1 / 1	1 / 0	0 / 0	0 / 0	1 / 1	1 / 1	1 / 0	1 / 1	0.64	0.327 above average	projection too low
SD	Ty France (1B)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.332 above average	—
SD	Xander Bogaerts (SS)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.32 average	—
NYY	Cody Bellinger (LF)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	0 / 1	-0.33	0.341 elite	—
NYY	Jazz Chisholm Jr. (2B)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	-0.33	0.295 below average	—
SD	Fernando Tatis Jr. (RF)	4 / 5	1 / 1	0 / 0	0 / 0	1 / 1	1 / 1	1 / 1	0 / 1	0 / 1	-0.33	0.375 elite	—
NYY	George Lombard Jr. (SS)	4 / 3	1 / 1	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0.31	0.283 weak	—
NYY	Jasson Domínguez (RF)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.31	0.291 below average	—
NYY	Austin Wells (C)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	2 / 0	1 / 1	0.19	0.292 below average	—
NYY	Luis García Jr. (1B)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	0	0.349 elite	—
NYY	Ryan McMahon (3B)	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0	0.305 average	—
SD	Ethan Salas (DH)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.147 weak	—
SD	Dustin Harris (LF)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0	0.328 above average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.776	positive favors home; negative favors away
Recent form	-0.628	positive favors home; negative favors away
Starting-pitcher matchup	-0.026	positive favors home; negative favors away
Composite	-0.82	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.11-4.01	n/a	Base ML historical model prediction
variance_applied	4-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-3	0.536	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	4.11-4.01	4-4	-0.104	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-4	4-3	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	4-3	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
NYY	Gerrit Cole	pitcher / Starter	better	predicted_SO 3
NYY	Brent Headrick	pitcher / Setup Man	worse	predicted_SO -2
SD	Michael King	pitcher / Starter	better	predicted_SO 2
SD	Bradgley Rodriguez	pitcher / Middle Reliever (High Leverage)	worse	predicted_SO -2
NYY	Ryan McMahon	batter / 3B	better	predicted_RBI 1.86
NYY	Ben Rice	batter / DH	worse	predicted_AB -1.45
NYY	Cody Bellinger	batter / LF	worse	predicted_SO 1.345
SD	Dustin Harris	batter / LF	worse	predicted_AB 1.238

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	NYY pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	NYY
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	NYY model-prediction batting totals differ from final simulation box score (AB: current=32, sim=33; H: current=8, sim=6; 2B: current=1, sim=0; HR: current=1, sim=2; R: current=5, sim=4; SO: current=9, sim=7).	NYY
info	CURRENT_PLAYER_MATCH_BOX	NYY current finalized hitter lines are model predictions and differ from the simulation box score: Ben Rice: AB current=4 box=3, HR current=1 box=0, SO current=1 box=2; Cody Bellinger: AB current=5 box=4, RBI current=1 box=0, BB current=1 box=0, SO current=1 box=2; Spencer Jones: RBI current=1 box=0, R current=1 box=0; Luis García Jr.: AB current=3 box=4, R current=1 box=0, SO current=1 box=0; Jazz Chisholm Jr.: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=1, BB current=1 box=0; George Lombard Jr.: H current=1 box=0, BB current=0 box=1, SO current=1 box=0; Jasson Domínguez: H current=1 box=0, 2B current=1 box=0; Ryan McMahon: AB current=3 box=4, H current=0 box=1, HR current=0 box=1, RBI current=0 box=2	NYY
info	CURRENT_FINALIZED_MATCH_BOX	SD model-prediction batting totals differ from final simulation box score (H: current=7, sim=6; 2B: current=1, sim=2; HR: current=1, sim=0; BB: current=3, sim=1; SO: current=8, sim=5).	SD
info	CURRENT_PLAYER_MATCH_BOX	SD current finalized hitter lines are model predictions and differ from the simulation box score: Fernando Tatis Jr.: AB current=5 box=4, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Jake Cronenworth: BB current=1 box=0, SO current=1 box=2; Manny Machado: AB current=4 box=3, H current=1 box=2, 2B current=0 box=1, SO current=1 box=0; Ty France: AB current=3 box=4, H current=1 box=0, SO current=1 box=0; Xander Bogaerts: RBI current=0 box=1, R current=0 box=1, SO current=1 box=0; Dustin Harris: AB current=3 box=4, H current=1 box=0, SO current=0 box=1; Jackson Merrill: AB current=5 box=4, H current=1 box=2, 2B current=1 box=0, RBI current=1 box=0; Luis Campusano: AB current=3 box=4, 2B current=0 box=1, RBI current=0 box=1, R current=0 box=1	SD
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

OAK@SEAHIT

Predicted: SEA 6-7 (conf +0.09, lean)Actual: SEA 0-2Brier 0.208total bias +11

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.09 (lean); this is not the half-scale distance from 50%
team-run absolute error 11; total bias +11
pregame home-minus-away evidence: season +0.352, recent -0.706, pitcher +0.011, composite +0.222
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, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 11 (13 vs 2)
OAK offense over-projected by 6
SEA offense over-projected by 5
Bryan Woo (SEA) allowed 3 fewer ER than projected (0 vs 3)

Team	Starter	Projected	Actual	ER err
OAK	Gage Jump	4 ER / 14 outs	2 ER / 12 outs	-2
SEA	Bryan Woo	3 ER / 17 outs	0 ER / 24 outs	-3

Decision internals: run factors home×1.016 / away×1.005; 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:

Zack Gelof (OAK 3B): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Cal Raleigh (SEA DH): 1-for-3, 0 HR, 1 RBI (projected 1.0 H, 1.0 HR) [run impact -0.93]
Lawrence Butler (OAK RF): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Alika Williams (OAK SS): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.78]

Starter miss: Bryan Woo (SEA) posted 0.00 ERA vs projected 4.76 (0 ER in 24 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

Gage Jump (OAK) threw 12 outs, 2 ER = 4.5 ERA against a projected 7.71 ERA (-3.21 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 2 runs on 7 hits and 0 HR, against a projected 4 runs / 8 hits (-2 runs vs projection).

No individual hitter beat his projection by a meaningful margin.

Verdict: offense: the lineup scored well below its projection

HOME pitching

Bryan Woo (SEA) threw 24 outs, 0 ER = 0 ERA against a projected 4.76 ERA (-4.76 ERA error)

Bullpen behind him: 0 ER over 3 outs.

The AWAY lineup he faced produced 0 runs on 1 hits and 0 HR, against a projected 4 runs / 9 hits (-4 runs vs projection).

No individual hitter beat his projection by a meaningful margin.

Verdict: offense: the lineup scored well below its projection

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
OAK	Gage Jump (Starter)	12 / 14	4.5 / 7.71 (-3.21)	2 / 4	6 / 5	0 / 2	0 / 2	7 / 2	venue away: 4.61 vs 5.62 ERA (ok) · opp run factor 1.02 (weak) · 18 prior appearances this season
SEA	Bryan Woo (Starter)	24 / 17	0 / 4.76 (-4.76)	0 / 3	1 / 4	0 / 1	2 / 0	9 / 7	venue home: 2.63 vs 5.62 ERA (ok) · opp run factor 1 (weak) · 26 prior appearances this season
SEA	Andrés Muñoz (Closer)	3 / 4	0 / 6.75 (-6.75)	0 / 1	0 / 1	0 / 1	0 / 0	1 / 1

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
OAK	Zack Gelof (3B)	4 / 5	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	2 / 1	-1.4	0.313 average	no signal
SEA	Cal Raleigh (DH)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	1 / 1	2 / 1	-0.93	0.316 average	no signal
OAK	Lawrence Butler (RF)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	1 / 1	-0.8	0.322 above average	bad day
OAK	Alika Williams (SS)	3 / 3	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.78	0.243 weak	projection too high
OAK	Jeff McNeil (1B)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 0	-0.47	0.306 average	—
OAK	Max Muncy (DH)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.291 below average	—
OAK	Carlos Cortes (LF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.304 below average	—
OAK	Brian Serven (C)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 1	-0.47	0.302 below average	—
SEA	Michael Arroyo (2B)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	-0.47	—	—
SEA	Jhonny Pereda (C)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.342 elite	—
OAK	Henry Bolte (CF)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	2 / 1	-0.33	0.311 average	—
SEA	J.P. Crawford (SS)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 1	1 / 1	-0.33	0.325 above average	—
SEA	Weston Wilson (3B)	2 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	2 / 1	-0.33	0.257 weak	—
SEA	Julio Rodríguez (CF)	4 / 5	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 0	2 / 1	-0.31	0.335 above average	—
SEA	Randy Arozarena (LF)	4 / 4	2 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	2 / 1	-0.17	0.348 elite	—
OAK	Donovan Walton (2B)	1 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	1 / 1	-0.14	0.333 above average	—
SEA	Dominic Canzone (RF)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	0 / 0	1 / 1	0	0.355 elite	—
SEA	Josh Naylor (1B)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0	0.314 average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.352	positive favors home; negative favors away
Recent form	-0.706	positive favors home; negative favors away
Starting-pitcher matchup	0.011	positive favors home; negative favors away
Composite	0.222	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.79-4.82	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-7	0.544	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, season_quality_guardrail)
embedded_savant_run_environment	6-7	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-7	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-7	n/a	Capped at max 12 realistic runs
simulation_result	6-7	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.79-4.82	4-5	-0.03	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	6-7	0	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, season_quality_guardrail)
embedded_savant_run_environment	6-7	6-7	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-7	6-7	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-7	6-7	0	Capped at max 12 realistic runs
simulation_result	6-7	6-7	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	Jeff McNeil	batter / 1B	better	predicted_RBI 2.581
SEA	Dominic Canzone	batter / RF	better	predicted_RBI 2.488
OAK	Zack Gelof	batter / 3B	better	predicted_2B 1.841
OAK	Carlos Cortes	batter / LF	boosted	predicted_RBI 1.635
SEA	Michael Arroyo	batter / 2B	better	predicted_RBI 1.607
OAK	Donovan Walton	batter / 2B	worse	predicted_SO 1.469
SEA	Randy Arozarena	batter / LF	better	predicted_RBI 1.46
OAK	Alika Williams	batter / SS	worse	predicted_SO 1.306

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	SEA pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	SEA
info	CURRENT_FINALIZED_MATCH_BOX	OAK model-prediction batting totals differ from final simulation box score (AB: current=32, sim=34; H: current=9, sim=7; 2B: current=1, sim=2; HR: current=1, sim=3; RBI: current=5, sim=6; R: current=4, sim=6; BB: current=3, sim=0; SO: current=8, sim=10).	OAK
info	CURRENT_PLAYER_MATCH_BOX	OAK current finalized hitter lines are model predictions and differ from the simulation box score: Henry Bolte: BB current=1 box=0; Jeff McNeil: H current=1 box=2, HR current=0 box=2, RBI current=1 box=3; Max Muncy: AB current=3 box=4, R current=0 box=1, SO current=1 box=0; Carlos Cortes: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=2, R current=0 box=1; Alika Williams: H current=1 box=0, 2B current=1 box=0, SO current=1 box=2; Donovan Walton: AB current=3 box=4, H current=1 box=0, BB current=1 box=0, SO current=1 box=2; Zack Gelof: AB current=5 box=4, H current=1 box=2, 2B current=0 box=2, HR current=1 box=0; Lawrence Butler: H current=1 box=0, RBI current=1 box=0, R current=0 box=1, BB current=1 box=0	OAK
info	CURRENT_FINALIZED_MATCH_BOX	SEA model-prediction batting totals differ from final simulation box score (AB: current=32, sim=30; H: current=8, sim=6; 2B: current=2, sim=0; HR: current=1, sim=3; RBI: current=4, sim=7; R: current=4, sim=7; SO: current=8, sim=4).	SEA
info	CURRENT_PLAYER_MATCH_BOX	SEA current finalized hitter lines are model predictions and differ from the simulation box score: Cal Raleigh: AB current=4 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Dominic Canzone: AB current=3 box=4, H current=1 box=3, HR current=0 box=1, RBI current=1 box=3; Josh Naylor: H current=1 box=0, SO current=0 box=1; J.P. Crawford: AB current=3 box=4, R current=0 box=1, BB current=1 box=0, SO current=1 box=0; Michael Arroyo: AB current=3 box=2, HR current=0 box=1, RBI current=0 box=2, BB current=0 box=1; Julio Rodríguez: AB current=5 box=4, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Randy Arozarena: AB current=4 box=3, 2B current=1 box=0, HR current=0 box=1, RBI current=1 box=2; Jhonny Pereda: H current=1 box=0, SO current=1 box=0	SEA
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

SF@NYMHIT

Predicted: NYM 4-6 (conf +0.03, toss_up)Actual: NYM 2-4Brier 0.234total bias +4

Root cause: Toss-up that went our way

Evidence-backed diagnosis

published system edge 0.03 (toss_up); this is not the half-scale distance from 50%
team-run absolute error 4; total bias +4
pregame home-minus-away evidence: season +0.029, recent -0.106, pitcher +0.024, composite +0.402
confidence pipeline reached 0.547 before later publication/recency controls; published home probability was 0.516
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 9 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 4 (10 vs 6)
Cesar Perdomo (SF) allowed 6 fewer ER than projected (0 vs 6)

Team	Starter	Projected	Actual	ER err
SF	Cesar Perdomo	6 ER / 12 outs	0 ER / 18 outs	-6
NYM	Christian Scott	3 ER / 15 outs	2 ER / 14 outs	-1

Decision internals: run factors home×1 / away×0.976; 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:

Rafael Devers (SF DH): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.26]
Juan Soto (NYM DH): 1-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.93]
Bo Bichette (NYM 3B): 1-for-4, 1 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Bryce Eldridge (SF 1B): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.64]

Starter miss: Cesar Perdomo (SF) posted 0.00 ERA vs projected 13.50 (0 ER in 18 outs). Thin prior record (1 prior appearances this season): high projection uncertainty.

What to do: Reduce game confidence for starters the feed has little on (scale by prior appearances); prefer priors and validate any run-param change in shadow.

Pitcher versus batter: what actually broke this game

AWAY pitching

Cesar Perdomo (SF) threw 18 outs, 0 ER = 0 ERA against a projected 13.5 ERA (-13.5 ERA error) only 1 prior appearances in the covered season - the projection had little of his own record to stand on

Bullpen behind him: 4 ER over 3 outs.

The HOME lineup he faced produced 3 runs on 6 hits and 1 HR, against a projected 4 runs / 8 hits (-1 runs vs projection).

Who did the damage to him:

Bo Bichette: 1-for-4, 1 HR, 3 RBI [+0.93 runs vs projection] season xwOBA 0.326, above average - real projection error
Carson Benge: 1-for-4, 0 HR, 0 RBI [+0.31 runs vs projection] season xwOBA 0.336, above average - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

HOME pitching

Christian Scott (NYM) threw 14 outs, 2 ER = 3.86 ERA against a projected 5.4 ERA (-1.54 ERA error)

Bullpen behind him: 0 ER over 4 outs.

The AWAY lineup he faced produced 2 runs on 5 hits and 0 HR, against a projected 4 runs / 8 hits (-2 runs vs projection).

No individual hitter beat his projection by a meaningful margin.

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
SF	Cesar Perdomo (Starter)	18 / 12	0 / 13.5 (-13.50)	0 / 6	3 / 5	0 / 2	1 / 1	4 / 4	opp run factor 1 (weak) · thin sample: 1 prior appearances this season (high uncertainty)
SF	Carson Seymour (Middle Reliever (High Leverage))	3 / 4	36 / 0 (+36.00)	4 / 0	3 / 1	2 / 0	1 / 0	0 / 1
NYM	Christian Scott (Starter)	14 / 15	3.86 / 5.4	2 / 3	4 / 3	0 / 1	0 / 3	7 / 3	venue home: 3.93 vs 3.98 ERA (ok) · opp run factor 0.98 (weak) · 20 prior appearances this season
NYM	Nathan Lavender (Middle Reliever (High Leverage))	1 / 4	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0
NYM	Kodai Senga (Closer)	3 / 4	0 / 6.75 (-6.75)	0 / 1	0 / 1	0 / 0	0 / 0	1 / 1

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
SF	Rafael Devers (DH)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	3 / 1	-1.26	0.339 elite	bad day
NYM	Juan Soto (DH)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	1 / 1	0 / 0	-0.93	0.416 elite	bad day
NYM	Bo Bichette (3B)	4 / 4	1 / 1	0 / 0	0 / 0	1 / 0	3 / 0	1 / 1	0 / 0	1 / 1	0.93	0.326 above average	projection too low
SF	Bryce Eldridge (1B)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 0	0 / 1	1 / 1	-0.64	0.356 elite	bad day
SF	Osleivis Basabe (3B)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	3 / 1	-0.47	0.273 weak	—
SF	Christian Koss (SS)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	-0.47	0.236 weak	—
SF	Drew Gilbert (CF)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	3 / 1	-0.47	0.325 above average	—
NYM	Mark Vientos (1B)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	-0.47	0.312 average	—
NYM	Luis Torrens (C)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.295 below average	—
SF	Grant McCray (LF)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	-0.33	0.311 average	—
NYM	Marcus Semien (2B)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	-0.33	0.306 average	—
NYM	Carson Benge (RF)	4 / 4	1 / 1	1 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	0.31	0.336 above average	—
NYM	A.J. Ewing (CF)	3 / 4	2 / 1	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0 / 1	-0.17	0.322 above average	—
NYM	Francisco Lindor (SS)	3 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 0	0 / 1	-0.14	0.338 above average	—
SF	Jonah Cox (2B)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	1 / 1	0	0.273 weak	—
SF	Drew Cavanaugh (C)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0	0.322 above average	—
SF	Jung Hoo Lee (RF)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 0	2 / 1	0 / 0	0 / 0	0 / 1	0	0.316 average	—
NYM	Nick Morabito (LF)	2 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 1	0	0.294 below average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.029	positive favors home; negative favors away
Recent form	-0.106	positive favors home; negative favors away
Starting-pitcher matchup	0.024	positive favors home; negative favors away
Composite	0.402	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.44-4.92	n/a	Base ML historical model prediction
variance_applied	3-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.516	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	4-6	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-6	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-6	n/a	Capped at max 12 realistic runs
simulation_result	4-6	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.44-4.92	3-5	0.521	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	3-5	4-5	-1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	4-5	4-6	1	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-6	4-6	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-6	4-6	0	Capped at max 12 realistic runs
simulation_result	4-6	4-6	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
NYM	Francisco Lindor	batter / SS	better	predicted_RBI 3.41
SF	Jonah Cox	batter / 2B	better	predicted_RBI 2.684
SF	Jung Hoo Lee	batter / RF	worse	predicted_AB -1.88
SF	Osleivis Basabe	batter / 3B	worse	predicted_SO 1.333
NYM	Bo Bichette	batter / 3B	worse	predicted_H -1.19
NYM	Mark Vientos	batter / 1B	better	predicted_H 1.12
SF	Bryce Eldridge	batter / 1B	better	predicted_SO -1.093
NYM	A.J. Ewing	batter / CF	better	predicted_SO -1.091

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	SF pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	SF
info	CURRENT_FINALIZED_MATCH_BOX	NYM model pitcher projections differ from final simulation pitching box score (H: current=4, sim=3).	NYM
info	CURRENT_FINALIZED_MATCH_BOX	NYM model-prediction batting totals differ from final simulation box score (AB: current=34, sim=31; H: current=8, sim=7; 2B: current=1, sim=0; 3B: current=0, sim=1; HR: current=1, sim=3; RBI: current=4, sim=6; R: current=4, sim=6; BB: current=3, sim=1).	NYM
info	CURRENT_PLAYER_MATCH_BOX	NYM current finalized hitter lines are model predictions and differ from the simulation box score: Juan Soto: H current=1 box=2, HR current=1 box=0, RBI current=1 box=0, BB current=1 box=0; Carson Benge: H current=1 box=0, R current=1 box=0; Mark Vientos: AB current=3 box=4, H current=1 box=2, HR current=0 box=1, R current=0 box=1; Nick Morabito: RBI current=1 box=0, SO current=1 box=2; A.J. Ewing: AB current=4 box=3, 2B current=1 box=0, HR current=0 box=1, RBI current=0 box=1; Bo Bichette: H current=1 box=0; Francisco Lindor: AB current=5 box=4, H current=1 box=2, 3B current=0 box=1, HR current=0 box=1; Marcus Semien: AB current=4 box=3, H current=1 box=0, BB current=1 box=0, SO current=1 box=0	NYM
info	CURRENT_FINALIZED_MATCH_BOX	SF model pitcher projections differ from final simulation pitching box score (H: current=6, sim=7; HR: current=2, sim=3).	SF
info	CURRENT_FINALIZED_MATCH_BOX	SF model-prediction batting totals differ from final simulation box score (AB: current=31, sim=30; H: current=8, sim=3; 2B: current=1, sim=0; SO: current=9, sim=4; HBP: current=0, sim=1).	SF
info	CURRENT_PLAYER_MATCH_BOX	SF current finalized hitter lines are model predictions and differ from the simulation box score: Rafael Devers: HR current=1 box=0, BB current=1 box=0, SO current=1 box=0; Bryce Eldridge: 2B current=1 box=0, RBI current=1 box=0, R current=0 box=1, BB current=1 box=0; Jonah Cox: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=3; Osleivis Basabe: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Grant McCray: SO current=1 box=0; Christian Koss: H current=1 box=0, R current=1 box=0, SO current=1 box=0; Drew Cavanaugh: H current=1 box=0, SO current=1 box=0; Drew Gilbert: AB current=3 box=2, H current=1 box=0, R current=0 box=1, BB current=0 box=1	SF
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

STL@COLHIT

Predicted: STL 5-4 (conf -0.12, moderate)Actual: STL 10-8Brier 0.194total bias -9

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.12 (moderate); this is not the half-scale distance from 50%
team-run absolute error 9; total bias -9
pregame home-minus-away evidence: season -0.729, recent -0.186, pitcher -0.001, composite -0.628
confidence pipeline reached 0.582 before later publication/recency controls; published home probability was 0.441
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

total runs under-projected by 9 (9 vs 18)
STL offense under-projected by 5
COL offense under-projected by 4
Tanner Gordon (COL) allowed 3 more ER than projected (5 vs 2)

Team	Starter	Projected	Actual	ER err
STL	Kyle Leahy	3 ER / 16 outs	5 ER / 8 outs	2
COL	Tanner Gordon	2 ER / 15 outs	5 ER / 14 outs	3

Decision internals: run factors home×1.012 / away×1.013; 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:

Jake McCarthy (COL LF): 3-for-5, 1 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 1.56]
Mickey Moniak (COL DH): 2-for-4, 1 HR, 4 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Alec Burleson (STL 1B): 3-for-5, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.25]
Nolan Gorman (STL 3B): 1-for-5, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Iván Herrera (STL DH): 3-for-5, 0 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 0.92]

Starter miss: Kyle Leahy (STL) posted 16.88 ERA vs projected 5.06 (5 ER in 8 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

Kyle Leahy (STL) threw 8 outs, 5 ER = 16.88 ERA against a projected 5.06 ERA (11.82 ERA error)

Bullpen behind him: 0 ER over 1 outs.

The HOME lineup he faced produced 6 runs on 9 hits and 2 HR, against a projected 5 runs / 9 hits (1 runs vs projection).

Who did the damage to him:

Jake McCarthy: 3-for-5, 1 HR, 3 RBI [+1.56 runs vs projection] season xwOBA 0.322, above average - real projection error
Mickey Moniak: 2-for-4, 1 HR, 4 RBI [+1.4 runs vs projection] season xwOBA 0.285, 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

Tanner Gordon (COL) threw 14 outs, 5 ER = 9.64 ERA against a projected 3.6 ERA (6.04 ERA error)

Bullpen behind him: 2 ER over 1 outs.

The AWAY lineup he faced produced 9 runs on 14 hits and 1 HR, against a projected 5 runs / 9 hits (4 runs vs projection).

Who did the damage to him:

Alec Burleson: 3-for-5, 0 HR, 1 RBI [+1.25 runs vs projection] season xwOBA 0.373, elite - real projection error
Nolan Gorman: 1-for-5, 1 HR, 2 RBI [+0.93 runs vs projection] season xwOBA 0.276, weak - a good day, not a model error
Iván Herrera: 3-for-5, 0 HR, 2 RBI [+0.92 runs vs projection] season xwOBA 0.332, above average - real projection error

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

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
STL	Kyle Leahy (Starter)	8 / 16	16.88 / 5.06 (+11.82)	5 / 3	6 / 2	1 / 0	0 / 0	6 / 6	venue away: 3.38 vs 2.91 ERA (ok) · opp run factor 1.01 (average) · 26 prior appearances this season
STL	George Soriano (Setup Man)	1 / 5	0 / 5.4 (-5.40)	0 / 1	1 / 1	0 / 1	0 / 0	1 / 1
COL	Tanner Gordon (Starter)	14 / 15	9.64 / 3.6 (+6.04)	5 / 2	8 / 4	0 / 0	1 / 3	4 / 5	venue home: 7.71 vs 1.59 ERA (worse here) · opp run factor 1.01 (average) · 20 prior appearances this season
COL	Jimmy Herget (Middle Reliever (High Leverage))	1 / 3	54 / 27 (+27.00)	2 / 3	2 / 4	1 / 0	0 / 0	1 / 0

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
COL	Jake McCarthy (LF)	5 / 5	3 / 1	0 / 1	0 / 0	1 / 0	3 / 1	2 / 1	0 / 0	0 / 1	1.56	0.322 above average	projection too low
COL	Mickey Moniak (DH)	4 / 4	2 / 1	0 / 0	0 / 0	1 / 0	4 / 1	1 / 0	0 / 0	2 / 1	1.4	0.285 below average	good day
STL	Alec Burleson (1B)	5 / 4	3 / 1	1 / 0	0 / 0	0 / 0	1 / 1	1 / 0	0 / 0	0 / 1	1.25	0.373 elite	projection too low
STL	Nolan Gorman (3B)	5 / 3	1 / 1	0 / 0	0 / 0	1 / 0	2 / 0	1 / 0	0 / 0	0 / 1	0.93	0.276 weak	good day
STL	Iván Herrera (DH)	5 / 5	3 / 1	1 / 0	0 / 0	0 / 0	2 / 1	2 / 1	0 / 1	1 / 1	0.92	0.332 above average	projection too low
COL	Adael Amador (2B)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 0	-0.8	0.253 weak	projection too high
COL	Cole Carrigg (CF)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	2 / 1	-0.64	0.32 average	no signal
COL	Troy Johnston (RF)	5 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	4 / 1	-0.47	0.326 above average	—
COL	Vimael Machín (3B)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	2 / 1	-0.47	—	—
COL	Brett Sullivan (C)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	-0.47	0.288 below average	—
COL	Ezequiel Tovar (SS)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.276 weak	—
STL	Jordan Walker (RF)	4 / 5	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 0	2 / 1	-0.45	0.347 elite	—
STL	Thomas Saggese (SS)	4 / 3	2 / 1	1 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0 / 1	1 / 1	0.45	0.313 average	—
STL	Joshua Baez (LF)	3 / 3	2 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	1 / 0	0 / 1	-0.13	0.337 above average	—
COL	TJ Rumfield (1B)	4 / 4	3 / 1	1 / 0	0 / 0	0 / 1	1 / 1	2 / 1	0 / 1	1 / 1	-0.01	0.334 above average	—
STL	Bryan Torres (2B)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.298 below average	—
STL	Leonardo Bernal (C)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 0	2 / 1	1 / 1	1 / 1	1 / 1	0	0.365 elite	—
STL	Nathan Church (CF)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	2 / 0	0	0.282 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.729	positive favors home; negative favors away
Recent form	-0.186	positive favors home; negative favors away
Starting-pitcher matchup	-0.001	positive favors home; negative favors away
Composite	-0.628	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	5.47-4.02	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	0.559	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	5.47-4.02	5-4	-0.456	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
COL	TJ Rumfield	batter / 1B	worse	predicted_RBI 2.44
STL	Nathan Church	batter / CF	better	predicted_RBI 1.66
STL	Leonardo Bernal	batter / C	worse	predicted_H -1.58
STL	Alec Burleson	batter / 1B	better	predicted_RBI 1.32
STL	Joshua Baez	batter / LF	worse	predicted_SO -1.3
COL	Jake McCarthy	batter / LF	worse	predicted_H -1.2
STL	Nolan Gorman	batter / 3B	worse	predicted_SO -1.197
STL	Iván Herrera	batter / DH	worse	predicted_SO 1.165

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	STL pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	STL
info	CURRENT_FINALIZED_MATCH_BOX	COL model-prediction batting totals differ from final simulation box score (AB: current=32, sim=30; H: current=9, sim=3; 2B: current=2, sim=0; RBI: current=5, sim=4; R: current=5, sim=4; BB: current=3, sim=0; SO: current=7, sim=8).	COL
info	CURRENT_PLAYER_MATCH_BOX	COL current finalized hitter lines are model predictions and differ from the simulation box score: Cole Carrigg: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Mickey Moniak: AB current=4 box=3, RBI current=1 box=0, R current=0 box=1, SO current=1 box=2; Troy Johnston: AB current=3 box=4, H current=1 box=0; Adael Amador: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Vimael Machín: HR current=0 box=1, RBI current=0 box=1; Jake McCarthy: AB current=5 box=4, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; TJ Rumfield: AB current=4 box=3, HR current=1 box=0, RBI current=1 box=3, BB current=1 box=0; Brett Sullivan: H current=1 box=0	COL
info	CURRENT_FINALIZED_MATCH_BOX	STL model-prediction batting totals differ from final simulation box score (AB: current=34, sim=36; 2B: current=1, sim=2; HR: current=1, sim=0; SO: current=8, sim=7; HBP: current=0, sim=2).	STL
info	CURRENT_PLAYER_MATCH_BOX	STL current finalized hitter lines are model predictions and differ from the simulation box score: Bryan Torres: H current=1 box=0, BB current=0 box=1, HBP current=0 box=1; Jordan Walker: H current=1 box=2, 2B current=1 box=0; Joshua Baez: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Alec Burleson: AB current=4 box=5, H current=1 box=2, RBI current=1 box=2, R current=0 box=1; Iván Herrera: RBI current=1 box=0, BB current=1 box=0, SO current=1 box=2; Thomas Saggese: 2B current=0 box=1, SO current=1 box=0; Leonardo Bernal: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Nolan Gorman: AB current=3 box=4, SO current=1 box=0	STL
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

TB@TEXMISS

Predicted: TB 5-4 (conf -0.05, lean)Actual: TEX 6-8Brier 0.278total bias -5

Root cause: Run environment far higher than projected (under-dispersed)

Evidence-backed diagnosis

published system edge 0.05 (lean); this is not the half-scale distance from 50%
team-run absolute error 5; total bias -5
pregame home-minus-away evidence: season -0.688, recent -0.634, pitcher -0.021, composite -0.763
confidence pipeline reached 0.538 before later publication/recency controls; published home probability was 0.473
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

total runs under-projected by 5 (9 vs 14)
TEX offense under-projected by 4
MacKenzie Gore (TEX) allowed 3 more ER than projected (5 vs 2)

Team	Starter	Projected	Actual	ER err
TB	Ian Seymour	3 ER / 16 outs	3 ER / 9 outs	0
TEX	MacKenzie Gore	2 ER / 16 outs	5 ER / 9 outs	3

Decision internals: run factors home×1.002 / away×1.022; 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:

Jake Burger (TEX 1B): 2-for-4, 2 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 2.66]
Junior Caminero (TB 3B): 0-for-2, 0 HR, 1 RBI (projected 1.0 H, 1.0 HR) [run impact -1.07]
Logan O'Hoppe (TEX DH): 2-for-3, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 0.94]
Jonathan Aranda (TB 1B): 2-for-4, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.78]
Corey Seager (TEX SS): 1-for-5, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.6]

Starter miss: MacKenzie Gore (TEX) posted 15.00 ERA vs projected 3.38 (5 ER in 9 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

Ian Seymour (TB) threw 9 outs, 3 ER = 9 ERA against a projected 5.06 ERA (3.94 ERA error)

Bullpen behind him: 0 ER over 2 outs.

The HOME lineup he faced produced 8 runs on 11 hits and 3 HR, against a projected 4 runs / 7 hits (4 runs vs projection).

Who did the damage to him:

Jake Burger: 2-for-4, 2 HR, 3 RBI [+2.66 runs vs projection] season xwOBA 0.317, average - a good day, not a model error
Logan O'Hoppe: 2-for-3, 0 HR, 0 RBI [+0.94 runs vs projection] season xwOBA 0.268, weak - a good day, not a model error
Corey Seager: 1-for-5, 1 HR, 1 RBI [+0.6 runs vs projection] season xwOBA 0.35, elite - real projection error

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

HOME pitching

MacKenzie Gore (TEX) threw 9 outs, 5 ER = 15 ERA against a projected 3.38 ERA (11.62 ERA error)

Bullpen behind him: 1 ER over 12 outs.

The AWAY lineup he faced produced 5 runs on 9 hits and 0 HR, against a projected 4 runs / 7 hits (1 runs vs projection).

Who did the damage to him:

Jonathan Aranda: 2-for-4, 0 HR, 1 RBI [+0.78 runs vs projection] season xwOBA 0.35, elite - real projection error
Chandler Simpson: 2-for-5, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.261, weak - a good day, not a model error
Jonny DeLuca: 2-for-5, 0 HR, 1 RBI [+0.47 runs vs projection] season xwOBA 0.268, weak - a good day, not a model error

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

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
TB	Ian Seymour (Starter)	9 / 16	9 / 5.06 (+3.94)	3 / 3	5 / 0	2 / 0	1 / 2	2 / 4	venue away: 3 vs 5.35 ERA (ok) · opp run factor 1 (weak) · 42 prior appearances this season
TB	Kevin Kelly (Setup Man)	2 / 3	0 / 0	0 / 0	1 / 1	0 / 0	0 / 0	1 / 0
TEX	MacKenzie Gore (Starter)	9 / 16	15 / 3.38 (+11.62)	5 / 2	7 / 2	0 / 1	3 / 2	1 / 6	venue home: 2.89 vs 5.62 ERA (ok) · opp run factor 1.02 (average) · 29 prior appearances this season
TEX	Jacob Latz (Closer)	3 / 3	0 / 0	0 / 0	1 / 1	0 / 0	0 / 0	1 / 0
TEX	Jakob Junis (Setup Man)	3 / 3	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 3
TEX	Tyler Alexander (Middle Reliever (High Leverage))	3 / 3	9 / 0 (+9.00)	1 / 0	2 / 1	0 / 0	0 / 0	1 / 0
TEX	Chase Silseth (Middle Reliever (High Leverage))	3 / 2	0 / 40.5 (-40.50)	0 / 3	0 / 4	0 / 0	0 / 1	1 / 2

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
TEX	Jake Burger (1B)	4 / 4	2 / 1	0 / 0	0 / 0	2 / 0	3 / 1	3 / 1	1 / 0	2 / 1	2.66	0.317 average	no signal
TB	Junior Caminero (3B)	2 / 5	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	2 / 1	0 / 1	-1.07	0.362 elite	bad day
TEX	Logan O'Hoppe (DH)	3 / 3	2 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	0 / 0	0 / 1	0.94	0.268 weak	good day
TB	Jonathan Aranda (1B)	4 / 4	2 / 1	1 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0 / 0	0 / 1	0.78	0.35 elite	projection too low
TEX	Corey Seager (SS)	5 / 4	1 / 1	0 / 0	0 / 0	1 / 0	1 / 1	1 / 1	0 / 1	0 / 1	0.6	0.35 elite	projection too low
TB	Chandler Simpson (LF)	5 / 5	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	0 / 0	0 / 0	0.47	0.261 weak	—
TB	Jonny DeLuca (CF)	5 / 3	2 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 1	0.47	0.268 weak	—
TEX	Elias Díaz (C)	3 / 3	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	1 / 1	0.47	0.275 weak	—
TEX	Cody Freeman (3B)	1 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	-0.47	0.232 weak	—
TEX	Ezequiel Duran (CF)	5 / 4	3 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	0 / 1	2 / 1	-0.32	0.307 average	—
TEX	Justin Foscue (2B)	4 / 3	1 / 1	1 / 0	0 / 0	0 / 0	2 / 0	0 / 0	0 / 0	0 / 1	0.31	0.334 above average	—
TEX	Brandon Nimmo (RF)	5 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 1	-0.31	0.353 elite	—
TB	Ryan Vilade (2B)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	2 / 0	1 / 1	0.19	0.335 above average	—
TB	Yandy Díaz (DH)	5 / 5	2 / 1	0 / 1	0 / 0	0 / 0	1 / 1	1 / 1	0 / 1	1 / 0	-0.17	0.358 elite	—
TEX	Wyatt Langford (LF)	2 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	1 / 1	3 / 1	1 / 1	-0.12	0.314 average	—
TB	Jorge Mateo (SS)	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	1 / 1	0	0.302 below average	—
TB	Nick Fortes (C)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	0 / 0	1 / 0	0	0.274 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.688	positive favors home; negative favors away
Recent form	-0.634	positive favors home; negative favors away
Starting-pitcher matchup	-0.021	positive favors home; negative favors away
Composite	-0.763	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.93-3.71	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	0.527	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	4.93-3.71	5-4	-0.219	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
TEX	Wyatt Langford	batter / LF	deboosted	predicted_RBI 3.72
TB	Ian Seymour	pitcher / Starter	better	predicted_SO 3
TB	Jorge Mateo	batter / SS	better	predicted_RBI 2.88
TB	Yandy Díaz	batter / DH	worse	predicted_SO 2.375
TB	Ryan Vilade	batter / 2B	worse	predicted_AB 2.316
TEX	Corey Seager	batter / SS	worse	predicted_AB -2.12
TB	Chandler Simpson	batter / LF	worse	predicted_AB -1.61
TB	Nick Fortes	batter / C	boosted	predicted_H 1.37

Structural audit findings

Severity	Rule	Finding	Subject
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_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	TB model-prediction batting totals differ from final simulation box score (AB: current=34, sim=36; H: current=8, sim=9; 2B: current=1, sim=2; RBI: current=4, sim=5; R: current=4, sim=5; BB: current=2, sim=3; SO: current=6, sim=11; 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: Ryan Vilade: AB current=3 box=5, RBI current=1 box=0; Chandler Simpson: AB current=5 box=3, SO current=0 box=1, HBP current=0 box=1; Jonny DeLuca: H current=1 box=0, BB current=0 box=1, SO current=1 box=0; Yandy Díaz: H current=1 box=2, 2B current=1 box=0, HR current=0 box=1, BB current=1 box=0; Junior Caminero: AB current=5 box=4, HR current=1 box=0, RBI current=1 box=0, SO current=1 box=0; Jonathan Aranda: AB current=4 box=5, RBI current=1 box=0, SO current=1 box=2; Victor Mesa Jr.: AB current=3 box=4, H current=1 box=0, SO current=1 box=2; Jorge Mateo: H current=0 box=1, 2B current=0 box=1, RBI current=0 box=3, BB current=0 box=1	TB
info	CURRENT_FINALIZED_MATCH_BOX	TEX model-prediction batting totals differ from final simulation box score (AB: current=32, sim=29; H: current=7, sim=2; 2B: current=2, sim=0; HR: current=1, sim=0; BB: current=3, sim=4; SO: current=8, sim=4).	TEX
info	CURRENT_PLAYER_MATCH_BOX	TEX current finalized hitter lines are model predictions and differ from the simulation box score: Ezequiel Duran: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Corey Seager: AB current=4 box=2, H current=1 box=0, RBI current=1 box=0, BB current=1 box=2; Justin Foscue: H current=1 box=0, R current=0 box=1, BB current=0 box=1, SO current=1 box=0; Wyatt Langford: 2B current=1 box=0, RBI current=0 box=4, BB current=1 box=0, SO current=1 box=0; Brandon Nimmo: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Elias Díaz: SO current=1 box=2; Logan O'Hoppe: SO current=1 box=0; Cody Freeman: AB current=3 box=2, R current=0 box=1, BB current=0 box=1	TEX
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

TOR@KCMISS

Predicted: TOR 5-4 (conf -0.07, lean)Actual: KC 1-6Brier 0.288total bias +2

Root cause: Small edge on the losing side

Evidence-backed diagnosis

published system edge 0.07 (lean); this is not the half-scale distance from 50%
team-run absolute error 6; total bias +2
pregame home-minus-away evidence: season -0.604, recent -0.842, pitcher +0.060, composite -0.592
confidence pipeline reached 0.572 before later publication/recency controls; published home probability was 0.463
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

TOR offense over-projected by 4
Randy Dobnak (KC) allowed 3 fewer ER than projected (1 vs 4)

Team	Starter	Projected	Actual	ER err
TOR	Spencer Arrighetti	4 ER / 15 outs	6 ER / 14 outs	2
KC	Randy Dobnak	4 ER / 18 outs	1 ER / 18 outs	-3

Decision internals: run factors home×0.984 / away×0.924; 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:

Kazuma Okamoto (TOR 3B): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Josh Smith (TOR 2B): 1-for-2, 1 HR, 1 RBI (projected 0.0 H, 0.0 HR) [run impact 1.07]
Vinnie Pasquantino (KC 1B): 1-for-3, 1 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 0.95]
Salvador Perez (KC DH): 1-for-4, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Brandon Valenzuela (TOR C): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]

Starter miss: Randy Dobnak (KC) posted 1.50 ERA vs projected 6.00 (1 ER in 18 outs); he is materially worse pitching home (4.44 ERA there vs 1.14 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

Spencer Arrighetti (TOR) threw 14 outs, 6 ER = 11.57 ERA against a projected 7.2 ERA (4.37 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 6 runs on 7 hits and 2 HR, against a projected 4 runs / 7 hits (2 runs vs projection).

Who did the damage to him:

Vinnie Pasquantino: 1-for-3, 1 HR, 3 RBI [+0.95 runs vs projection] season xwOBA 0.328, above average - real projection error
Salvador Perez: 1-for-4, 1 HR, 2 RBI [+0.93 runs vs projection] season xwOBA 0.309, average - a good day, not a model error
Luke Maile: 0-for-2, 0 HR, 0 RBI [+0.66 runs vs projection] season xwOBA 0.196, weak - a good day, not a model error

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

HOME pitching

Randy Dobnak (KC) threw 18 outs, 1 ER = 1.5 ERA against a projected 6 ERA (-4.5 ERA error) only 11 prior appearances in the covered season - the projection had little of his own record to stand on

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 4 runs / 8 hits (-3 runs vs projection).

Who did the damage to him:

Josh Smith: 1-for-2, 1 HR, 1 RBI [+1.07 runs vs projection] season xwOBA 0.287, below average - a good day, not a model error
Brett Bateman: 3-for-3, 0 HR, 0 RBI [+0.61 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

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
TOR	Spencer Arrighetti (Starter)	14 / 15	11.57 / 7.2 (+4.37)	6 / 4	6 / 4	2 / 1	3 / 2	3 / 4	venue away: 5.98 vs 3.38 ERA (worse here) · opp run factor 0.98 (weak) · 20 prior appearances this season
KC	Randy Dobnak (Starter)	18 / 18	1.5 / 6 (-4.50)	1 / 4	4 / 6	1 / 1	0 / 2	3 / 5	venue home: 4.44 vs 1.14 ERA (worse here) · opp run factor 0.92 (weak) · 11 prior appearances this season

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
TOR	Kazuma Okamoto (3B)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	1 / 1	-1.4	0.317 average	no signal
TOR	Josh Smith (2B)	2 / 3	1 / 0	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 1	0 / 1	1.07	0.287 below average	good day
KC	Vinnie Pasquantino (1B)	3 / 4	1 / 1	0 / 1	0 / 0	1 / 0	3 / 1	2 / 1	2 / 1	0 / 1	0.95	0.328 above average	projection too low
KC	Salvador Perez (DH)	4 / 4	1 / 1	0 / 0	0 / 0	1 / 0	2 / 1	1 / 0	0 / 0	0 / 1	0.93	0.309 average	no signal
TOR	Brandon Valenzuela (C)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	-0.8	0.312 average	no signal
TOR	Jesús Sánchez (LF)	2 / 3	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.78	0.324 above average	bad day
KC	Luke Maile (C)	2 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	1 / 1	0.66	0.196 weak	good day
TOR	Brett Bateman (CF)	3 / 5	3 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0.61	0.3 below average	good day
TOR	Vladimir Guerrero Jr. (DH)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 1	-0.47	0.332 above average	—
TOR	Sean Keys (1B)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	2 / 1	-0.47	0.236 weak	—
TOR	Andrés Giménez (SS)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.269 weak	—
KC	Josh Rojas (3B)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	-0.47	0.219 weak	—
KC	Nick Loftin (2B)	4 / 3	2 / 1	0 / 1	0 / 0	0 / 0	1 / 0	0 / 1	0 / 0	1 / 0	0.16	0.297 below average	—
KC	Bobby Witt Jr. (SS)	4 / 5	2 / 1	1 / 0	0 / 0	0 / 1	0 / 1	2 / 1	1 / 1	1 / 1	-0.15	0.385 elite	—
KC	Isaac Collins (LF)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	1 / 1	-0.14	0.296 below average	—
TOR	Nathan Lukes (RF)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 0	0	0.289 below average	—
KC	John Rave (RF)	5 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	2 / 1	0	0.316 average	—
KC	Kyle Isbel (CF)	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.262 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.604	positive favors home; negative favors away
Recent form	-0.842	positive favors home; negative favors away
Starting-pitcher matchup	0.06	positive favors home; negative favors away
Composite	-0.592	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	5.1-4.65	n/a	Base ML historical model prediction
variance_applied	5-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	0.537	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by starter_quality_favours_the_other_side, 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	5.1-4.65	5-5	-0.452	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-5	5-4	1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by starter_quality_favours_the_other_side, 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
KC	John Rave	batter / RF	better	predicted_RBI 2.63
TOR	Nathan Lukes	batter / RF	better	predicted_RBI 2.48
KC	Vinnie Pasquantino	batter / 1B	worse	predicted_SO 2.347
TOR	Jesús Sánchez	batter / LF	worse	predicted_AB 2.232
TOR	Josh Smith	batter / 2B	better	predicted_H 1.8
KC	Bobby Witt Jr.	batter / SS	worse	predicted_AB -1.55
TOR	Vladimir Guerrero Jr.	batter / DH	worse	predicted_SO 1.465
TOR	Kazuma Okamoto	batter / 3B	deboosted	predicted_AB -1.43

Structural audit findings

Severity	Rule	Finding	Subject
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_PITCHER_DUPLICATION	TOR pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	TOR
info	CURRENT_FINALIZED_MATCH_BOX	KC model pitcher projections differ from final simulation pitching box score (H: current=7, sim=9).	KC
info	CURRENT_FINALIZED_MATCH_BOX	KC model-prediction batting totals differ from final simulation box score (AB: current=31, sim=32; H: current=7, sim=5; 2B: current=2, sim=0; RBI: current=3, sim=4; BB: current=3, sim=2).	KC
info	CURRENT_PLAYER_MATCH_BOX	KC current finalized hitter lines are model predictions and differ from the simulation box score: John Rave: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=3, SO current=1 box=0; Vinnie Pasquantino: AB current=4 box=3, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Nick Loftin: AB current=3 box=4, H current=1 box=0, 2B current=1 box=0, R current=1 box=0; Bobby Witt Jr.: AB current=5 box=3, HR current=1 box=0, RBI current=1 box=0; Salvador Perez: H current=1 box=2, R current=0 box=1; Josh Rojas: AB current=3 box=4, R current=0 box=1; Luke Maile: SO current=1 box=0; Isaac Collins: AB current=3 box=4, H current=1 box=0, BB current=1 box=0	KC
info	CURRENT_FINALIZED_MATCH_BOX	TOR model-prediction batting totals differ from final simulation box score (AB: current=32, sim=36; H: current=8, sim=9; 2B: current=1, sim=4; HR: current=1, sim=2; RBI: current=4, sim=5; R: current=4, sim=5; HBP: current=0, sim=1).	TOR
info	CURRENT_PLAYER_MATCH_BOX	TOR current finalized hitter lines are model predictions and differ from the simulation box score: Vladimir Guerrero Jr.: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=0 box=1; Jesús Sánchez: AB current=3 box=5, SO current=1 box=2; Sean Keys: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Brett Bateman: H current=1 box=2, 2B current=0 box=1, BB current=1 box=0, SO current=1 box=2; Nathan Lukes: H current=1 box=2, 2B current=0 box=1, HR current=0 box=1, RBI current=1 box=3; Kazuma Okamoto: AB current=4 box=3, H current=1 box=0, HR current=1 box=0, BB current=0 box=1; Josh Smith: AB current=3 box=4, H current=0 box=2, HR current=0 box=1, RBI current=0 box=1; Brandon Valenzuela: AB current=3 box=4, H current=1 box=2, 2B current=0 box=1, R current=0 box=1	TOR
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

WAS@LADHIT

Predicted: LAD 4-5 (conf +0.06, lean)Actual: LAD 5-7Brier 0.221total bias -3

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 -3
pregame home-minus-away evidence: season +0.789, recent +0.414, pitcher -0.082, composite +0.537
confidence pipeline reached 0.577 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

total runs under-projected by 3 (9 vs 12)
Justin Wrobleski (LAD) allowed 3 more ER than projected (4 vs 1)

Team	Starter	Projected	Actual	ER err
WAS	Andrew Alvarez	2 ER / 15 outs	3 ER / 13 outs	1
LAD	Justin Wrobleski	1 ER / 16 outs	4 ER / 9 outs	3

Decision internals: run factors home×1.022 / away×1.105; 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:

Will Smith (LAD C): 1-for-3, 0 HR, 1 RBI (projected 0.0 H, 0.0 HR) [run impact 1.13]
Harry Ford (WAS C): 1-for-2, 0 HR, 2 RBI (projected 0.0 H, 0.0 HR) [run impact 1.11]
Tommy Edman (LAD CF): 2-for-4, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.11]
Daylen Lile (WAS LF): 2-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 1.09]
CJ Abrams (WAS 2B): 1-for-4, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]

Starter miss: Justin Wrobleski (LAD) posted 12.00 ERA vs projected 1.69 (4 ER in 9 outs); he faced a strong offense (opponent run factor 1.10, 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

Andrew Alvarez (WAS) threw 13 outs, 3 ER = 6.23 ERA against a projected 3.6 ERA (2.63 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 6 runs on 9 hits and 1 HR, against a projected 5 runs / 8 hits (1 runs vs projection).

Who did the damage to him:

Will Smith: 1-for-3, 0 HR, 1 RBI [+1.13 runs vs projection] season xwOBA 0.38, elite - real projection error
Tommy Edman: 2-for-4, 0 HR, 1 RBI [+1.11 runs vs projection] season xwOBA 0.316, average - a good day, not a model error
Max Muncy: 3-for-4, 0 HR, 0 RBI [+0.92 runs vs projection] season xwOBA 0.349, elite - real projection error

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

HOME pitching

Justin Wrobleski (LAD) threw 9 outs, 4 ER = 12 ERA against a projected 1.69 ERA (10.31 ERA error)

Bullpen behind him: 0 ER over 6 outs.

The AWAY lineup he faced produced 5 runs on 9 hits and 1 HR, against a projected 5 runs / 8 hits (0 runs vs projection).

Who did the damage to him:

Harry Ford: 1-for-2, 0 HR, 2 RBI [+1.11 runs vs projection] season xwOBA 0.285, below average - a good day, not a model error
Daylen Lile: 2-for-4, 0 HR, 0 RBI [+1.09 runs vs projection] season xwOBA 0.33, above average - real projection error
CJ Abrams: 1-for-4, 1 HR, 1 RBI [+0.93 runs vs projection] season xwOBA 0.329, above average - real projection error

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

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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
WAS	Andrew Alvarez (Starter)	13 / 15	6.23 / 3.6	3 / 2	4 / 3	0 / 0	2 / 1	5 / 5	venue away: 3.48 vs 3.26 ERA (ok) · opp run factor 1.02 (strong) · 21 prior appearances this season
LAD	Justin Wrobleski (Starter)	9 / 16	12 / 1.69 (+10.31)	4 / 1	8 / 2	1 / 0	1 / 4	6 / 7	venue home: 2.96 vs 3.86 ERA (ok) · opp run factor 1.1 (strong) · 22 prior appearances this season
LAD	Brock Stewart (Middle Reliever (High Leverage))	3 / 4	0 / 0	0 / 0	0 / 1	0 / 0	1 / 0	0 / 2
LAD	Edgardo Henriquez (Setup Man)	3 / 3	0 / 18 (-18.00)	0 / 2	0 / 2	0 / 0	1 / 1	2 / 0

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
LAD	Will Smith (C)	3 / 3	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	2 / 0	1 / 1	1.13	0.38 elite	projection too low
WAS	Harry Ford (C)	2 / 3	1 / 0	1 / 0	0 / 0	0 / 0	2 / 0	2 / 0	2 / 1	0 / 1	1.11	0.285 below average	good day
LAD	Tommy Edman (CF)	4 / 4	2 / 1	1 / 0	0 / 0	0 / 0	1 / 1	2 / 1	1 / 0	1 / 1	1.11	0.316 average	no signal
WAS	Daylen Lile (LF)	4 / 4	2 / 1	0 / 0	1 / 0	0 / 0	0 / 1	1 / 1	0 / 0	2 / 1	1.09	0.33 above average	projection too low
WAS	CJ Abrams (2B)	4 / 4	1 / 1	0 / 0	0 / 0	1 / 0	1 / 1	1 / 1	1 / 1	2 / 1	0.93	0.329 above average	projection too low
LAD	Max Muncy (DH)	4 / 3	3 / 1	1 / 0	0 / 0	0 / 0	0 / 0	2 / 0	0 / 1	0 / 1	0.92	0.349 elite	projection too low
LAD	Kyle Tucker (RF)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	-0.8	0.326 above average	bad day
WAS	Dylan Crews (RF)	4 / 5	0 / 1	0 / 1	0 / 0	0 / 0	1 / 0	0 / 1	0 / 0	2 / 1	-0.78	0.307 average	no signal
LAD	Miguel Rojas (2B)	2 / 3	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	-0.78	0.309 average	no signal
LAD	Freddie Freeman (1B)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	-0.64	0.372 elite	bad day
WAS	Nasim Nuñez (SS)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.266 weak	—
WAS	Jacob Young (CF)	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0 / 0	0 / 1	0.47	0.32 average	—
LAD	Enrique Hernández (3B)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.258 weak	—
LAD	Teoscar Hernández (LF)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	1 / 0	1 / 1	0.33	0.325 above average	—
LAD	Mookie Betts (SS)	5 / 5	1 / 1	0 / 0	0 / 0	1 / 1	4 / 1	1 / 1	0 / 1	0 / 0	-0.33	0.33 above average	—
WAS	Brady House (3B)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 0	0 / 0	1 / 1	-0.31	0.336 above average	—
WAS	James Wood (DH)	2 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	3 / 1	0 / 1	-0.27	0.417 elite	—
WAS	Andrés Chaparro (1B)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0	0.348 elite	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.789	positive favors home; negative favors away
Recent form	0.414	positive favors home; negative favors away
Starting-pitcher matchup	-0.082	positive favors home; negative favors away
Composite	0.537	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.35-4.07	n/a	Base ML historical model prediction
variance_applied	4-4	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, season_quality_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	4.35-4.07	4-4	0.276	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-4	4-5	1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail, season_quality_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
WAS	Andrés Chaparro	batter / 1B	worse	predicted_AB 2.5
LAD	Freddie Freeman	batter / 1B	worse	predicted_SO 2.294
WAS	Dylan Crews	batter / RF	worse	predicted_RBI 1.54
LAD	Mookie Betts	batter / SS	worse	predicted_SO 1.524
LAD	Max Muncy	batter / DH	better	predicted_RBI 1.469
WAS	Harry Ford	batter / C	better	predicted_H 1.42
WAS	James Wood	batter / DH	worse	predicted_SO -1.342
LAD	Kyle Tucker	batter / RF	better	predicted_H 1.25

Structural audit findings

Severity	Rule	Finding	Subject
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_PITCHER_DUPLICATION	WAS pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	WAS
info	CURRENT_FINALIZED_MATCH_BOX	LAD model-prediction batting totals differ from final simulation box score (AB: current=33, sim=29; H: current=8, sim=6; 2B: current=2, sim=3; 3B: current=0, sim=1; BB: current=4, sim=1; SO: current=7, sim=8; SF: current=0, sim=1).	LAD
info	CURRENT_PLAYER_MATCH_BOX	LAD current finalized hitter lines are model predictions and differ from the simulation box score: Tommy Edman: 2B current=0 box=1, SO current=1 box=0; Will Smith: AB current=3 box=4, R current=1 box=0, SO current=1 box=0; Teoscar Hernández: AB current=4 box=3, 2B current=0 box=1, RBI current=1 box=0; Freddie Freeman: 2B current=1 box=0, RBI current=1 box=0, BB current=1 box=0, SO current=1 box=3; Max Muncy: AB current=3 box=2, HR current=0 box=1, RBI current=0 box=2, R current=0 box=1; Mookie Betts: AB current=5 box=4, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Kyle Tucker: AB current=4 box=3, H current=1 box=2, 2B current=0 box=1, 3B current=0 box=1; Enrique Hernández: H current=1 box=0	LAD
info	CURRENT_FINALIZED_MATCH_BOX	WAS model-prediction batting totals differ from final simulation box score (AB: current=33, sim=34; 2B: current=2, sim=1; 3B: current=0, sim=1; HR: current=1, sim=0; R: current=5, sim=4; BB: current=3, sim=6; SF: current=0, sim=1).	WAS
info	CURRENT_PLAYER_MATCH_BOX	WAS current finalized hitter lines are model predictions and differ from the simulation box score: James Wood: AB current=4 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; CJ Abrams: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, SO current=1 box=2; Daylen Lile: AB current=4 box=3, 3B current=0 box=1, BB current=0 box=1, SO current=1 box=0; Andrés Chaparro: AB current=3 box=5; Brady House: AB current=4 box=5, H current=1 box=2, RBI current=1 box=0, R current=0 box=1; Nasim Nuñez: H current=1 box=0, BB current=0 box=1; Dylan Crews: H current=1 box=2, 2B current=1 box=0, RBI current=0 box=2, SO current=1 box=2; Harry Ford: AB current=3 box=2, H current=0 box=2, RBI current=0 box=1, R current=0 box=1	WAS
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.

Luna deep post-mortem (full)

> Coverage receipt: COMPLETE · 9/8 sections; 15/15 games; 4/4 causes; 6/6 sources

1. Evidence receipt and gaps

Sources received

Six of seven manifest entries were loaded:

Prediction review — /app/output/reports/prediction_review_20260906.json, 377,236 bytes; official outcomes and scoring errors.
Forensic deep audit — /app/output/reports/PREDICTION_SYSTEM_DEEP_AUDIT_20260906.json, 240,848 bytes; probability, provenance, starter, and decision contracts.
Continuous learning — /app/output/learning/continuous_learning_daily_20260906.json, 2,489,405 bytes; pipeline quality, calibration, lineage, score distribution, and starter comparisons.
Detailed report — /app/output/reports/detailed_20260906.html, 349,460 bytes.
Ultra-detailed player audit — /app/output/reports/20260906_ultra_detailed_player_audit.html, 30,202,030 bytes.
Structural audit — /app/output/reports/audit_report_20260906.html, 1,864,642 bytes.
Missing: /app/output/reports/audit_report_20260906_integrity_revision.html, purpose “post-Luna remediation proof for immutable revision artifacts.” Its status is explicitly missing.

Coverage

All 15 expected games were reviewed:

ARI@HOU, ATL@PHI, BOS@BAL, CHC@MIA, DET@CLE, LAA@PIT, MIL@CIN, MIN@CHW, NYY@SD, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC, and WAS@LAD.

Coverage counts are complete for the supplied artifacts:

Canonical artifacts loaded: 15/15
Continuous-learning games: 15/15
Deep-audit games: 15/15
Starter comparisons: 15/15
Structural audits: 15/15
All games structurally audited: true

Gaps and contradictions

The integrity-revision audit is missing. This is a provenance/artifact gap, although the supplied per-game structural audits all passed.
The player feed had no rows for slate 20260906. Therefore, player-level adjustments describe forecast changes, but actual player-level causal attribution is unavailable for every game.
No source-labelled moneyline was available: market cross-check coverage was 0/15.
No matchup-pattern breakdown was available.
Probability-lineage text is internally inconsistent in four entries:
ARI@HOU
ATL@PHI
LAA@PIT
OAK@SEA

Each says "no recorded transformation explains the publication delta" while also reporting unexplained_material_delta: false. The deltas are small, but the wording and flag do not agree.

The causal register states that all eight material probability movements were explained by score-coherence and/or recency transformations. That is compatible with the four small deltas being below the materiality threshold, but the lineage text should explicitly say “below material threshold,” not “no recorded transformation.”
The high-raw cluster threshold was 0.65, but the only listed calibrated probability was 0.5833. The row is not itself above that threshold; it appears to refer to a high raw model probability of 0.652. The cluster definition should distinguish raw from calibrated probabilities.

2. Executive diagnosis

Winner accuracy

The slate went 9–6, 60.0%, versus a walk-forward baseline of 56.0% and model-validation winner accuracy of 59.6244%. This is a positive result, but only modestly above baseline and not evidence that the score forecasts were reliable.

The winner model was correct in:

ARI@HOU, BOS@BAL, DET@CLE, LAA@PIT, MIN@CHW, OAK@SEA, SF@NYM, STL@COL, and WAS@LAD.

It missed:

ATL@PHI, CHC@MIA, MIL@CIN, NYY@SD, TB@TEX, and TOR@KC.

Calibration quality

Calibration was mixed and materially imperfect:

Weighted absolute calibration error: 0.0937
0.50–0.55 bucket: average confidence 0.5325, observed hit rate 0.6364, calibration error −0.1039 — underconfident.
0.55–0.60 bucket: average confidence 0.5656, observed hit rate 0.5000, calibration error +0.0656 — overconfident.
High-conviction miss rate: 0.5, with 2 misses among 4 high-conviction games.

The published confidence was therefore not uniformly conservative. The lower-confidence bucket was underconfident, but the higher-confidence bucket and high-conviction subset were too aggressive. This supports bounded confidence-cap work, not a blanket conclusion that every probability was overconfident.

Score accuracy

Score accuracy was the principal weakness:

Holdout R²:
Away runs: 0.018161
Home runs: 0.007377
Total runs: 0.014682
Model-margin MAE: 2.667 runs
Model-margin signed bias: −2.0, meaning actual margins exceeded predicted margins on average.
All 15 predicted margins were 1 or 2 runs.
Four actual margins were at least 4 runs: CHC@MIA, MIL@CIN, MIN@CHW, and TOR@KC, with actual margins of 7, 4, 9, and 5.

The registered weak_run_model_signal issue is therefore substantiated. The model beat average guessing, but its exact run estimates and tails were weak. The small positive R² must not be interpreted as reliable score prediction.

Starter accuracy

Starter projections were highly variable. Identity matching was successful for all listed starters, but actual earned runs often diverged materially from projection. Examples include:

Clay Holmes: projected 2 ER, actual 8 ER.
Cesar Perdomo: projected 6 ER, actual 0 ER.
Bailey Ober: projected 5 ER, actual 1 ER.
Justin Wrobleski: projected 1 ER, actual 4 ER.
Kyle Leahy: projected 3 ER, actual 5 ER.

The lever-tension receipt records the pitcher-weighting lever as helping 10 games and hurting 2. That is useful directional evidence, but not a controlled causal result. Cross-slate evidence does not support changing context.opponent_strength_weight: starter ERA error has the same sign on only 4 of 7 measured slates, with z = 0.65.

Artifact integrity

Artifact integrity was strong for the delivered slate:

Structural audit summary: 0 critical, 0 errors, 0 warnings, 114 informational findings
Every per-game structural audit passed.
All game artifacts passed critical/error gates and were eligible for learning.
No winner probability reversion or overwrite was detected: 15/15 blend sides survived to publication, with zero blend flips and zero overwritten games.
Simulation produced zero winner flips, but that is a persistent functional concern, not evidence that the simulation was useful.

The missing integrity-revision audit remains a documentation gap. It does not invalidate the supplied structural passes, but it prevents a fully closed artifact chain.

Causal issue register reconciliation

High: `weak_run_model_signal`

Evidence: home/away/total R² = 0.0074/0.0182/0.0147.
Affected games: all 15.
Action: treat exact runs as low-confidence; investigate stronger features or a better target through isolated training backtests.
Guardrail: a small positive R² only shows performance above average guessing; it does not make exact scores reliable.
Assessment: confirmed and escalated across 10 consecutive slates. This is a material model-risk issue, not a contract failure.

Info: `complete_slate_audit_coverage`

Evidence: 15/15 games structurally audited.
Affected games: all 15.
Action: retain per-game audit receipts.
Guardrail: full coverage does not cancel substantive findings inside audited games.
Assessment: confirmed as coverage, not as a clean-slate conclusion.

Info: `probability_publication_delta_explained`

Evidence: eight material movements reportedly explained by score coherence and/or recency.
Affected games: CHC@MIA, DET@CLE, MIL@CIN, MIN@CHW, SF@NYM, STL@COL, TOR@KC, WAS@LAD.
Action: continue exposing each probability layer and transformation.
Guardrail: a published probability above calibrated probability is not a governor bypass if a named transformation explains it.
Assessment: generally supported, but lineage wording should be corrected for the four small unexplained-text deltas.

Info: `simulation_inert_on_slate`

Evidence: zero flip candidates and zero published flips.
Affected games: all 15.
Action: measure simulation contribution over rolling slates and separate score realization from winner flipping.
Guardrail: zero winner flips is not automatically a defect and does not prove no score work occurred.
Assessment: escalated to a persistent, established failure over 10 consecutive slates because the supplied self-assessment identifies a likely early-exit/no-op mechanism.

3. Every game

Conventions: Published probability is the probability of the published side, while the raw field published_pick_probability is the home probability. “Edge” is home minus away. Team-run error is the absolute team-run error. Total bias is actual total minus predicted total; positive means the model underpredicted total runs.

All player actuals were unavailable because the slate player feed had no rows.

ARI@HOU — correct winner, wrong score

Prediction: ARI 4, HOU 3; ARI winner.
Actual: ARI 3, HOU 2; ARI winner.
Published probability: ARI 0.5373; home probability 0.463.
Confidence edge: home minus away −0.0746; lean, away side.
Brier / team-run error / total bias: 0.214084 / 2 / +2.
Starter lines:
Eduardo Rodriguez: 2.0 projected ER, 18 outs; 1 ER, 18 outs; ER error −1.
Peter Lambert: 2.0 projected ER, 16 outs; 3 ER, 21 outs; ER error +1.
Edges: season −0.026, recent +0.420, pitcher +0.122, composite +0.180 for HOU; recent and pitcher factors partially offset the season edge.
Controls: Savant adjustment applied; starter-experience shrinkage not applied; score volatility no_change.
Score path: 4–3 historical/base direction, ultimately published 4–3. Confidence reached 0.544 before controls; publication was 0.463 home.
Player adjustments: none available.
Audit: structural PASS; 0 defects, 8 informational notes. Forecast artifact was not causal.
Primary/secondary cause: none; correct call. This is a correct-winner/wrong-score game.

ATL@PHI — wrong winner

Prediction: ATL 3, PHI 4; PHI winner.
Actual: ATL 5, PHI 4; ATL winner.
Published probability: PHI 0.550646; home probability 0.551.
Confidence edge: +0.1013, moderate PHI edge.
Brier / team-run error / total bias: 0.303211 / 2 / −2.
Starters:
Tyler Mahle: 2 projected ER, 17 outs; 0 ER, 18 outs; error −2.
Aaron Nola: 3 projected ER, 16 outs; 1 ER, 21 outs; error −2.
Edges: season −0.391, recent −0.010, pitcher −0.101, composite −0.260 for PHI; evidence actually leaned ATL.
Controls: Savant applied; no starter-experience shrinkage; no score-volatility change.
Score path: historical 4.44–4.49, variance 4–4, enhanced 3–4, then final 3–4.
Player adjustments: 22 material changes, including Trea Turner projected RBI +2.55, Michael Harris II projected AB −1.89, and Brandon Marsh projected hits +1.70.
Audit: PASS; 0 defects, 9 info.
Causes: primary wrong_conviction_side; secondary none. The model committed to PHI despite season and composite edges favoring ATL. The score was close, but the winner direction was wrong.

BOS@BAL — correct winner, wrong score

Prediction: BOS 4, BAL 3; BOS winner.
Actual: BOS 3, BAL 1; BOS winner.
Published probability: BOS 0.54926; home probability 0.451.
Confidence edge: −0.0985, lean BOS.
Brier / team-run error / total bias: 0.203167 / 3 / +3.
Starters:
Payton Tolle: 0 projected ER, 18 outs; 1 ER, 18 outs; error +1.
Kyle Bradish: 3 projected ER, 16 outs; 0 ER, 12 outs; error −3.
Edges: season −0.555, recent −0.178, pitcher −0.066, composite −0.675 for BOS.
Controls: starter quality moved score toward BOS; recency guardrail applied; no score-volatility change.
Score path: 4–3 final from a 4–3-type low-margin path.
Player adjustments: none available.
Audit: PASS; 0 defects, 7 info.
Causes: no winner cause; correct winner. Score miss was primarily run-environment/starter realization, especially Bradish allowing 0 rather than 3 ER. This is a correct-winner/wrong-score game.

CHC@MIA — wrong winner and major score miss

Prediction: CHC 6, MIA 4; CHC winner.
Actual: CHC 3, MIA 10; MIA winner.
Published probability: MIA 0.420; CHC published-side probability 0.580.
Confidence edge: −0.1600, moderate CHC edge. This is not a near coin flip; it is a moderate 0.16 edge.
Brier / team-run error / total bias: 0.336400 / 9 / −3.
Starters:
Clay Holmes: 2 projected ER, 18 outs; 8 ER, 13 outs; error +6.
Tyler Phillips: 4 projected ER, 15 outs; 3 ER, 15 outs; error −1.
Edges: season −0.841, recent −0.858, pitcher −0.079, composite −1.211 for CHC.
Controls: high-total environment lift and supported blowout expansion; recency guardrail; no starter-experience shrinkage.
Score path: historical 4.17–3.69, variance 4–4, enhanced 6–4, then final 6–4. The model amplified the CHC side after recent-form and H2H processing.
Player adjustments: 21 material changes, including Owen Caissie RBI +2.62 and Michael Busch RBI +2.44; no actual player feed to verify them.
Audit: PASS; 0 defects, 8 info.
Causes: primary starter_projection; secondary wrong_conviction_side. Holmes’ six-ER projection miss was the earliest clear game-specific failure. The model also gave a strong CHC conviction despite MIA’s actual offensive breakout.

DET@CLE — correct winner, wrong score

Prediction: DET 3, CLE 4; CLE winner.
Actual: DET 2, CLE 3; CLE winner.
Published probability: CLE 0.520296; home probability 0.520.
Confidence edge: +0.0406, lean CLE.
Brier / team-run error / total bias: 0.230116 / 2 / +2.
Starters:
Jackson Jobe: 1 projected ER, 15 outs; 2 ER, 15 outs; error +1.
Gavin Williams: 2 projected ER, 16 outs; 2 ER, 18 outs; error 0.
Edges: season −0.130, recent +0.434, pitcher +0.005, composite +0.229 for CLE.
Controls: inexperienced-starter shrinkage from 0.545249 to 0.528092, shrink factor 0.6208; winner unchanged; no volatility change.
Score path: 3–4 final.
Player adjustments: none available.
Audit: PASS; 0 defects, 7 info.
Causes: no winner cause; correct call. Score error was small and consistent with run variance. Correct-winner/wrong-score.

LAA@PIT — correct winner, wrong score

Prediction: LAA 3, PIT 5; PIT winner.
Actual: LAA 0, PIT 1; PIT winner.
Published probability: PIT 0.572315; home probability 0.572.
Confidence edge: +0.1446, moderate PIT edge.
Brier / team-run error / total bias: 0.182914 / 7 / +7.
Starters:
Walbert Ureña: 4 projected ER, 17 outs; 1 ER, 24 outs; error −3.
Paul Skenes: 0 projected ER, 17 outs; 0 ER, 16 outs; error 0.
Edges: season +0.719, recent +0.432, pitcher −0.131, composite +0.575 for PIT.
Controls: Savant applied; no experience shrinkage; no volatility change.
Score path: 3–5 from a PIT-favored score path, but both offenses were heavily overprojected.
Player adjustments: none available.
Audit: PASS; 0 defects, 8 info.
Causes: winner correct; score miss was a run-environment and starter-outperformance issue, especially Ureña allowing only 1 rather than 4 ER. Correct-winner/wrong-score.

MIL@CIN — wrong winner and severe total miss

Prediction: MIL 7, CIN 5; MIL winner.
Actual: MIL 8, CIN 12; CIN winner.
Published probability: MIL 0.530; home probability 0.470.
Confidence edge: −0.0600, lean MIL.
Brier / team-run error / total bias: 0.280900 / 8 / −8.
Starters:
Kyle Harrison: 2 projected ER, 15 outs; 4 ER, 11 outs; error +2.
Brady Singer: 5 projected ER, 15 outs; 6 ER, 15 outs; error +1.
Edges: season −1.480, recent −0.258, pitcher −0.023, composite −1.289 for MIL.
Controls: supported blowout-margin expansion; recency guardrail; large Savant run-environment swing.
Score path: historical 5–3, variance 5–3, enhanced 7–5, final 7–5. The model correctly raised total scoring somewhat but still missed by eight runs.
Player adjustments: 21 material changes, including Dane Myers hits +2.40 and Elly De La Cruz hits +1.892 for CIN.
Audit: PASS; 0 defects, 7 info.
Causes: primary run_environment_under. Both clubs scored materially above forecast, especially CIN, which was underprojected by 7. The winner error followed the run-tail failure.

MIN@CHW — correct winner, wrong score

Prediction: MIN 4, CHW 6; CHW winner.
Actual: MIN 1, CHW 10; CHW winner.
Published probability: CHW 0.530; home probability 0.530.
Confidence edge: +0.0600, lean CHW.
Brier / team-run error / total bias: 0.220900 / 7 / −1.
Starters:
Bailey Ober: 5 projected ER, 16 outs; 1 ER, 13 outs; error −4.
Bryan Hudson: 0 projected ER, 15 outs; 0 ER, 5 outs; error 0.
Edges: season +0.372, recent −0.018, pitcher −0.011, composite +0.537 for CHW.
Controls: recency guardrail; high-raw cluster row; no score-volatility change. Raw pick probability was 0.652, calibrated 0.5833, published 0.530.
Score path: 4–6 final despite actual 1–10.
Player adjustments: none available.
Audit: PASS; 0 defects, 8 info.
Causes: winner correct; score failure came from offense/run realization and Ober outperforming the 5-ER projection. Correct-winner/wrong-score.

NYY@SD — wrong winner

Prediction: NYY 4, SD 3; NYY winner.
Actual: NYY 3, SD 4; SD winner.
Published probability: NYY 0.535664; home probability 0.464.
Confidence edge: −0.0713, lean NYY.
Brier / team-run error / total bias: 0.286936 / 2 / 0.
Starters:
Gerrit Cole: 3 projected ER, 19 outs; 4 ER, 21 outs; error +1.
Michael King: 2 projected ER, 18 outs; 0 ER, 18 outs; error −2.
Edges: season −0.776, recent −0.628, pitcher −0.026, composite −0.820 for NYY.
Controls: recency guardrail; no volatility change.
Score path: historical 4–4, enhanced 4–3, final 4–3; winner direction regressed toward NYY at the enhanced stage.
Player adjustments: 23 material changes, including Gerrit Cole strikeout +3 and Michael King strikeout +2; no actual player-level validation.
Audit: PASS; 0 defects, 7 info.
Causes: primary mixed. The model had strong season/recent NYY evidence, but King’s shutout performance and a one-run game produced the reversal. This is not principally a run-total miss; it is a close winner-direction miss.

OAK@SEA — correct winner, very wrong score

Prediction: OAK 6, SEA 7; SEA winner.
Actual: OAK 0, SEA 2; SEA winner.
Published probability: SEA 0.5442; home probability 0.544.
Confidence edge: +0.0884, lean SEA.
Brier / team-run error / total bias: 0.207754 / 11 / +11.
Starters:
Gage Jump: 4 projected ER, 14 outs; 2 ER, 12 outs; error −2.
Bryan Woo: 3 projected ER, 17 outs; 0 ER, 24 outs; error −3.
Edges: season +0.352, recent −0.706, pitcher +0.011, composite +0.222 for SEA.
Controls: no volatility change.
Score path: 6–7 final; actual scoring was 0–2.
Player adjustments: none available.
Audit: PASS; 0 defects, 7 info.
Causes: winner correct; extreme run-environment overprojection, especially both offenses and Woo’s dominance. Correct-winner/wrong-score.

SF@NYM — correct winner, wrong score

Prediction: SF 4, NYM 6; NYM winner.
Actual: SF 2, NYM 4; NYM winner.
Published probability: NYM 0.516328; home probability 0.516.
Confidence edge: +0.0327, toss-up.
Brier / team-run error / total bias: 0.233939 / 4 / +4.
Starters:
Cesar Perdomo: 6 projected ER, 12 outs; 0 ER, 18 outs; error −6.
Christian Scott: 3 projected ER, 15 outs; 2 ER, 14 outs; error −1.
Edges: season +0.029, recent −0.106, pitcher +0.024, composite +0.402 for NYM.
Controls: inexperienced-starter shrinkage from 0.547531 to 0.51921, factor 0.4042; winner unchanged; no volatility change.
Score path: 4–6 final.
Player adjustments: 23 material changes, including pitcher strikeout adjustments; no actual player feed.
Audit: PASS; 0 defects, 9 info.
Causes: winner correct in a toss-up; score error was driven chiefly by Perdomo’s six-ER projection miss. Correct-winner/wrong-score.

STL@COL — correct winner, wrong score

Prediction: STL 5, COL 4; STL winner.
Actual: STL 10, COL 8; STL winner.
Published probability: STL 0.55935; home probability 0.441.
Confidence edge: −0.1187, moderate STL edge. A 0.12 edge is moderate, not a near coin flip.
Brier / team-run error / total bias: 0.194172 / 9 / −9.
Starters:
Kyle Leahy: 3 projected ER, 16 outs; 5 ER, 8 outs; error +2.
Tanner Gordon: 2 projected ER, 15 outs; 5 ER, 14 outs; error +3.
Edges: season −0.729, recent −0.186, pitcher −0.001, composite −0.628 for STL.
Controls: recency guardrail; no volatility change.
Score path: 5–4 final despite an 18-run actual total.
Player adjustments: none available.
Audit: PASS; 0 defects, 7 info.
Causes: winner correct; both starters and both offenses exceeded the score environment. Correct-winner/wrong-score.

TB@TEX — wrong winner and total miss

Prediction: TB 5, TEX 4; TB winner.
Actual: TB 6, TEX 8; TEX winner.
Published probability: TB 0.527394; home probability 0.473.
Confidence edge: −0.0548, lean TB.
Brier / team-run error / total bias: 0.278144 / 5 / −5.
Starters:
Ian Seymour: 3 projected ER, 16 outs; 3 ER, 9 outs; ER error 0.
MacKenzie Gore: 2 projected ER, 16 outs; 5 ER, 9 outs; error +3.
Edges: season −0.688, recent −0.634, pitcher −0.021, composite −0.763 for TB.
Controls: recency guardrail; no volatility change.
Score path: 5–4 throughout the final stages.
Player adjustments: 24 material changes, including Wyatt Langford RBI deboost −3.72 and multiple lineup AB/RBI adjustments.
Audit: PASS; 0 defects, 7 info.
Causes: primary run_environment_under; Gore’s projection was also materially wrong. The stronger TB evidence did not survive the actual run distribution.

TOR@KC — wrong winner and score

Prediction: TOR 5, KC 4; TOR winner.
Actual: TOR 1, KC 6; KC winner.
Published probability: TOR 0.537096; home probability 0.463.
Confidence edge: −0.0742, lean TOR.
Brier / team-run error / total bias: 0.288472 / 6 / +2.
Starters:
Spencer Arrighetti: 4 projected ER, 15 outs; 6 ER, 14 outs; error +2.
Randy Dobnak: 4 projected ER, 18 outs; 1 ER, 18 outs; error −3.
Edges: season −0.604, recent −0.842, pitcher +0.060, composite −0.592 for TOR.
Controls: starter-experience control applied but did not change the winner; supported blowout expansion; score path remained 5–4.
Player adjustments: 22 material changes, including John Rave RBI +2.63 and Nathan Lukes RBI +2.48.
Audit: PASS; 0 defects, 8 info.
Causes: primary lean_side_miss; small losing-side edge rather than high conviction. Secondary causal evidence is starter/run realization: KC’s Dobnak allowed only 1 versus 4 projected, while Arrighetti allowed 6.

WAS@LAD — correct winner, wrong score

Prediction: WAS 4, LAD 5; LAD winner.
Actual: WAS 5, LAD 7; LAD winner.
Published probability: LAD 0.530; home probability 0.530.
Confidence edge: +0.0600, lean LAD.
Brier / team-run error / total bias: 0.220900 / 3 / −3.
Starters:
Andrew Alvarez: 2 projected ER, 15 outs; 3 ER, 13 outs; error +1.
Justin Wrobleski: 1 projected ER, 16 outs; 4 ER, 9 outs; error +3.
Edges: season +0.789, recent +0.414, pitcher −0.082, composite +0.537 for LAD.
Controls: recency guardrail; no volatility change.
Score path: 4–5 final.
Player adjustments: none available.
Audit: PASS; 0 defects, 7 info.
Causes: winner correct; underpredicted total, with Wrobleski’s three-ER excess the main starter miss. Correct-winner/wrong-score.

4. Miss reconstruction

Earliest wrong-direction evidence for each loss

ATL@PHI

Bad input: no structural or identity defect identified.
Bad weighting: likely. Season, recent, pitcher, and composite edges all favored ATL, but the published side was PHI.
Calibration: moderate 0.10 confidence was too strong for the evidence direction.
Starter error: both starters outperformed ER projections, but this did not explain the side reversal.
Offense/bullpen: actual ATL scored 5 versus projected 3; no bullpen-specific evidence supplied.
Run environment: total was 9 versus projected 7, a modest miss.
Earliest wrong-direction stage: the enhanced prediction changed from a 4–4 variance state to 3–4, committing to PHI despite a negative composite home edge. This is a winner-direction regression in evidence weighting, not merely irreducible variance.

CHC@MIA

Bad input: no structural defect, but no player actuals.
Bad weighting: the model gave dominant weight to CHC season/recent form and enhanced CHC’s score from 4 to 6.
Calibration: published CHC probability 0.58 was materially more confident than raw 0.58? Here governor output was 0.6773 while published side was 0.58; the side remained moderately confident despite the extreme outcome risk.
Starter error: earliest decisive game-specific failure. Holmes projected 2 ER and allowed 8.
Offense/bullpen: MIA scored 10 versus projected 4; CHC scored 3 versus projected 6.
Run environment: total 13 projected versus 10 actual, but the side-specific allocation was severely wrong.
Earliest wrong-direction stage: enhanced_prediction moved CHC from 4 to 6 and retained CHC as leader. The later run-volatility logic expanded the wrong leader; it did not create the initial side choice.

MIL@CIN

Bad input: no structural defect.
Bad weighting: season and composite evidence strongly favored MIL, but the model did not represent CIN’s high-run tail adequately.
Calibration: lean confidence was not excessive, but the score distribution was too narrow.
Starter error: both starters allowed more than projected, especially Harrison by 2 ER and Singer by 1.
Offense/bullpen: CIN was underprojected by 7 runs.
Run environment: earliest and dominant failure. Total actual was 20 versus predicted 12.
Earliest wrong-direction stage: the historical and enhanced stages already made MIL the leader. supported_blowout_margin_expansion widened the MIL-led score, so it reinforced the wrong winner rather than correcting it.

NYY@SD

Bad input: no structural defect.
Bad weighting: strong NYY season/recent evidence outweighed a close matchup in which the home side could win.
Calibration: lean 0.07 edge was not extreme.
Starter error: King allowed 0 versus 2 projected; Cole allowed 4 versus 3.
Offense/bullpen: no specific bullpen evidence.
Run environment: total was exactly 7 in both forecast and actual; not a total-run failure.
Earliest wrong-direction stage: enhanced_prediction changed the path to NYY 4–3. This was a winner-direction regression from a near-even 4–4 baseline, and later stages did not recover it.

TB@TEX

Bad input: no structural defect.
Bad weighting: season/recent/composite evidence strongly favored TB, but it did not sufficiently account for the possibility of TEX offense breaking out.
Calibration: lean 0.05 edge was appropriately modest.
Starter error: Gore allowed 5 versus 2 projected; Seymour’s ER count was accurate but he lasted only 9 outs.
Offense/bullpen: TEX scored 8 versus projected 4.
Run environment: actual total 14 versus projected 9.
Earliest wrong-direction stage: the baseline and enhanced stages retained TB 5–4. The miss was primarily a tail/run-environment failure, not a late publication artifact.

TOR@KC

Bad input: no structural defect.
Bad weighting: season/recent evidence favored TOR, but the model treated the recent matchup context and starter matchup insufficiently.
Calibration: lean 0.07 edge, not high conviction.
Starter error: Dobnak allowed 1 versus 4 projected; Arrighetti allowed 6 versus 4.
Offense/bullpen: TOR scored 1 versus projected 5; KC scored 6 versus 4.
Run environment: actual total 7 versus projected 9, so this was not principally a high-total miss.
Earliest wrong-direction stage: the variance_applied and enhanced stages retained TOR as leader despite a negative composite home edge. Starter-quality controls reduced confidence but left the winner unchanged. This is a lean-side weighting/starter miss.

Pipeline-stage quality across all stages

Stage	Winner accuracy	Team-run MAE	Total-run MAE	Margin MAE	Interpretation
historical_ml	8/15 = 53.33%	2.6762	4.6454	2.4455	Baseline
variance_applied	8/15 = 53.33%	2.6667	4.8000	2.2667	Margin improved; total error worsened
enhanced_prediction	9/15 = 60.00%	2.7000	4.2000	2.7333	Winner accuracy improved, but margin error regressed versus baseline
embedded_savant_run_environment	9/15	2.6667	4.1333	2.6667	Small score recovery
fundamental_edge_target_review	9/15	2.6667	4.1333	2.6667	No change
capped_target	9/15	2.6667	4.1333	2.6667	No change
simulation_result	9/15	2.6667	4.1333	2.6667	No change
published_final	9/15	2.6667	4.1333	2.6667	No change

Important transitions:

historical_ml → variance_applied: five directions changed to ties; no winner gain; margin MAE improved by 0.1788.
variance_applied → enhanced_prediction: five ties were resolved; WAS@LAD improved, while four other winner directions moved toward the eventual wrong side or otherwise did not improve. Winner accuracy increased by one game, but team-run MAE worsened by 0.0333 and margin MAE worsened by 0.4666 versus the prior stage.
enhanced_prediction → Savant environment: no winner changes; team-run and total-run errors improved modestly.
Savant through publication: no winner changes and no score changes.

A later recovery does not erase an intermediate winner-direction regression. In particular, enhanced prediction improved the aggregate hit count but still introduced or preserved wrong-direction states in ATL@PHI, CHC@MIA, NYY@SD, TOR@KC, and WAS@LAD transitions as documented by the transition receipt.

5. Cross-game counterfactuals

Lever: widen run distributions

Lever: score.variance_multiplier / score.variance_floor

Helps: ATL@PHI and CHC@MIA according to the registered tension. Wider tails would reduce overconfidence and better represent the observed outcome range.
Hurts: none listed in the lever-tension receipt.
Net evidence: 2 helps, 0 hurts on this slate. However, this is not enough to justify a large adjustment because score R² is near zero and the cross-slate lineup_run_delta bias is not persistent.
Guardrail: do not interpret the two-game result as proof that a broad variance increase improves all games. It could worsen already-close cases such as ARI@HOU, ATL@PHI’s score component, DET@CLE, or NYY@SD.

Lever: pitcher form weights / prior games

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

Helps: BOS@BAL, CHC@MIA, LAA@PIT, MIN@CHW, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC, WAS@LAD.
Hurts: ARI@HOU and DET@CLE, where starter projections were off by only 1 ER.
Net evidence: 10 helps versus 2 hurts in the supplied tension.
Cross-slate decision: no immediate parameter move. The cross-slate starter ERA error pattern is not persistent and is explicitly marked hold. A bounded diagnostic or shadow test is more appropriate than changing the weight live.

Lever: recent/season run-mean weighting

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

Helps: CHC@MIA, LAA@PIT, MIL@CIN, MIN@CHW, OAK@SEA, SF@NYM, STL@COL, TB@TEX, TOR@KC.
Hurts: ARI@HOU, ATL@PHI, DET@CLE, NYY@SD.
Net evidence: 9 helps versus 4 hurts.
Cross-slate decision: do not globally increase this lever from one slate. It helped some large-margin outcomes but hurt close games, and cross-slate lineup-run bias is explicitly nonpersistent.

Tempting fixes that must not be made

Do not simply widen every score distribution substantially because of CHC@MIA, MIL@CIN, MIN@CHW, or OAK@SEA; that could degrade close, well-estimated games.
Do not increase pitcher-form weighting globally solely because ten games are listed as helped; ARI@HOU and DET@CLE show the downside, and cross-slate persistence is absent.
Do not reverse the recent/season weighting based on ATL@PHI, NYY@SD, or TOR@KC; that would sacrifice the nine games where the lever is credited with helping.
Do not remove the starter-experience guardrail. It shrank confidence in DET@CLE, SF@NYM, and TOR@KC without changing the winner; it is not shown to be the cause of the misses.
Do not treat the simulation’s zero flips as evidence that it should be forced to flip winners. The proper fix is diagnostic and independent-distribution construction, not manual winner intervention.
Do not tune score tails as though that alone can solve the near-zero-R² feature problem.

6. Data and contract defects

Critical, error, warning, and info findings

Across the structural audit:

Critical: 0
Errors: 0
Warnings: 0
Informational findings: 114

Every per-game audit reported PASS, with zero defects and zero warnings. Thus, there were no supplied structural contract failures that explain the misses.

The major non-structural defects are:

Missing integrity-revision audit artifact — presentation/provenance gap.
Empty player feed — data-coverage gap affecting causal player attribution.
Simulation functional inertness — likely code/process defect, established across 10 slates.
Near-zero run-model signal — model/data/feature-contract defect, established across 10 slates.
Probability-lineage wording inconsistency — artifact/presentation defect for four small deltas.

Model defects versus artifact bugs

Prediction-model/process defects

Run models explain almost no holdout variance.
Simulation distribution is likely built from target-search draws and exits after approximately two draws, making the reported win-rate signal statistically vacuous.
Confidence is overaggressive in the higher-confidence subset and high-conviction misses.

Presentation/artifact defects

Missing integrity-revision audit.
Four lineage records say “no recorded transformation explains” while marking the delta as not materially unexplained.
The high-raw cluster presentation mixes raw and calibrated thresholds without an explicit distinction.

These should not be conflated: clean structural artifacts do not imply a strong predictive model.

Probability-lineage audit

The required layers are raw model → calibrated → governor → score coherence/recency → published.

Examples:

CHC@MIA: raw 0.5979 → governor input 0.58 → calibrated/governed 0.6773 → score coherence 0.671156 → published CHC 0.58. The large reduction from calibrated to published is recorded as score coherence plus matchup-recency guardrail.
MIL@CIN: raw 0.6356 → governor input 0.53 → calibrated 0.6257 → score coherence 0.625697 → published MIL 0.53; named recency transformation.
MIN@CHW: raw 0.652 → governor input 0.53 → calibrated 0.5833 → score coherence 0.583263 → published CHW 0.53; recency guardrail recorded.
WAS@LAD: raw 0.5293 → governor input 0.53 → calibrated 0.5772 → score coherence 0.577206 → published LAD 0.53; recency guardrail recorded.

For the eight material movements identified in the causal register, a named transformation is present. No material governor bypass is demonstrated.

However, ARI@HOU, ATL@PHI, LAA@PIT, and OAK@SEA have small publication deltas and text stating that no transformation explains them. Because their unexplained_material_delta flags are false, the likely intended interpretation is that the deltas are below materiality. The lineage contract should record the exact transformation or explicitly mark them as nonmaterial rounding/serialization differences.

7. Cross-slate persistence

The raw structural and trend evidence contains both persistent signals and nonpersistent noise.

Finding streaks

Finding	Consecutive-slate count	Present now?	Interpretation
high_raw_probability_cluster	0	No	3 incidences in 10 slates, not consecutive
pipeline_winner_instability	0	No	4 incidences in 10 slates, not currently present
score_margin_compression	0	No	3 incidences in 10 slates, not currently present
simulation_inert_on_slate	10	Yes	Established persistent failure
single_slate_calibration_instability	0	No	5 incidences in 10 slates, but no current streak
weak_run_model_signal	10	Yes	Established persistent failure

The two ten-slate findings must be treated as established failures, not single-slate noise.

Escalated findings

`simulation_inert_on_slate`

Streak: 10 consecutive slates.
Target lever: simulation.allow_score, although the register says it is not parameter-actionable.
Metric harmed: team_run_mae, total_run_mae; also winner-direction intervention is effectively absent.
Direction: confirm that the simulation runs and does not silently no-op.
Correct action: code/diagnostic investigation, not a forced parameter move.

The self-assessment identifies the likely mechanism: target-realization draws are reused as the win-probability distribution, and the loop exits early after finding a play-by-play realization. A win rate based on approximately two selected draws is not a probability.

`weak_run_model_signal`

Streak: 10 consecutive slates.
Target levers listed: training.xgb.n_estimators, training.xgb.max_depth, training.xgb.learning_rate.
Metric harmed: team-run MAE and total-run MAE.
Direction: investigate features and target through an isolated training backtest.
Correct action: add or improve feature contract components such as valid no-vig market probability, park/weather environment, bullpen fatigue, and platoon-adjusted lineup quality; retrain and compare same-fingerprint holdouts. Do not blindly tune nightly parameters.

The register marks this as not parameter-actionable despite listing training parameters. That means the listed parameters are candidate experiment controls, not a justified live change.

Metric-trend directions

All requested directions:

Margin bias: improving. Series ends around −1.467 versus earlier values as low as −3.5.
Mean team-run error on losses: improving. Current 5.333 is below several earlier values, though still large.
Weighted calibration error: improving. Current 0.0937 is the best/latest value in the supplied series.
Winner accuracy: improving. Current 0.600, above the nine-slate mean of 0.5688.
Top-confidence-pick loss streak: 1. The latest top-confidence pick was a miss, but the preceding slate was a hit. This is not an extended streak.

Improving trend labels do not erase the persistent run-model and simulation failures.

8. Adjustments for today

Recommendations are ranked by persistence and scope.

1. Simulation code/diagnostic repair — strongest persistent action

Registered target: simulation.allow_score
Change: do not flip this parameter blindly. Perform the registered diagnostic/code action: verify that allow_score invokes an independent simulation distribution and that the draw loop cannot early-exit before a meaningful sample.
Bounded implementation: preserve the target-realization search for score rendering, but generate an independent 200–500 draw winner distribution, or use a closed-form Poisson/Skellam difference based on expected runs and variance. Make the draw count registered.
Expected metric: improve simulation contribution, reduce team_run_mae and total_run_mae, and restore meaningful winner-probability intervention.
Helps: games with extreme or nonlinear score paths, especially CHC@MIA, MIL@CIN, MIN@CHW, OAK@SEA, and STL@COL.
Hurts/regression risk: independent simulations may alter close-game probabilities such as ARI@HOU, DET@CLE, NYY@SD, and TOR@KC. Use a shadow gate and rollback monitor.
Reason for rank: 10-slate persistence and a concrete identified code mechanism.

2. Run-model feature/target backtest — persistent model failure

Registered target: the shared feature contract in historical_game_features.py and training in historical_game_analyzer.py; candidate XGB parameters are training.xgb.n_estimators, training.xgb.max_depth, and training.xgb.learning_rate.
Change: do not live-tune one XGB parameter. Run an isolated same-fingerprint backtest after adding validated park/weather, bullpen-rest, platoon-lineup, and source-labelled no-vig market features where available. Bump the model schema version if semantics change.
Expected metric: improve home, away, and total-run R² and reduce team-run and total-run MAE.
Helps: all games, with greatest expected value in high-total and blowout cases: CHC@MIA, MIL@CIN, MIN@CHW, OAK@SEA, STL@COL, and TB@TEX.
Hurts/regression risk: feature additions can overfit or degrade close low-total games such as ARI@HOU, DET@CLE, and NYY@SD. Require holdout and rollback gates.
Reason for rank: 10-slate persistence and near-zero holdout R².

3. Bounded confidence/variance shadow experiment — strongest single-slate driver

Registered parameters: confidence.base_cap, confidence.moderate_cap, score.variance_multiplier, and score.variance_floor.
Change: apply a small downward adjustment to confidence caps and a small upward adjustment to the variance multiplier/floor in shadow mode only. The exact production magnitude is not justified by the supplied evidence; the change should remain small and reversible.
Expected metrics: lower Brier score and calibration error in overconfident moderate/high-conviction games; potentially reduce margin and total-run underdispersion.
Helps: ATL@PHI and CHC@MIA directly according to the lever tension; also potentially MIL@CIN, TB@TEX, STL@COL, and other tail misses.
Hurts/regression risk: may widen or dilute correct close calls such as ARI@HOU, DET@CLE, NYY@SD, and TOR@KC. It could also worsen score estimates where the model already tracks runs closely.
Reason for rank: current-slate evidence supports it, but it is weaker than the two persistent findings because calibration is mixed and the cross-slate variance bias is not persistent.

Changes explicitly not recommended today

Do not change context.opponent_strength_weight; cross-slate starter bias is not persistent.
Do not change score.variance_floor aggressively based only on the two-game help count; the cross-slate lineup-run bias is nonpersistent.
Do not force the simulation to flip winners without repairing its sampling logic.
Do not promote any player-level adjustment based on actual player attribution because the actual player feed is empty.

9. Bottom line

Known:

The system went 9–6, or 60%, slightly above the 56% walk-forward baseline.
Six losses occurred, including two large wrong-side or tail failures: CHC@MIA and MIL@CIN.
Exact score quality was poor: all predicted margins were only 1–2 runs, while four actual margins were at least 4.
The run models have near-zero holdout R² and this failure has persisted for 10 slates.
Simulation winner intervention has been inert for 10 consecutive slates.
Structural artifact checks passed with zero critical, error, or warning findings.
The winner probability lineage is mostly documented, but four small deltas have inconsistent explanatory text.
Starter identity matching was clean, but actual starter performance was often far from projections.

Uncertain:

Player-level causal attribution is unavailable because the slate player feed had no rows.
No market benchmark is available.
The cross-game lever counts are observational, not controlled experiments.
The exact production magnitude for confidence-cap or variance changes is not established.
The missing integrity-revision audit prevents complete provenance closure.

Safest next action: repair and instrument the simulation distribution in shadow mode, using an independent sufficiently large draw batch and a rollback monitor. This addresses the strongest persistent process failure without forcing an unvalidated live parameter change, while preserving the ability to measure whether the simulation actually improves probability calibration, winner accuracy, and run-error metrics on the next slate.

Deep slate report — 20260906

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 (15)

ARI@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

ATL@PHIMISS

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

BOS@BALHIT

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

CHC@MIAMISS

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@CLEHIT

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

LAA@PITHIT

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@CINMISS

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@CHWHIT

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

NYY@SDMISS

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@SEAHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

High: `weak_run_model_signal`

Info: `complete_slate_audit_coverage`

Info: `probability_publication_delta_explained`

Info: `simulation_inert_on_slate`