Deep slate report 20260927

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_20260927.json	n/a
forensic deep audit	loaded	probability, provenance, starter and decision contracts	/app/output/reports/PREDICTION_SYSTEM_DEEP_AUDIT_20260927.json	n/a
continuous learning	loaded	pipeline-stage quality, calibration, probability lineage	/app/output/learning/continuous_learning_daily_20260927.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 (14)

Severity	Domain	Finding and evidence	Affected	Action & guardrail
CRITICAL	probability_lineage	Published probability differs materially without a recorded transformation 1 game(s) have an unexplained ≥2pp calibrated-to-published delta	TB@PHI	Do thisBlock any promotion until the missing probability transformation is recorded and reconciled. Don't concludeDo not assume the probability governor was bypassed until you have traced the pick-side math.
HIGH	artifact_integrity	Blocking canonical artifact error structural audit totals: {'critical': 0, 'error': 2, 'warning': 1, 'info': 111}	20260927_CHC@BOS	Do thisExclude this slate from parameter learning until the artifact-generation step that wrote the broken file is fixed. Don't concludeA malformed artifact is a file bug, not proof the prediction model was wrong; do not blame accuracy changes on it.
HIGH	score_distribution	Published margins are compressed relative to realized margins 11/14 predictions were ≤2 runs; 7/14 actual margins were ≥4; mean predicted margin 1.79 vs actual 3.36	ARI@SD, CHC@BOS, CIN@TOR, HOU@OAK, TB@PHI	Do thisCheck whether predicted margins are too narrow across several recent out-of-sample slates (measuring tail coverage and margin MAE) before widening the score spread. Don't concludeA single blowout game is not a reason to widen every score prediction.
HIGH	model_validation	Validated run models explain very little holdout variance home/away/total R² = 0.0021/0.0309/0.0024	ARI@SD, ATL@MIA, CHC@BOS, CIN@TOR, CLE@KC, COL@CHW, HOU@OAK, LAA@SEA, LAD@SF, NYM@WAS, PIT@DET, STL@MIL, TB@PHI, TEX@MIN	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 15/14 games structurally audited	ARI@SD, ATL@MIA, CHC@BOS, CIN@TOR, CLE@KC, COL@CHW, HOU@OAK, LAA@SEA, LAD@SF, NYM@WAS, PIT@DET, STL@MIL, TB@PHI, TEX@MIN	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	simulation	Simulation produced no winner-direction intervention on this slate flip candidates 0; published flips 0	ARI@SD, ATL@MIA, CHC@BOS, CIN@TOR, CLE@KC, COL@CHW, HOU@OAK, LAA@SEA, LAD@SF, NYM@WAS, PIT@DET, STL@MIL, TB@PHI, TEX@MIN	Do thisMeasure the simulation's contribution across rolling slates, separating score realization from winner flipping. Don't concludeZero winner flips is not a defect and does not mean the simulation did no score-reconciliation work.

Finding	Persistence	Target lever(s)	Direction	Guardrail
simulation_inert_on_slate	11 consecutive · 11 in window	simulation.allow_score	confirm the simulation module is running, not silently no-oping	simulation_inert_on_slate has persisted 11 consecutive slates; this is no longer single-slate noise. Treat it as an actionable pattern and raise the code/diagnostic action named; do not force a parameter move.
weak_run_model_signal	11 consecutive · 11 in window	training.xgb.n_estimators, training.xgb.max_depth, training.xgb.learning_rate	investigate features/target via isolated training backtest (manual)	weak_run_model_signal has persisted 11 consecutive slates; this is no longer single-slate noise. Treat it as an actionable pattern and raise the code/diagnostic action named; do not force a parameter move.

Finding	Consecutive slates	Incidence	Present now	Fix type
high_raw_probability_cluster	0	2/11	no	param
pipeline_winner_instability	0	6/11	no	code/diagnostic
score_margin_compression	2	2/11	yes	param
simulation_inert_on_slate	11	11/11	yes	code/diagnostic
single_slate_calibration_instability	0	6/11	no	param
weak_run_model_signal	11	11/11	yes	code/diagnostic

Metric	Direction	Series (old → new)
winner_accuracy	improving	[0.4444, 0.6667, 0.7333, 0.6, 0.3333, 0.4, 0.5, 0.6667, 0.6471, 0.5385, 0.6429]
weighted_calibration_error	improving	[0.1016, 0.1247, 0.3458, 0.3184, 0.2028, 0.1881, 0.1108, 0.1701, 0.1302, 0.1524, 0.1031]
margin_bias	improving	[-2.111, -2.0, -0.667, -1.533, -2.667, -1.867, -1.188, -1.083, -1.706, -3.0, -1.571]
team_run_error_on_losses	improving	[3.6, 5.2, 6.5, 6.5, 3.0, 6.222, 5.75, 3.5, 5.333, 6.167, 4.8]

Stage	Winner hits	Accuracy	Team-run MAE	Total-run MAE	Margin MAE
historical_ml	8/14	57.1%	1.7	1.757	2.881
variance_applied	8/14	57.1%	1.679	1.786	2.786
enhanced_prediction	9/14	64.3%	1.857	2.714	3
embedded_savant_run_environment	9/14	64.3%	1.893	2.929	2.929
fundamental_edge_target_review	9/14	64.3%	1.929	2.857	3
capped_target	9/14	64.3%	1.929	2.857	3
simulation_result	9/14	64.3%	1.929	2.857	3
published_final	9/14	64.3%	1.929	2.857	3

Transition	Δ accuracy	Δ hits	Δ team-run MAE	Δ margin MAE	Games whose winner/tie changed
historical_ml → variance_applied	0	0	-0.0215	-0.0948	ARI@SD: ARI → TIE; ATL@MIA: ATL → TIE
variance_applied → enhanced_prediction	0.0715	1	0.1785	0.2143	ARI@SD: TIE → ARI; ATL@MIA: TIE → MIA; BAL@NYY: BAL → NYY; CHC@BOS: BOS → CHC; TB@PHI: PHI → TB
enhanced_prediction → embedded_savant_run_environment	0	0	0.0358	-0.0714	TB@PHI: TB → TIE
embedded_savant_run_environment → fundamental_edge_target_review	0	0	0.0357	0.0714	TB@PHI: TIE → TB
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@SD	miss	0.5176	0.5127	0.5127	0.5127	0.5108	published probability matches the calibrated/governed value
ATL@MIA	hit	0.5071	0.505	0.505	0.505	0.505	published probability matches the calibrated/governed value
CHC@BOS	hit	0.5	0.5071	0.5071	0.5071	0.5051	published probability matches the calibrated/governed value
CIN@TOR	hit	0.5332	0.5212	0.5212	0.5212	0.5116	score coherence followed by matchup-recency guardrail
CLE@KC	miss	0.6363	0.6785	0.6785	0.6605	0.56	score coherence followed by matchup-recency guardrail
COL@CHW	hit	0.6596	0.6745	0.6745	0.6745	0.56	score coherence followed by matchup-recency guardrail
HOU@OAK	hit	0.6674	0.6476	0.6476	0.645	0.56	score coherence followed by matchup-recency guardrail
LAA@SEA	hit	0.6485	0.5973	0.5973	0.5848	0.56	score coherence followed by matchup-recency guardrail
LAD@SF	hit	0.7304	0.6405	0.6405	0.6064	0.58	score coherence followed by matchup-recency guardrail
NYM@WAS	miss	0.6402	0.6055	0.6055	0.5889	0.5306	score coherence followed by matchup-recency guardrail
PIT@DET	hit	0.5182	0.5131	0.5131	0.5131	0.505	score coherence followed by matchup-recency guardrail
STL@MIL	hit	0.7139	0.6314	0.6314	0.6314	0.58	score coherence followed by matchup-recency guardrail
TB@PHI	miss	0.5	0.6056	0.6056	0.589	0.5756	no recorded transformation explains the publication delta
TEX@MIN	miss	0.5247	0.5166	0.5166	0.5166	0.512	published probability matches the calibrated/governed value

Sides analyzed	Broke on pitching	Broke on offense	Both	Variance	Mean starter ERA error	Mean lineup runs vs proj
18	2	5	0	11	-2.33	-0.11

Game	Trained p(home)	Blended p(home)	Blend favoured	Published p(home)	Published side	Artifact says winner changed	Outcome
CHC@BOS	0.56	0.526	home	0.495	away	yes	hit
TB@PHI	0.673	0.525	home	0.424	away	yes	miss

Slate	Games	Winner acc	Mean starter ERA error	Mean lineup runs vs proj	Sides broke on pitching	Sides broke on offense
20261003	4	1.000	-0.95	-2.25		4
20261002		—	—	—
20261001	1	1.000	0.00	0.50
20260930	4	0.750	0.78	0.12		1
20260929	4	1.000	-0.76	-0.50		3
20260928		—	—	—
20260927	14	0.643	-2.16	-0.11	2	5
20260926	13	0.538	0.55	0.69	3	4
20260925	17	0.647	1.13	0.50	8	10
20260924	12	0.667	1.68	-0.67	4	11
20260923	16	0.500	-0.49	-0.91	4	11
20260922	15	0.400	-0.32	-0.97	3	11

Game	Pred margin	Actual margin	Pred − actual
ARI@SD	2	5	-3
ATL@MIA	1	2	-1
CHC@BOS	1	4	-3
CIN@TOR	1	4	-3
CLE@KC	2	1	1
COL@CHW	3	2	1
HOU@OAK	2	9	-7
LAA@SEA	2	4	-2
LAD@SF	3	4	-1
NYM@WAS	2	2	0
PIT@DET	1	2	-1
STL@MIL	3	2	1
TB@PHI	1	4	-3
TEX@MIN	1	2	-1

Team	Starter	Projected	Actual	ER err
ARI	Michael Soroka	3 ER / 16 outs	3 ER / 14 outs	0
SD	Randy Vásquez	1 ER / 15 outs	1 ER / 18 outs	0

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
SD	Randy Vásquez (Starter)	18 / 15	1.5 / 1.8	1 / 1	3 / 5	0 / 0	0 / 1	1	2 / 3	2.52	venue home: 4.13 vs 3.48 ERA (worse here) · opp run factor 0.98 (average) · 31 prior appearances this season

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Manny Machado (3B)	SD	2 / 4	1 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	0 / 1	1 / 1	-1.26	0.328 above average	bad day
Gavin Sheets (DH)	SD	2 / 3	1 / 0	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 1	0 / 1	1.07	0.316 average	no signal
Austin Hays (LF)	SD	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	2 / 1	2 / 0	1 / 0	1 / 1	0.8	0.295 below average	good day
Freddy Fermin (C)	SD	4 / 3	2 / 1	1 / 0	0 / 0	0 / 0	3 / 0	0 / 0	0 / 0	1 / 1	0.78	0.27 weak	good day
Ty France (1B)	SD	4 / 4	2 / 1	0 / 1	0 / 0	0 / 0	1 / 1	2 / 1	1 / 0	0 / 1	0.49	0.327 above average	—
Dustin Harris (RF)	SD	5 / 4	2 / 1	0 / 0	0 / 0	0 / 0	0 / 1	2 / 1	0 / 0	0 / 0	0.47	0.314 average	—
Sung-Mun Song (SS)	SD	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	1 / 1	-0.14	0.253 weak	—
Jake Cronenworth (2B)	SD	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	2 / 1	1 / 1	-0.14	0.275 weak	—
Jase Bowen (CF)	SD	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0	0.244 weak	—

Signal	Home-away edge	Meaning
Season quality	0.218	positive favors home; negative favors away
Recent form	0.37	positive favors home; negative favors away
Starting-pitcher matchup	-0.05	positive favors home; negative favors away
Composite	-0.258	positive favors home; negative favors away

Stage	Away-home runs	Confidence	What changed
historical_ml	4.37-3.56	n/a	Base ML historical model prediction
variance_applied	4-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	0.511	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

Stage	Before	After	Toward final pick	What changed
variance_applied	4.37-3.56	4-4	-0.813	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-4	5-4	1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	5-4	6-4	1	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

Team	Player	Role	Net effect	Largest change
SD	Manny Machado	batter / 3B	worse	predicted_AB -2.7
ARI	Lars Nootbaar	batter / LF	better	predicted_RBI 2.686
ARI	Gabriel Moreno	batter / DH	better	predicted_RBI 2.362
SD	Dustin Harris	batter / RF	boosted	predicted_3B 1.897
SD	Ty France	batter / 1B	better	predicted_RBI 1.35
ARI	Pavin Smith	batter / 1B	worse	predicted_AB 1.237
SD	Jake Cronenworth	batter / 2B	worse	predicted_SO 1.103
ARI	Corbin Carroll	batter / RF	worse	predicted_H 1.05

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	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	ARI model-prediction batting totals differ from final simulation box score (AB: current=33, sim=38; H: current=8, sim=11; 2B: current=1, sim=3; HR: current=1, sim=2; RBI: current=4, sim=6; R: current=5, sim=6; BB: current=3, sim=2; SO: current=6, sim=7).	ARI
info	CURRENT_PLAYER_MATCH_BOX	ARI current finalized hitter lines are model predictions and differ from the simulation box score: Lars Nootbaar: AB current=4 box=5, H current=1 box=3, 2B current=0 box=1, HR current=0 box=1; Gabriel Moreno: H current=1 box=2, 2B current=0 box=1, HR current=0 box=1, RBI current=1 box=3; Geraldo Perdomo: BB current=1 box=0, SO current=0 box=1, HBP current=0 box=1; Ketel Marte: AB current=4 box=5, RBI current=1 box=0, SO current=0 box=1; Nolan Arenado: 2B current=0 box=1, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Tim Tawa: R current=0 box=1, BB current=0 box=1; Corbin Carroll: AB current=4 box=5, H current=1 box=2, 2B current=1 box=0, R current=0 box=1; James McCann: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0	ARI
info	CURRENT_FINALIZED_MATCH_BOX	SD model-prediction batting totals differ from final simulation box score (H: current=7, sim=4; 3B: current=0, sim=3; HR: current=1, sim=0; BB: current=3, sim=4; SO: current=8, sim=9).	SD
info	CURRENT_PLAYER_MATCH_BOX	SD current finalized hitter lines are model predictions and differ from the simulation box score: Dustin Harris: H current=1 box=2, 3B current=0 box=2, RBI current=1 box=2, SO current=0 box=1; Gavin Sheets: AB current=3 box=4, BB current=1 box=0; Austin Hays: H current=1 box=0, RBI current=1 box=0, R current=0 box=1, BB current=0 box=1; Ty France: H current=1 box=2, 3B current=0 box=1, RBI current=1 box=2; Sung-Mun Song: H current=1 box=0, BB current=0 box=1; Manny Machado: AB current=4 box=2, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Freddy Fermin: AB current=3 box=4, H current=1 box=0, SO current=1 box=0; Jake Cronenworth: H current=1 box=0, R current=1 box=0, BB current=1 box=0, SO current=1 box=2	SD
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

Team	Starter	Projected	Actual	ER err
ATL	JR Ritchie	3 ER / 14 outs	4 ER / 17 outs	1
MIA	Janson Junk	3 ER / 14 outs	1 ER / 14 outs	-2

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
ATL	JR Ritchie (Starter)	17 / 14	6.35 / 5.79	4 / 3	7 / 6	1 / 0	2 / 2	2.76	2 / 4	5.43	venue away: 7.01 vs 2.89 ERA (worse here) · opp run factor 1.08 (average) · 16 prior appearances this season

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Matt Olson (1B)	ATL	2 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	1 / 1	-1.73	0.351 elite	bad day
Mauricio Dubón (LF)	ATL	4 / 4	2 / 1	0 / 1	0 / 0	1 / 0	2 / 1	1 / 1	0 / 0	0 / 0	1.09	0.304 average	no signal
Mike Yastrzemski (DH)	ATL	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	-0.8	0.309 average	no signal
Sean Murphy (C)	ATL	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0.47	0.273 weak	—
Ozzie Albies (2B)	ATL	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 1	-0.47	0.273 weak	—
Brewer Hicklen (RF)	ATL	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0 / 1	-0.33	0.306 average	—
Ha-Seong Kim (SS)	ATL	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	2 / 0	1 / 1	0.19	0.222 weak	—
Austin Riley (3B)	ATL	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 0	1 / 1	-0.14	0.299 below average	—
DaShawn Keirsey Jr. (CF)	ATL	4 / 2	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 2	0 / 0	2 / 0	0	0.222 weak	—

Signal	Home-away edge	Meaning
Season quality	-0.649	positive favors home; negative favors away
Recent form	-0.454	positive favors home; negative favors away
Starting-pitcher matchup	0.068	positive favors home; negative favors away
Composite	-0.141	positive favors home; negative favors away

Stage	Away-home runs	Confidence	What changed
historical_ml	4.35-4.25	n/a	Base ML historical model prediction
variance_applied	4-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.505	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	n/a	Capped at max 12 realistic runs
simulation_result	4-5	n/a	Final inning-by-inning simulation result

Stage	Before	After	Toward final pick	What changed
variance_applied	4.35-4.25	4-4	0.095	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)
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

Team	Player	Role	Net effect	Largest change
MIA	Jared Serna	batter / 2B	better	predicted_RBI 1.808
MIA	Deyvison De Los Santos	batter / 1B	better	predicted_H 1.667
ATL	DaShawn Keirsey Jr.	batter / CF	neutral	predicted_R -1.5
MIA	Brian Navarreto	batter / C	worse	predicted_AB 1.5
MIA	Heriberto Hernández	batter / LF	deboosted	predicted_AB -1.37
MIA	Graham Pauley	batter / 3B	worse	predicted_AB 1.346
ATL	Mike Yastrzemski	batter / DH	worse	predicted_AB 1.313
ATL	Austin Riley	batter / 3B	deboosted	predicted_SO -1.248

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	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	ATL model-prediction batting totals differ from final simulation box score (AB: current=30, sim=34; H: current=8, sim=7; 2B: current=1, sim=3; HR: current=1, sim=2; R: current=6, sim=4; BB: current=3, sim=1; SO: current=7, sim=6).	ATL
info	CURRENT_PLAYER_MATCH_BOX	ATL current finalized hitter lines are model predictions and differ from the simulation box score: Austin Riley: H current=1 box=2, 2B current=0 box=1, RBI current=1 box=0, R current=1 box=0; Mike Yastrzemski: AB current=3 box=4, H current=1 box=0, BB current=1 box=0, SO current=1 box=2; Matt Olson: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Mauricio Dubón: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Sean Murphy: AB current=3 box=4, 2B current=0 box=1, RBI current=0 box=1, R current=0 box=1; Ozzie Albies: H current=1 box=2, 2B current=0 box=1; Ha-Seong Kim: H current=1 box=0, BB current=0 box=1; Brewer Hicklen: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=1, R current=0 box=1	ATL
info	CURRENT_FINALIZED_MATCH_BOX	MIA model-prediction batting totals differ from final simulation box score (AB: current=31, sim=34; H: current=7, sim=10; 3B: current=0, sim=1; RBI: current=4, sim=5; R: current=3, sim=5; BB: current=3, sim=2; SO: current=8, sim=6).	MIA
info	CURRENT_PLAYER_MATCH_BOX	MIA current finalized hitter lines are model predictions and differ from the simulation box score: Javier Sanoja: H current=1 box=2, 3B current=0 box=1, R current=0 box=1; Graham Pauley: AB current=3 box=4, RBI current=1 box=0, SO current=1 box=2; Esteury Ruiz: H current=1 box=0, BB current=0 box=1, SO current=1 box=0; Deyvison De Los Santos: AB current=3 box=4, H current=1 box=2, R current=0 box=1; Jared Serna: AB current=3 box=4, H current=0 box=1, HR current=0 box=1, RBI current=0 box=2; Agustín Ramírez: H current=1 box=0, R current=1 box=0, BB current=1 box=0; Heriberto Hernández: AB current=4 box=3, H current=1 box=2, R current=0 box=1, SO current=1 box=0; Griffin Conine: H current=1 box=2, HR current=1 box=0, BB current=1 box=0, SO current=1 box=0	MIA
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

CHC@BOSHIT

Predicted: CHC 4-3 (conf -0.01, toss_up)Actual: CHC 6-2Brier 0.245total bias -1

Root cause: Toss-up that went our way

Evidence-backed diagnosis

published system edge 0.01 (toss_up); this is not the half-scale distance from 50%
team-run absolute error 3; total bias -1
pregame home-minus-away evidence: season -0.261, recent +0.326, pitcher -0.053, composite -0.359
confidence pipeline reached 0.507 before later publication/recency controls; published home probability was 0.495
score-volatility actions: no_change
forecast-file integrity: 2 structural defect(s) on the saved artifact; see the per-game file-integrity table (this is a file check, not why the pick missed)

Quantified contributing factors

No large single-factor errors.

Team	Starter	Projected	Actual	ER err
CHC	Shota Imanaga	1 ER / 17 outs	1 ER / 3 outs	0
BOS	Tanner Houck	4 ER / 3 outs	3 ER / 9 outs	-1

Decision internals: run factors home×0.947 / away×1; 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:

Alex Bregman (CHC DH): 4-for-4, 2 HR, 4 RBI (projected 1.0 H, 0.0 HR) [run impact 3.6]
Willson Contreras (BOS DH): 0-for-2, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.73]
Pete Crow-Armstrong (CHC CF): 1-for-1, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.26]
Miguel Amaya (CHC C): 2-for-4, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 1.25]
Adley Rutschman (BOS C): 1-for-2, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.95]

Starter miss: Tanner Houck (BOS) posted 9.00 ERA vs projected 36.00 (3 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

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

Shota Imanaga (CHC) threw 3 outs, 1 ER = 9 ERA against a projected 1.59 ERA (7.41 ERA error)

Bullpen behind him: 0 ER over 3 outs.

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

Who did the damage to him:

Adley Rutschman: 1-for-2, 1 HR, 1 RBI [+0.95 runs vs projection] season xwOBA 0.327, above average - real projection error
Caleb Durbin: 1-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.291, below average - a good day, not a model error
Nick Sogard: 1-for-3, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.311, 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 Houck (BOS) threw 9 outs, 3 ER = 9 ERA against a projected 36 ERA (-27 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The AWAY lineup he faced produced 4 runs on 11 hits and 2 HR, against a projected 5 runs / 8 hits (-1 runs vs projection).

Who did the damage to him:

Alex Bregman: 4-for-4, 2 HR, 4 RBI [+3.6 runs vs projection] season xwOBA 0.327, above average - real projection error
Miguel Amaya: 2-for-4, 0 HR, 0 RBI [+1.25 runs vs projection] season xwOBA 0.295, below average - a good day, not a model error
Michael Conforto: 2-for-5, 0 HR, 1 RBI [+0.78 runs vs projection] season xwOBA 0.332, above average - real projection 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. exp BB and exp SO are the continuous rate expectations. The integer beside each is a simulation draw reconciled to the opposing simulated batting box, so it is not the line that ranks pitchers: on the 2026-09-07 slate the integers published 39 walks and 121 strikeouts against 51 and 130 real, while these expectations landed on 50.4 and 132.6.)

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
CHC	Shota Imanaga (Starter)	3 / 17	9 / 1.59 (+7.41)	1 / 1	1 / 2	1 / 0	0 / 3	0.95	3 / 5	5.38	venue away: 3.45 vs 3.9 ERA (ok) · opp run factor 0.95 (average) · 31 prior appearances this season
CHC	Jacob Webb (Closer)	3 / 3	0 / 9 (-9.00)	0 / 1	0 / 1	0 / 1	0 / 0	0.47	1 / 0	1.07
BOS	Tanner Houck (Starter)	9 / 3	9 / 36 (-27.00)	3 / 4	4 / 8	2 / 2	0 / 4	0.37	1 / 3	4.71	opp run factor 1 (strong)

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

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Alex Bregman (DH)	CHC	4 / 4	4 / 1	1 / 1	0 / 0	2 / 0	4 / 1	3 / 1	1 / 0	0 / 1	3.6	0.327 above average	projection too low
Willson Contreras (DH)	BOS	2 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	1 / 1	-1.73	0.376 elite	bad day
Pete Crow-Armstrong (CF)	CHC	1 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	0 / 1	-1.26	0.374 elite	bad day
Miguel Amaya (C)	CHC	4 / 3	2 / 0	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 1	1.25	0.295 below average	good day
Adley Rutschman (C)	BOS	2 / 4	1 / 1	0 / 1	0 / 0	1 / 0	1 / 1	1 / 0	1 / 0	0 / 1	0.95	0.327 above average	projection too low
Ian Happ (LF)	CHC	5 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	3 / 1	-0.8	0.313 average	no signal
Michael Busch (1B)	CHC	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	-0.8	0.331 above average	bad day
Roman Anthony (LF)	BOS	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	2 / 1	-0.8	0.341 elite	bad day
Michael Conforto (RF)	CHC	5 / 3	2 / 1	1 / 0	0 / 0	0 / 0	1 / 1	0 / 0	0 / 0	0 / 1	0.78	0.332 above average	projection too low
Nico Hoerner (2B)	CHC	2 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	-0.78	0.322 above average	bad day
Pedro Ramirez (3B)	CHC	4 / 4	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0.47	0.284 below average	—
Andruw Monasterio (1B)	BOS	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.287 below average	—
Caleb Durbin (3B)	BOS	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	0.47	0.291 below average	—
Trevor Story (SS)	BOS	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.248 weak	—
Nick Sogard (2B)	BOS	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	0 / 1	0.33	0.311 average	—
Ceddanne Rafaela (CF)	BOS	2 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	-0.31	0.285 below average	—
Dansby Swanson (SS)	CHC	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 0	2 / 1	-0.14	0.301 below average	—
Wilyer Abreu (RF)	BOS	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 0	1 / 1	-0.14	0.346 elite	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.261	positive favors home; negative favors away
Recent form	0.326	positive favors home; negative favors away
Starting-pitcher matchup	-0.053	positive favors home; negative favors away
Composite	-0.359	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.99-4.52	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-3	0.505	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	3.99-4.52	4-5	-0.476	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	4-3	2	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
CHC	Shota Imanaga	pitcher / Starter	neutral	predicted_SO 3
BOS	Adley Rutschman	batter / C	worse	predicted_AB -1.73
CHC	Pete Crow-Armstrong	batter / CF	worse	predicted_AB -1.51
CHC	Michael Conforto	batter / RF	better	predicted_AB 1.5
BOS	Wilyer Abreu	batter / RF	worse	predicted_AB -1.4
BOS	Andruw Monasterio	batter / 1B	worse	predicted_H 1.28
CHC	Michael Busch	batter / 1B	worse	predicted_H -1.25
CHC	Dansby Swanson	batter / SS	better	predicted_H 1.17

Structural audit findings

Severity	Rule	Finding	Subject
error	PITCHING_REALISM	current:BOS has implausible starter workload/result line(s): Tanner Houck: 1.0 IP starter outing	BOS
error	PITCHING_REALISM	simulation:BOS has implausible starter workload/result line(s): Tanner Houck: 1.0 IP starter outing	BOS
info	CURRENT_PITCHER_DUPLICATION	BOS pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	BOS
info	CURRENT_PITCHER_DUPLICATION	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_FINALIZED_MATCH_BOX	BOS model-prediction batting totals differ from final simulation box score (AB: current=34, sim=32; H: current=8, sim=5; 2B: current=2, sim=1; RBI: current=4, sim=3; R: current=4, sim=3; BB: current=2, sim=4; SO: current=9, sim=5).	BOS
info	CURRENT_PLAYER_MATCH_BOX	BOS current finalized hitter lines are model predictions and differ from the simulation box score: Willson Contreras: AB current=4 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Roman Anthony: H current=1 box=0, R current=1 box=0, BB current=1 box=0; Adley Rutschman: AB current=4 box=3, 2B current=1 box=0, RBI current=1 box=2, R current=0 box=1; Wilyer Abreu: AB current=4 box=3, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Ceddanne Rafaela: 2B current=1 box=0, HR current=0 box=1, R current=0 box=1; Andruw Monasterio: AB current=3 box=4, H current=1 box=2, 2B current=0 box=1, R current=0 box=1; Nick Sogard: AB current=4 box=3, H current=1 box=0, BB current=0 box=1; Caleb Durbin: AB current=3 box=4, H current=0 box=1, R current=1 box=0, SO current=1 box=0	BOS
info	CURRENT_FINALIZED_MATCH_BOX	CHC model-prediction batting totals differ from final simulation box score (AB: current=33, sim=35; HR: current=1, sim=2; R: current=5, sim=4; BB: current=3, sim=4; SO: current=8, sim=6).	CHC
info	CURRENT_PLAYER_MATCH_BOX	CHC current finalized hitter lines are model predictions and differ from the simulation box score: Alex Bregman: AB current=4 box=5, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Michael Conforto: AB current=3 box=4, H current=1 box=2, 2B current=0 box=1, HR current=0 box=1; Ian Happ: H current=1 box=0, R current=1 box=0, BB current=1 box=0, SO current=1 box=2; Pedro Ramirez: 2B current=0 box=1, RBI current=0 box=1, R current=0 box=1; Michael Busch: H current=1 box=0, R current=1 box=0, SO current=1 box=0; Nico Hoerner: 2B current=1 box=0, HR current=0 box=1, RBI current=0 box=1, R current=0 box=1; Pete Crow-Armstrong: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Dansby Swanson: H current=1 box=2, BB current=0 box=1, SO current=1 box=0	CHC
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

Team	Starter	Projected	Actual	ER err
CIN	Brandon Williamson	1 ER / 13 outs	4 ER / 9 outs	3
TOR	Max Scherzer	4 ER / 15 outs	1 ER / 21 outs	-3

Signal	Home-away edge	Meaning
Season quality	0.303	positive favors home; negative favors away
Recent form	-0.408	positive favors home; negative favors away
Starting-pitcher matchup	0.115	positive favors home; negative favors away
Composite	0.499	positive favors home; negative favors away

Stage	Away-home runs	Confidence	What changed
historical_ml	4.2-4.62	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.512	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by winner_starter_is_below_average, projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	n/a	Capped at max 12 realistic runs
simulation_result	4-5	n/a	Final inning-by-inning simulation result

Stage	Before	After	Toward final pick	What changed
variance_applied	4.2-4.62	4-5	0.578	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	4-5	0	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by winner_starter_is_below_average, projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	4-5	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	4-5	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	4-5	0	Capped at max 12 realistic runs
simulation_result	4-5	4-5	0	Final inning-by-inning simulation result

Team	Player	Role	Net effect	Largest change
CIN	JJ Bleday	batter / LF	better	predicted_RBI 2.62
TOR	Charles McAdoo	batter / 1B	worse	predicted_SO 2.242
TOR	Davis Schneider	batter / LF	worse	predicted_SO 1.967
TOR	Brett Bateman	batter / RF	worse	predicted_AB -1.85
CIN	Will Banfield	batter / C	worse	predicted_AB 1.5
TOR	Kazuma Okamoto	batter / 3B	better	predicted_RBI 1.35
TOR	Ernie Clement	batter / 2B	boosted	predicted_H 1.32
CIN	Carlos Jorge	batter / CF	worse	predicted_SO 1.133

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	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	CIN model-prediction batting totals differ from final simulation box score (H: current=8, sim=6; HR: current=1, sim=0).	CIN
info	CURRENT_PLAYER_MATCH_BOX	CIN current finalized hitter lines are model predictions and differ from the simulation box score: Elly De La Cruz: AB current=4 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Juan Brito: AB current=4 box=3, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Héctor Rodríguez: H current=1 box=0, RBI current=1 box=0; Carlos Jorge: R current=0 box=1, BB current=1 box=0, SO current=1 box=2; JJ Bleday: 2B current=0 box=1, RBI current=0 box=3, BB current=1 box=0, SO current=1 box=2; Oswaldo Cabrera: AB current=3 box=4; Matt McLain: BB current=0 box=1, SO current=1 box=0; Sal Stewart: H current=1 box=2, SO current=1 box=0	CIN
info	CURRENT_FINALIZED_MATCH_BOX	TOR model-prediction batting totals differ from final simulation box score (AB: current=30, sim=34; H: current=7, sim=8; 2B: current=2, sim=3; 3B: current=0, sim=1; RBI: current=4, sim=5; R: current=4, sim=5; BB: current=4, sim=2; SO: current=7, sim=12).	TOR
info	CURRENT_PLAYER_MATCH_BOX	TOR current finalized hitter lines are model predictions and differ from the simulation box score: Brett Bateman: AB current=4 box=2, 2B current=0 box=1, BB current=0 box=1, SO current=1 box=0; Charles McAdoo: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; George Springer: AB current=4 box=5, BB current=1 box=0; Kazuma Okamoto: RBI current=1 box=2, BB current=1 box=0; Davis Schneider: AB current=3 box=4, BB current=1 box=0, SO current=1 box=3; Brandon Valenzuela: AB current=3 box=4, BB current=1 box=0, SO current=1 box=2; Myles Straw: H current=0 box=1, 3B current=0 box=1, BB current=0 box=1; Andrés Giménez: AB current=3 box=4, 2B current=0 box=1, R current=0 box=1	TOR
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

Team	Starter	Projected	Actual	ER err
CLE	Parker Messick	3 ER / 19 outs	1 ER / 9 outs	-2
KC	Daniel Lynch IV	1 ER / 16 outs	2 ER / 16 outs	1

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
CLE	Parker Messick (Starter)	9 / 19	3 / 4.26	1 / 3	6 / 4	0 / 1	1 / 1	1.21	2 / 8	6.86	venue away: 2.19 vs 2.97 ERA (ok) · opp run factor 0.97 (weak) · 31 prior appearances this season

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Jo Adell (RF)	CLE	3 / 4	2 / 1	0 / 1	0 / 0	1 / 0	1 / 1	1 / 1	1 / 0	0 / 1	1.42	0.329 above average	projection too low
Angel Martínez (LF)	CLE	4 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	1 / 1	-1.4	0.299 below average	projection too high
Daniel Schneemann (3B)	CLE	4 / 3	2 / 1	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 0	0 / 1	1.4	0.292 below average	good day
Austin Hedges (DH)	CLE	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 0	-0.8	0.28 weak	projection too high
Travis Bazzana (2B)	CLE	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	-0.8	0.299 below average	projection too high
David Fry (C)	CLE	4 / 3	2 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	0.61	0.271 weak	good day
Nathaniel Lowe (1B)	CLE	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	0 / 1	-0.31	0.335 above average	—
Angel Genao (SS)	CLE	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0	0.286 below average	—
Petey Halpin (CF)	CLE	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0	0.295 below average	—

Signal	Home-away edge	Meaning
Season quality	-1.096	positive favors home; negative favors away
Recent form	-1.374	positive favors home; negative favors away
Starting-pitcher matchup	-0.028	positive favors home; negative favors away
Composite	-1.368	positive favors home; negative favors away

Stage	Away-home runs	Confidence	What changed
historical_ml	4.62-3.59	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-4	0.56	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	6-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-4	n/a	Capped at max 12 realistic runs
simulation_result	6-4	n/a	Final inning-by-inning simulation result

Stage	Before	After	Toward final pick	What changed
variance_applied	4.62-3.59	5-4	-0.026	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	6-4	1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	6-4	6-4	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-4	6-4	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-4	6-4	-0	Capped at max 12 realistic runs
simulation_result	6-4	6-4	-0	Final inning-by-inning simulation result

Team	Player	Role	Net effect	Largest change
KC	Tyler Tolbert	batter / 2B	better	predicted_RBI 3.841
CLE	Petey Halpin	batter / CF	worse	predicted_H 1.7
CLE	Travis Bazzana	batter / 2B	better	predicted_RBI 1.63
KC	Bobby Witt Jr.	batter / SS	worse	predicted_AB -1.59
CLE	Jo Adell	batter / RF	neutral	predicted_RBI 1.47
KC	Jac Caglianone	batter / RF	worse	predicted_AB -1.43
KC	Maikel Garcia	batter / 3B	worse	predicted_H -1.23
KC	Isaac Collins	batter / LF	worse	predicted_SO 1.156

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	KC pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	KC
info	CURRENT_FINALIZED_MATCH_BOX	CLE model-prediction batting totals differ from final simulation box score (AB: current=32, sim=35; 2B: current=2, sim=1; 3B: current=0, sim=1; HR: current=1, sim=2; RBI: current=4, sim=6; R: current=4, sim=6; BB: current=3, sim=5).	CLE
info	CURRENT_PLAYER_MATCH_BOX	CLE current finalized hitter lines are model predictions and differ from the simulation box score: Angel Martínez: AB current=4 box=5, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Angel Genao: HR current=0 box=1, RBI current=0 box=1, R current=0 box=1, SO current=0 box=1; Austin Hedges: 2B current=0 box=1; Daniel Schneemann: BB current=0 box=1, SO current=1 box=0; Travis Bazzana: RBI current=1 box=2, SO current=1 box=0; Jo Adell: AB current=4 box=5, 2B current=1 box=0, 3B current=0 box=1, RBI current=1 box=2; Nathaniel Lowe: 2B current=1 box=0, HR current=0 box=1, R current=0 box=1, BB current=0 box=1; Petey Halpin: AB current=3 box=4, H current=1 box=2, R current=0 box=1, SO current=0 box=1	CLE
info	CURRENT_FINALIZED_MATCH_BOX	KC model pitcher projections differ from final simulation pitching box score (H: current=9, sim=8).	KC
info	CURRENT_FINALIZED_MATCH_BOX	KC model-prediction batting totals differ from final simulation box score (H: current=8, sim=5; 2B: current=2, sim=1; HR: current=1, sim=2; BB: current=3, sim=2; SO: current=7, sim=8).	KC
info	CURRENT_PLAYER_MATCH_BOX	KC current finalized hitter lines are model predictions and differ from the simulation box score: Jac Caglianone: AB current=4 box=3, HR current=1 box=0, RBI current=1 box=0, BB current=0 box=1; Maikel Garcia: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Bobby Witt Jr.: AB current=4 box=3, 2B current=1 box=0; Salvador Perez: H current=1 box=0, RBI current=1 box=0; Tyler Tolbert: AB current=3 box=4, H current=1 box=2, HR current=0 box=2, RBI current=0 box=4; Vinnie Pasquantino: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Kyle Isbel: H current=1 box=0; Luke Maile: SO current=1 box=0	KC
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

HOU@OAKHIT

Predicted: HOU 8-6 (conf -0.12, moderate)Actual: HOU 9-0Brier 0.194total bias +5

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 7; total bias +5
pregame home-minus-away evidence: season -1.052, recent -0.490, pitcher +0.063, composite -0.922
confidence pipeline reached 0.648 before later publication/recency controls; published home probability was 0.440
score-volatility actions: supported_blowout_margin_expansion
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 5 (14 vs 9)
OAK offense over-projected by 6
Peter Lambert (HOU) allowed 4 fewer ER than projected (0 vs 4)
Seth Johnson (OAK) allowed 7 fewer ER than projected (0 vs 7)

Team	Starter	Projected	Actual	ER err
HOU	Peter Lambert	4 ER / 16 outs	0 ER / 6 outs	-4
OAK	Seth Johnson	7 ER / 12 outs	0 ER / 5 outs	-7

Decision internals: run factors home×1.063 / away×1; 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:

Christian Walker (HOU 1B): 2-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 2.06]
Lucas Spence (HOU CF): 1-for-2, 0 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 1.28]
Yordan Alvarez (HOU DH): 1-for-4, 0 HR, 1 RBI (projected 1.0 H, 1.0 HR) [run impact -1.26]
Cam Smith (HOU RF): 2-for-4, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.8]

Starter miss: Peter Lambert (HOU) posted 0.00 ERA vs projected 6.75 (0 ER in 6 outs); he faced a strong offense (opponent run factor 1.06, bucket weak).

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

Peter Lambert (HOU) threw 6 outs, 0 ER = 0 ERA against a projected 6.75 ERA (-6.75 ERA error)

Bullpen behind him: 0 ER over 3 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: neither side broke materially - the result is closer to variance

HOME pitching

No settled pitching line for this club. The forecast named Seth Johnson (OAK) at a projected 15.75 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 8 runs on 11 hits and 1 HR, against a projected 5 runs / 9 hits (3 runs vs projection).

Who did the damage to him:

Christian Walker: 2-for-3, 1 HR, 1 RBI [+2.06 runs vs projection] season xwOBA 0.305, average - a good day, not a model error
Lucas Spence: 1-for-2, 0 HR, 2 RBI [+1.28 runs vs projection] season xwOBA 0.305, average - a good day, not a model error
Cam Smith: 2-for-4, 0 HR, 1 RBI [+0.8 runs vs projection] season xwOBA 0.316, average - a good day, not a model error

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

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
HOU	Peter Lambert (Starter)	6 / 16	0 / 6.75 (-6.75)	0 / 4	2 / 7	0 / 0	0 / 3	1.73	1 / 6	5.66	venue away: 3.65 vs 3.99 ERA (ok) · opp run factor 1.06 (weak) · 28 prior appearances this season
HOU	Logan VanWey (Setup Man)	3 / 4	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0	0 / 2	0.94

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

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Christian Walker (1B)	HOU	3 / 4	2 / 1	0 / 0	0 / 0	1 / 0	1 / 1	3 / 0	2 / 0	0 / 1	2.06	0.305 average	no signal
Lucas Spence (CF)	HOU	2 / 4	1 / 1	0 / 0	1 / 0	0 / 0	2 / 0	1 / 1	3 / 1	0 / 1	1.28	0.305 average	no signal
Yordan Alvarez (DH)	HOU	4 / 4	1 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	0 / 1	1 / 1	-1.26	0.442 elite	bad day
Cam Smith (RF)	HOU	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	1 / 1	0.8	0.316 average	no signal
Isaac Paredes (3B)	HOU	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0 / 0	-0.47	0.323 above average	—
Yainer Diaz (C)	HOU	5 / 4	2 / 1	1 / 1	0 / 0	0 / 0	1 / 1	1 / 0	0 / 0	1 / 1	0.47	0.312 average	—
Taylor Trammell (LF)	HOU	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	3 / 1	-0.47	0.283 weak	—
Jeremy Peña (SS)	HOU	5 / 4	2 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	0 / 0	2 / 1	-0.46	0.326 above average	—
Jose Altuve (2B)	HOU	5 / 4	1 / 1	0 / 1	0 / 0	0 / 0	1 / 0	0 / 1	0 / 0	1 / 1	-0.31	0.271 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-1.052	positive favors home; negative favors away
Recent form	-0.49	positive favors home; negative favors away
Starting-pitcher matchup	0.063	positive favors home; negative favors away
Composite	-0.922	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	5.65-4.14	n/a	Base ML historical model prediction
variance_applied	6-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	8-6	0.56	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	8-6	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	8-6	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	8-6	n/a	Capped at max 12 realistic runs
simulation_result	8-6	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	5.65-4.14	6-4	0.493	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-4	8-6	-0	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	8-6	8-6	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	8-6	8-6	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	8-6	8-6	-0	Capped at max 12 realistic runs
simulation_result	8-6	8-6	-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
HOU	Lucas Spence	batter / CF	better	predicted_RBI 3.375
HOU	Jeremy Peña	batter / SS	better	predicted_RBI 3.04
OAK	Jeff McNeil	batter / 1B	better	predicted_RBI 2.582
OAK	Donovan Walton	batter / 2B	worse	predicted_SO 2.451
OAK	Henry Bolte	batter / CF	better	predicted_H 1.991
OAK	Shea Langeliers	batter / DH	worse	predicted_H 1.952
OAK	Joshua Kuroda-Grauer	batter / SS	worse	predicted_AB 1.65
HOU	Jose Altuve	batter / 2B	better	predicted_AB -1.64

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	HOU pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	HOU
info	CURRENT_PITCHER_DUPLICATION	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_FINALIZED_MATCH_BOX	HOU model-prediction batting totals differ from final simulation box score (AB: current=34, sim=35; H: current=9, sim=8; 2B: current=2, sim=0; HR: current=2, sim=3; RBI: current=5, sim=8; R: current=5, sim=8; BB: current=4, sim=3; SO: current=8, sim=9).	HOU
info	CURRENT_PLAYER_MATCH_BOX	HOU current finalized hitter lines are model predictions and differ from the simulation box score: Isaac Paredes: H current=1 box=0, RBI current=1 box=0, SO current=0 box=2; Yainer Diaz: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=0 box=1; Jose Altuve: AB current=4 box=3, H current=1 box=2, 2B current=1 box=0, BB current=0 box=1; Lucas Spence: HR current=0 box=1, RBI current=0 box=4, BB current=1 box=0; Jeremy Peña: H current=1 box=3, HR current=1 box=2, RBI current=1 box=4, BB current=0 box=1; Yordan Alvarez: AB current=4 box=5, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Christian Walker: AB current=4 box=3, H current=1 box=0, RBI current=1 box=0, SO current=1 box=2; Taylor Trammell: AB current=3 box=4, R current=0 box=1, BB current=1 box=0, SO current=1 box=2	HOU
info	CURRENT_FINALIZED_MATCH_BOX	OAK model pitcher projections differ from final simulation pitching box score (ER: current=7, sim=8).	OAK
info	CURRENT_FINALIZED_MATCH_BOX	OAK model-prediction batting totals differ from final simulation box score (AB: current=32, sim=38; H: current=9, sim=11; 2B: current=2, sim=1; HR: current=2, sim=0; RBI: current=5, sim=6; R: current=4, sim=6; BB: current=3, sim=4; 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: Shea Langeliers: AB current=4 box=5, H current=1 box=3, HR current=1 box=0, SO current=1 box=0; Henry Bolte: AB current=4 box=5, H current=1 box=3, 2B current=1 box=0, BB current=1 box=0; Jeff McNeil: AB current=4 box=3, H current=1 box=2, RBI current=1 box=3, R current=0 box=1; Joshua Kuroda-Grauer: AB current=3 box=5, RBI current=0 box=1, R current=0 box=1, SO current=1 box=2; Carlos Cortes: AB current=3 box=4, R current=0 box=1, SO current=1 box=0; Zack Gelof: HR current=1 box=0, HBP current=0 box=1; Lawrence Butler: H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Donovan Walton: AB current=3 box=4, H current=1 box=0, SO current=1 box=3	OAK
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

TB@PHIMISS

Predicted: TB 4-3 (conf -0.15, moderate)Actual: PHI 3-7Brier 0.331total bias -3

Root cause: Backed the wrong side with real conviction

Evidence-backed diagnosis

published system edge 0.15 (moderate); this is not the half-scale distance from 50%
team-run absolute error 5; total bias -3
pregame home-minus-away evidence: season -0.631, recent -1.050, pitcher +0.002, composite -0.604
confidence pipeline reached 0.606 before later publication/recency controls; published home probability was 0.424
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 (7 vs 10)
PHI offense under-projected by 4
committed to TB at 0.15 confidence and it lost

Team	Starter	Projected	Actual	ER err
TB	Nick Martinez	3 ER / 18 outs	4 ER / 8 outs	1
PHI	Zack Wheeler	3 ER / 18 outs	1 ER / 18 outs	-2

Decision internals: run factors home×0.978 / 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:

Jonny DeLuca (TB RF): 2-for-4, 2 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 2.33]
Bryson Stott (PHI 2B): 3-for-3, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.89]
Junior Caminero (TB 3B): 0-for-2, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.73]
Edmundo Sosa (PHI 3B): 2-for-4, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.71]
Liam Hicks (TB DH): 2-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 1.09]

Starter miss: Nick Martinez (TB) posted 13.50 ERA vs projected 4.50 (4 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

Nick Martinez (TB) threw 8 outs, 4 ER = 13.5 ERA against a projected 4.5 ERA (9 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 7 runs on 12 hits and 2 HR, against a projected 4 runs / 8 hits (3 runs vs projection).

Who did the damage to him:

Bryson Stott: 3-for-3, 0 HR, 1 RBI [+1.89 runs vs projection] season xwOBA 0.33, above average - real projection error
Edmundo Sosa: 2-for-4, 1 HR, 1 RBI [+1.71 runs vs projection] season xwOBA 0.328, above average - real projection error
J.T. Realmuto: 1-for-4, 0 HR, 1 RBI [+0.78 runs vs projection] season xwOBA 0.317, average - a good day, not a model error

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

HOME pitching

Zack Wheeler (PHI) threw 18 outs, 1 ER = 1.5 ERA against a projected 4.5 ERA (-3 ERA error)

Bullpen behind him: 0 ER over 3 outs.

The AWAY lineup he faced produced 3 runs on 8 hits and 2 HR, against a projected 4 runs / 7 hits (-1 runs vs projection).

Who did the damage to him:

Jonny DeLuca: 2-for-4, 2 HR, 2 RBI [+2.33 runs vs projection] season xwOBA 0.261, weak - a good day, not a model error
Liam Hicks: 2-for-4, 0 HR, 0 RBI [+1.09 runs vs projection] season xwOBA 0.319, average - a good day, not a model error
Jorge Mateo: 1-for-3, 0 HR, 0 RBI [+0.78 runs vs projection] season xwOBA 0.272, weak - a good day, not a model error

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

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
TB	Nick Martinez (Starter)	8 / 18	13.5 / 4.5 (+9.00)	4 / 3	7 / 8	0 / 0	1 / 0	0.67	1 / 4	3.6	venue away: 2.05 vs 3.9 ERA (ok) · opp run factor 0.98 (average) · 30 prior appearances this season
PHI	Zack Wheeler (Starter)	18 / 18	1.5 / 4.5 (-3.00)	1 / 3	5 / 3	1 / 0	0 / 2	1.81	5 / 3	7.05	venue home: 3.01 vs 3.1 ERA (ok) · opp run factor 0.98 (strong) · 27 prior appearances this season
PHI	Jhoan Duran (Closer)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0.18	2 / 0	1.37

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

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Jonny DeLuca (RF)	TB	4 / 4	2 / 1	0 / 0	0 / 0	2 / 0	2 / 1	2 / 0	0 / 0	0 / 1	2.33	0.261 weak	good day
Bryson Stott (2B)	PHI	3 / 4	3 / 1	2 / 0	0 / 0	0 / 0	1 / 0	1 / 1	1 / 0	0 / 1	1.89	0.33 above average	projection too low
Junior Caminero (3B)	TB	2 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	1 / 1	-1.73	0.364 elite	bad day
Edmundo Sosa (3B)	PHI	4 / 3	2 / 1	1 / 0	0 / 0	1 / 0	1 / 1	2 / 0	0 / 0	1 / 1	1.71	0.328 above average	projection too low
Liam Hicks (DH)	TB	4 / 4	2 / 1	2 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	0 / 1	1.09	0.319 average	no signal
Richie Palacios (2B)	TB	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	-0.8	0.306 average	no signal
Jorge Mateo (SS)	TB	3 / 3	1 / 0	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 1	0.78	0.272 weak	good day
J.T. Realmuto (C)	PHI	4 / 3	1 / 0	1 / 0	0 / 0	0 / 0	1 / 0	1 / 0	0 / 0	1 / 1	0.78	0.317 average	no signal
Cedric Mullins (CF)	TB	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0.47	0.264 weak	—
Nick Fortes (C)	TB	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.273 weak	—
Trea Turner (SS)	PHI	5 / 4	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	1 / 1	0.47	0.293 below average	—
Brandon Marsh (LF)	PHI	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.301 below average	—
Jonathan Aranda (1B)	TB	2 / 4	1 / 1	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	-0.33	0.343 elite	—
Bryce Harper (RF)	PHI	5 / 4	1 / 1	0 / 1	0 / 0	1 / 0	1 / 1	1 / 1	0 / 1	1 / 1	0.29	0.379 elite	—
Alec Bohm (1B)	PHI	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	2 / 0	0 / 0	0.19	0.327 above average	—
Kyle Schwarber (DH)	PHI	4 / 4	2 / 1	1 / 0	0 / 0	0 / 1	2 / 1	1 / 1	1 / 1	1 / 1	-0.15	0.364 elite	—
Victor Mesa Jr. (LF)	TB	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0 / 0	2 / 1	0	0.329 above average	—
Justin Crawford (CF)	PHI	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0	0.284 below average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.631	positive favors home; negative favors away
Recent form	-1.05	positive favors home; negative favors away
Starting-pitcher matchup	0.002	positive favors home; negative favors away
Composite	-0.604	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.03-5.39	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-3	0.576	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run)
embedded_savant_run_environment	3-3	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	4.03-5.39	4-5	0.355	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	4-3	2	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run)
embedded_savant_run_environment	4-3	3-3	-1	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-3	4-3	1	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
PHI	Alec Bohm	batter / 1B	better	predicted_H 2.066
PHI	Bryce Harper	batter / RF	worse	predicted_H 2.038
PHI	Zack Wheeler	pitcher / Starter	better	predicted_SO 2
TB	Jonathan Aranda	batter / 1B	worse	predicted_AB -1.57
PHI	Kyle Schwarber	batter / DH	worse	predicted_SO 1.569
TB	Liam Hicks	batter / DH	worse	predicted_RBI 1.418
PHI	Edmundo Sosa	batter / 3B	worse	predicted_AB 1.198
TB	Richie Palacios	batter / 2B	worse	predicted_AB 1.085

Structural audit findings

Severity	Rule	Finding	Subject
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_PITCHER_DUPLICATION	TB pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	TB
info	CURRENT_FINALIZED_MATCH_BOX	PHI 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; 3B: current=0, sim=1; HR: current=1, sim=0; RBI: current=4, sim=3; R: current=4, sim=3; BB: current=2, 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=3, 3B current=0 box=1, RBI current=1 box=0, BB current=1 box=0; Alec Bohm: H current=1 box=3, 2B current=0 box=1, RBI current=1 box=2, R current=0 box=1; Bryson Stott: H current=1 box=0, R current=1 box=0, SO current=1 box=0; Edmundo Sosa: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0, SO current=1 box=0; Trea Turner: AB current=4 box=5, 2B current=0 box=1, RBI current=0 box=1, SO current=1 box=0; Kyle Schwarber: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Brandon Marsh: AB current=3 box=4, SO current=1 box=0; Justin Crawford: H current=1 box=0, BB current=0 box=1, SO current=0 box=1	PHI
info	CURRENT_FINALIZED_MATCH_BOX	TB model-prediction batting totals differ from final simulation box score (AB: current=32, sim=31; H: current=7, sim=4; 2B: current=1, sim=0; 3B: current=0, sim=1; BB: current=3, sim=2; SO: current=9, sim=3).	TB
info	CURRENT_PLAYER_MATCH_BOX	TB current finalized hitter lines are model predictions and differ from the simulation box score: Victor Mesa Jr.: 3B current=0 box=1, RBI current=0 box=1, SO current=1 box=0; Liam Hicks: AB current=4 box=3, HR current=0 box=1, RBI current=1 box=2, BB current=0 box=1; Jonathan Aranda: AB current=4 box=3, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Junior Caminero: HR current=1 box=0, BB current=1 box=0, SO current=1 box=0; Jonny DeLuca: H current=1 box=0, RBI current=1 box=0, SO current=1 box=0; Richie Palacios: AB current=3 box=4, H current=1 box=0, BB current=1 box=0; Cedric Mullins: SO current=1 box=0; Nick Fortes: H current=1 box=0, SO current=1 box=0	TB
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; 14/14 games; 6/6 causes; 5/6 sources

1. Evidence receipt and gaps

Sources received

All seven source_manifest entries were checked:

Source	Status	Purpose
`prediction_review_20260927.json`	Loaded	Official outcomes and scoring errors
`PREDICTION_SYSTEM_DEEP_AUDIT_20260927.json`	Loaded	Probability, provenance, starter, and decision contracts
`continuous_learning_daily_20260927.json`	Loaded	Stage quality, calibration, probability lineage, score distribution, and starter comparisons
`detailed_20260927.html`	Loaded	Full game and simulation rendering
`20260927_ultra_detailed_player_audit.html`	Loaded	Player-level prediction trace
`audit_report_20260927.html`	Loaded	Invariant findings
`audit_report_20260927_integrity_revision.html`	Missing	Post-Luna remediation proof for immutable revision artifacts

The missing integrity-revision audit is material to remediation verification, although it is not evidence that the prediction model itself failed.

Games reviewed

The games array contains 14 games, all reviewed:

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

The game-level results reconcile to 9 winner hits and 5 misses. However, the evidence contains multiple coverage contradictions:

games_reviewed: 14, hits: 9, and continuous_learning_evidence.summary reports 14 settled games.
coverage_matrix.expected_games: 14.
But canonical_artifacts_loaded: 15, continuous_learning_games: 15, forensic_deep_audit_games: 15, and structural_audits_completed: 15.
complete_slate_audit_coverage states "15/14 games structurally audited".
coverage_matrix.all_games_structurally_audited is true despite the 15-versus-14 mismatch.
The pipeline-stage tables consistently operate on 14 games, while some slate-level summaries use 15.
continuous_learning_evidence.summary.accuracy is 0.6429, which is 9/14, but also reports games: 15, misses: 5, and settled: 14.

This is an evidence-contract defect. It does not change the 14 game outcomes presented, but it prevents calling the slate fully clean.

Other gaps and contradictions

Starter comparisons are reported for 14 games, but the coverage matrix reports 15 deep-audit and structural-audit games.
Player-level forensics are unavailable for CIN@TOR, CLE@KC, COL@CHW, LAA@SEA, LAD@SF, NYM@WAS, PIT@DET, STL@MIL, and TEX@MIN because no matching settled per-player rows were found.
Several starter lines contain implausible workload/result combinations, especially CHC@BOS, where Tanner Houck is represented as a 1.0-inning starter.
TB@PHI has an unexplained calibrated-to-published probability transformation and no recency guardrail recorded.
The artifact-integrity revision proof is missing.
No source-labelled moneyline was available for market cross-checking.
The report uses “high conviction” inconsistently with game conviction labels: the detailed games use moderate, lean, and toss_up; the aggregate spread analysis labels seven games as high-conviction.

2. Executive diagnosis

Winner accuracy

The slate produced 9 winner hits out of 14 settled games: 64.29%.

That exceeds:

The reported walk-forward baseline accuracy of 0.570149.
The historical model validation winner accuracy of 0.600414.
Random guessing.

This is a positive directional result, but it is not sufficient evidence of a good overall prediction. Four of the nine winner hits had substantial score errors, and several hits were toss-ups that happened to resolve favorably.

The five misses were:

ARI@SD, CLE@KC, NYM@WAS, TB@PHI, TEX@MIN.

Calibration quality

Calibration was weak despite acceptable winner accuracy:

Weighted absolute calibration error: 0.1031.
0.50–0.55 confidence bucket: average confidence 0.5114, observed hit rate 0.5714.
0.55–0.60 bucket: average confidence 0.5679, observed hit rate 0.7143.
Both buckets were labeled underconfident under the supplied definition, but the overall calibration error remains material.
High-confidence miss rate was 0.2857, or 2 of 7 games.
Brier performance varied materially: best 0.1764, worst 0.331361.

The system’s published confidence was generally reduced by governors, score coherence, or recency controls. That does not mean the underlying probability path was always correct. TB@PHI has an unexplained publication transformation and is the critical probability-lineage issue.

A confidence edge of 0.12 is moderate, not near a coin flip. CLE@KC and COL@CHW both carried a 0.12 edge; CLE@KC lost and COL@CHW won. Those outcomes should not be treated as equivalent to the toss-up games.

Score accuracy

Score accuracy was the clearest model weakness:

Mean predicted margin: 1.79.
Mean actual margin: 3.36.
Seven of 14 actual margins were at least four runs.
Eleven of 14 predicted margins were two runs or less.
Margin MAE: 3.0.
Signed margin bias: -1.0, indicating under-dispersed predicted margins.
Holdout run-model R²:
Away runs: 0.030936
Home runs: 0.002143
Total runs: 0.002415

The model therefore had some directional value but almost no validated exact-run explanatory power. Correct winners with poor scores must not be counted as strong forecasts.

Starter accuracy

Starter performance was mixed and strongly contaminated by outliers and missing or implausible lines:

The slate-level matchup analysis reports mean starter ERA error of -4.85 over six analyzed starts, but this is driven by a small number of extreme observations.
CHC@BOS contains a structural error involving Tanner Houck’s 1.0-inning starter workload.
CIN@TOR had large starter errors in both directions: Williamson +3 ER versus projection and Scherzer -3 ER.
HOU@OAK had very large negative errors for both starters, but the winner was still correct.
TB@PHI was materially affected by Zack Wheeler outperforming his projected ERA, while Nick Martinez underperformed.
The persistent cross-slate starter ERA sign has not passed the persistence test: 7 of 12 slates shared the sign, with mean 0.713, standard error 0.47, and z=1.51.

The pitcher-form lever therefore should not be changed from this slate alone. The lever tension was 4 games helped versus 5 hurt.

Artifact integrity

Artifact integrity is not clean:

Structural audit summary: 0 critical, 2 errors, 1 warning, 104 informational findings.
The two errors are both on CHC@BOS and concern the implausible Tanner Houck workload/result line.
The warning is on NYM@WAS: Sean Manaea pitch-count batters faced differ between summary and box score, 26 versus 24.
The critical causal issue is not a structural game artifact but an unexplained probability publication delta on TB@PHI.
The missing integrity-revision audit further limits post-remediation verification.

Causal issue register reconciliation

unexplained_probability_publication_delta — Critical

Affected game: TB@PHI.
Evidence: One game has an unexplained calibrated-to-published delta of at least two percentage points.
Action: Block promotion until the missing probability transformation is recorded and reconciled.
Guardrail: Do not assume the probability governor was bypassed until the pick-side math is traced.
Diagnosis: Supported. TB@PHI records calibrated probability 0.6056, governor output 0.6056, score-coherence probability 0.588988, and published pick probability 0.57564, with "no recorded transformation explains the publication delta". This is a contract failure, not proof that the governor was bypassed.

artifact_integrity_failure — High

Affected game: CHC@BOS.
Evidence: Two structural errors and one warning slate-wide.
Action: Exclude the slate from parameter learning until the artifact-generation step is fixed.
Guardrail: Do not blame accuracy changes on a malformed artifact.
Diagnosis: Supported. The CHC@BOS file defect is separate from the correct winner call.

score_margin_compression — High

Affected games: ARI@SD, CHC@BOS, CIN@TOR, HOU@OAK, TB@PHI.
Evidence: 11/14 predicted margins ≤2, 7/14 actual margins ≥4, predicted mean margin 1.79 versus actual 3.36.
Action: Check multiple recent out-of-sample slates before widening score distributions.
Guardrail: Do not widen every score prediction because of one blowout.
Diagnosis: Supported and now persistent for two consecutive slates. Cross-game evidence favors a bounded shadow increase in variance, not an unrestricted widening.

weak_run_model_signal — High

Affected: All 14 games.
Evidence: Near-zero holdout R².
Action: Investigate features and target through an isolated backtest.
Guardrail: Do not treat small positive R² as exact-score reliability.
Diagnosis: Strongly supported and escalated after 11 consecutive slates. This is a code/data/training problem, not a nightly tuning problem.

complete_slate_audit_coverage — Info

Evidence: The slate says every settled game received QA, but counts are internally inconsistent at 15 versus 14.
Action: Retain per-game audit receipts.
Diagnosis: The existence of per-game receipts is supported, but the coverage contract itself is not reconciled.

simulation_inert_on_slate — Info in the slate register, escalated across slates

Evidence: Zero flip candidates and zero published flips.
Action: Measure simulation contribution over rolling slates.
Guardrail: Zero flips alone is not proof that the simulator did no score work.
Diagnosis: As a single slate, this is only diagnostic. Across 11 consecutive slates, it is an established code/diagnostic failure requiring operator action.

3. Every game

The season, recent, pitcher, and composite edges below are home-minus-away edges. “Controls” includes governor, recency, starter-experience, or score-volatility handling. Player-level adjustments are reported only where supplied; absent data are not inferred.

Game-by-game table

Game	Predicted / actual	Winner	Published probability; edge; bucket	Brier; team-run error; total bias	Starter projected → actual	Edges: season / recent / pitcher / composite	Primary / secondary cause
ARI@SD	ARI 6–4 / SD 9–4	Miss	Home 0.489; edge -0.0216; toss-up	.260932; 7; -3	Soroka 3 ER, 5.06 ERA, 16 outs → 3 ER, 5.79, 14; Vásquez 1 ER, 1.80, 15 → 1 ER, 1.50, 18	+.218 / +.370 / -.050 / -.258	Toss-up variance / lean-side miss
ATL@MIA	MIA 5–4 / MIA 5–3	Hit	Home .505; edge +.010; toss-up	.245025; 1; +1	Ritchie 3 ER, 5.79, 14 → 4 ER, 6.35, 17; Junk 3 ER, 5.79, 14 → 1 ER, 1.93, 14	-.649 / -.454 / +.068 / -.141	No primary; toss-up went correctly
CHC@BOS	CHC 4–3 / CHC 6–2	Hit; correct winner, wrong score	Away .505; edge -.0103; toss-up	.244878; 3; -1	Imanaga 1 ER, 1.59, 17 → 1 ER, 9.00, 3; Houck 4 ER, 36.00, 3 → 3 ER, 9.00, 9	-.261 / +.326 / -.053 / -.359	No primary; artifact error and starter/workload anomaly
CIN@TOR	TOR 5–4 / TOR 5–1	Hit; correct winner, wrong score	Home .512; edge +.0232; toss-up	.238543; 3; +3	Williamson 1 ER, 2.08, 13 → 4 ER, 12.00, 9; Scherzer 4 ER, 7.20, 15 → 1 ER, 1.29, 21	+.303 / -.408 / +.115 / +.499	No primary; starter error
CLE@KC	CLE 6–4 / KC 3–2	Miss	Away .560; edge -.120; moderate	.313600; 5; +5	Messick 3 ER, 4.26, 19 → 1 ER, 3.00, 9; Lynch 1 ER, 1.69, 16 → 2 ER, 3.38, 16	-1.096 / -1.374 / -.028 / -1.368	Run environment over / wrong conviction side
COL@CHW	CHW 8–5 / CHW 4–2	Hit; correct winner, wrong score	Home .560; edge +.120; moderate	.193600; 7; +7	Freeland 5 ER, 10.38, 13 → 4 ER, 9.00, 12; Kay 3 ER, 5.79, 14 → 1 ER, 1.80, 15	+1.494 / +1.020 / +.048 / +1.722	Correct winner, severe score miss
HOU@OAK	HOU 8–6 / HOU 9–0	Hit; correct winner, wrong score	Away .560; edge -.120; moderate	.193600; 7; +5	Lambert 4 ER, 6.75, 16 → 0 ER, 0.00, 6; Johnson 7 ER, 15.75, 12 → 0 ER, 0.00, 5	-1.052 / -.490 / +.063 / -.922	Correct winner, starter/run-tail miss
LAA@SEA	SEA 5–3 / SEA 7–3	Hit	Home .560; edge +.120; moderate	.193600; 2; -2	Kikuchi 4 ER, 7.20, 15 → 3 ER, 5.40, 15; Gilbert 0 ER, 0.00, 17 → 0 ER, 0.00, 12	+.316 / +.250 / +.097 / +.809	Strong correct call
LAD@SF	LAD 6–3 / LAD 5–1	Hit; correct winner, wrong score	Away .580; edge -.160; moderate	.176400; 3; +3	Dreyer 1 ER, 1.80, 15 → 1 ER, 4.50, 6; Seymour 4 ER, 9.00, 12 → 0 ER, 0.00, 3	-1.631 / -1.146 / +.003 / -1.717	Strong correct call; score/starter variance
NYM@WAS	NYM 7–5 / WAS 6–4	Miss	Away .531; edge -.0611; lean	.281491; 4; +2	Manaea 4 ER, 6.35, 17 → 5 ER, 7.50, 18; Herz 4 ER, 9.82, 11 → 4 ER, 12.00, 9	+.155 / -.194 / +.059 / -.193	Lean-side miss
PIT@DET	PIT 4–3 / PIT 4–2	Hit; correct winner, close score	Away .505; edge -.010; toss-up	.245025; 1; +1	Jones 1 ER, 1.59, 17 → 0 ER, 0.00, 12; Ryan 1 ER, 3.00, 9 → 2 ER, 13.50, 4	+.036 / -.282 / +.006 / -.218	Toss-up went correctly
STL@MIL	MIL 5–2 / MIL 6–4	Hit; correct winner, wrong score	Home .580; edge +.160; moderate	.176400; 3; -3	Pallante 4 ER, 6.75, 16 → 6 ER, 10.80, 15; Misiorowski 2 ER, 3.00, 18 → 0 ER, 0.00, 16	+1.486 / +1.266 / +.101 / +1.896	Strong correct call; score under-dispersion
TB@PHI	TB 4–3 / PHI 7–3	Miss	Away .576; edge -.1513; moderate	.331361; 5; -3	Martinez 3 ER, 4.50, 18 → 4 ER, 13.50, 8; Wheeler 3 ER, 4.50, 18 → 1 ER, 1.50, 18	-.631 / -1.050 / +.002 / -.604	Wrong conviction side / lean-side miss
TEX@MIN	TEX 5–4 / MIN 6–4	Miss	Away .512; edge -.0239; toss-up	.262113; 3; -1	Gore 4 ER, 7.71, 14 → 1 ER, 3.86, 7; Kremer 2 ER, 3.60, 15 → 1 ER, 1.29, 21	+.020 / +.522 / -.028 / -.100	Toss-up variance

Controls, score paths, and player adjustments

ARI@SD: Governor and matchup-recency guardrail reduced the home probability to approximately 0.513 before publication; published home probability was 0.489. Score path moved from historical ARI 4.37–SD 3.56 to final ARI 6–SD 4, including a Savant/environment lift. Player adjustments included Machado (-2.7 predicted AB), Nootbaar (+2.686 RBI), Moreno (+2.362 RBI), and a 24-player material-change count. The audit passed with zero defects. The actual result was classified as variance rather than a material starter or bullpen break.
ATL@MIA: Governor and recency guardrail applied; final probability remained 0.505. Score path ended at ATL 4–MIA 5. Player-level evidence identified Olson underperformance and Dubón overperformance. Audit passed.
CHC@BOS: Governor and recency guardrail applied; published away probability was approximately 0.505. Score path was CHC 4–BOS 3; winner was right but score wrong. Bregman’s +3.6 run impact was the largest supplied player deviation. The artifact contained two errors, both tied to Houck’s implausible 1.0-inning starter line.
CIN@TOR: Starter-experience shrink factor 0.9458 was applied; winner unchanged. Score path remained CIN 4–TOR 5. No player-level actuals were joined. Audit passed.
CLE@KC: Governor, matchup-recency controls, and blowout-margin expansion applied. The score path expanded to CLE 6–KC 4, despite the actual low-scoring CLE 2–KC 3. Player adjustments had 23 material changes, including Tolbert and Halpin. Audit passed. This is a correct moderate-edge interpretation of the supplied pregame evidence but a wrong result.
COL@CHW: Governor and matchup-recency guardrail applied; score path expanded to COL 5–CHW 8. Actual score was only 2–4, making this a correct-winner/wrong-score game. No player-level actuals were joined. Audit passed.
HOU@OAK: Governor and matchup-recency guardrail applied; score path was HOU 8–OAK 6, actual 9–0. Walker was +2.06 runs versus projection. Audit passed. The actual starter outcomes were extreme in both directions.
LAA@SEA: Governor and recency guardrail applied; score path LAA 3–SEA 5, actual 3–7. Audit passed. No player-level actuals were joined.
LAD@SF: Governor and recency guardrail applied; score path LAD 6–SF 3, actual 5–1. The largest supplied starter deviation was Seymour at -4 ER. Audit passed.
NYM@WAS: Governor, recency guardrail, and starter-experience shrink factor 0.4042 applied; the inexperienced-starter control changed probability but not the winner. Score path ended NYM 7–WAS 5. There was one warning for Manaea’s pitch-count mismatch. No player-level actuals were joined.
PIT@DET: Governor, recency guardrail, and starter-experience control applied; score path PIT 4–DET 3, actual 4–2. Audit passed.
STL@MIL: Governor and recency guardrail applied; score path STL 2–MIL 5, actual 4–6. Audit passed. No player-level actuals were joined.
TB@PHI: Governor was applied, but recency guardrail was recorded as false. Score path changed TB 4–PHI 5 to TB 4–PHI 3 and then published TB 4–PHI 3; actual was 3–7. Player adjustments included Stott +1.89, Sosa +1.71, and Wheeler strikeout changes. Audit passed structurally, but probability lineage failed the critical contract.
TEX@MIN: Governor and recency guardrail applied; score path TEX 5–MIN 4, actual 4–6. No player-level actuals were joined. Audit passed.

4. Miss reconstruction

ARI@SD

Earliest wrong-direction movement: The historical model initially favored ARI, and variance application produced a tie. Enhanced prediction moved back toward ARI, and later score/environment stages increased ARI’s projected score to 6 versus 4.
Bad input: Partial starter evidence: the detailed player feed lacked a settled pitching line for Michael Soroka, although the game-level starter line supplied a projection and actual.
Bad weighting: Composite home edge was -0.258, while season and recent edges favored SD. The evidence was internally mixed; no clear weighting defect is proven.
Calibration: Appropriate toss-up treatment was not the principal problem; published home probability was only 0.489.
Starter error: Not material by ER: both starters had ER error 0.
Offense/bullpen: SD’s offense was under-projected by 5 runs; the detailed forensics classify the result as closer to variance.
Run environment: Total was under-projected by 3 and team-run absolute error was 7.
Irreducible variance: Strong contributor. This was a low-confidence direction that flipped.

CLE@KC

Earliest wrong-direction movement: Historical and variance stages favored CLE; enhanced prediction expanded CLE to 6–4. The score-volatility module then explicitly applied blowout-margin expansion.
Bad input: No direct artifact defect. Player-level actuals were unavailable.
Bad weighting: Season, recent, and composite edges all strongly favored CLE, so the direction was not an arbitrary small-edge mistake.
Calibration: A published 0.12 edge is moderate, not near a coin flip. It was a meaningful commitment and lost.
Starter error: Messick performed better than projected by 2 ER; Lynch was 1 ER worse. This modestly favored CLE and does not explain the miss.
Offense/bullpen: CLE offense was over-projected by 4.
Run environment: Primary cause. Ten projected total runs became five actual runs; total bias was +5.
Irreducible variance: Some, but the model also overcommitted to a wide high-scoring outcome.

NYM@WAS

Earliest wrong-direction movement: Historical ML favored NYM, and later stages increased NYM to 7–5. The starter-experience adjustment shrank confidence but did not change the winner.
Bad input: No player-level actuals were joined. One pitch-count warning exists for Manaea.
Bad weighting: Composite edge was only -0.193, despite season edge favoring NYM and recent edge favoring WAS. The lean-side call was not strongly supported after reconciliation.
Calibration: Published away probability 0.531 and edge 0.0611 were appropriately lean rather than moderate.
Starter error: Manaea was 1 ER worse than projected; Herz’s ER matched projection, though his ERA was worse because of workload.
Offense/bullpen: NYM offense was over-projected by 3.
Run environment: Total was over-projected by 2.
Irreducible variance: Present, but the principal model issue was a small edge on the wrong side.

TB@PHI

Earliest wrong-direction movement: Historical ML favored PHI, but the published pathway committed to TB. The probability-lineage evidence is incomplete, so the precise earliest probability-stage regression cannot be certified.
Bad input: The probability publication record is incomplete; recency guardrail is recorded false despite the general matchup context. Starter feed is missing Nick Martinez’s settled player line.
Bad weighting: Season and recent edges favored PHI, while the system backed TB. The composite home edge was -0.604, not a marginal disagreement.
Calibration: Published away probability was 0.57564, a moderate 0.1513 edge. This was not a coin flip and was materially overconfident relative to the loss.
Starter error: Martinez was 1 ER worse than projected; Wheeler was 2 ER better. Wheeler’s performance materially helped PHI.
Offense/bullpen: PHI offense was under-projected by 4. The player audit identified Stott and Sosa as positive deviations. Bullpen ER was not the primary issue in supplied data.
Run environment: Total was under-projected by 3; team-run error was 5.
Irreducible variance: Wheeler’s outlier performance contributed, but the critical publication-lineage defect must be resolved before causality is finalized.

TEX@MIN

Earliest wrong-direction movement: Historical and enhanced stages favored TEX. The recency and matchup controls did not reverse the side.
Bad input: No player-level actuals were joined.
Bad weighting: Recent edge favored MIN by +0.522, while season edge was nearly neutral and composite favored neither side strongly. The final TEX lean was not robust.
Calibration: Published away probability 0.51197, edge 0.0239, correctly treated as toss-up.
Starter error: Gore was 3 ER better than projected; Kremer was 1 ER better. The good pitching suppressed scoring and altered the close game.
Offense/bullpen: No decisive lineup failure was documented.
Run environment: Total bias was -1 and team-run error 3.
Irreducible variance: Primary contributor.

Pipeline-stage quality across all stages

A later correct winner does not erase an earlier winner-direction regression.

Stage	Winner accuracy	Team-run MAE	Total-run MAE	Margin MAE
Historical ML	8/14 = 57.14%	1.7001	1.7568	2.8805
Variance applied	8/14	1.6786	1.7857	2.7857
Enhanced prediction	9/14 = 64.29%	1.8571	2.7143	3.0000
Embedded Savant environment	9/14	1.8929	2.9286	2.9286
Fundamental edge target review	9/14	1.9286	2.8571	3.0000
Capped target	9/14	1.9286	2.8571	3.0000
Simulation result	9/14	1.9286	2.8571	3.0000
Published final	9/14	1.9286	2.8571	3.0000

Important transitions:

Historical ML → variance applied improved margin MAE by 0.0948 and team-run MAE by 0.0215; it changed ARI and ATL to ties.
Variance applied → enhanced prediction improved winner accuracy by one hit, but worsened margin MAE by 0.2143 and team-run MAE by 0.1785. It helped ATL@MIA and CHC@BOS but regressed TB@PHI. It also changed several winner directions, including CHC@BOS and TB@PHI.
Enhanced → embedded Savant changed TB from a tie to a TB lead but did not improve winner accuracy and worsened team-run MAE.
Embedded Savant → fundamental review again changed TB toward TB without improving winner accuracy.
Fundamental review, capped target, simulation result, and published final were identical in winner and score metrics.

The pipeline improved winner accuracy after enhanced prediction, but the improvement came with worse score quality and a documented regression on TB@PHI. That intermediate regression remains a model-process finding.

5. Cross-game counterfactuals

Lever tension 1: widen run distributions

Lever: score.variance_multiplier / score.variance_floor

Helps: CLE@KC, where a wider distribution should reduce overconfidence in the six-to-four projection and better represent a low-scoring reversal; TB@PHI, where the narrow score range failed to represent PHI’s possible offensive tail.
Hurts: No games were listed as hurt in the supplied lever tension.
Net evidence: Supports a bounded increase. The evidence is strengthened by two consecutive slates of margin compression, but the change should be shadowed because a single slate cannot prove the optimal magnitude.
Guardrail: Do not widen all predictions aggressively based on HOU@OAK or one blowout.

Lever tension 2: pitcher form weights / prior-games weighting

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

Helps: CIN@TOR, HOU@OAK, LAD@SF, and TEX@MIN according to starter-error attribution.
Hurts: ARI@SD, CHC@BOS, LAA@SEA, NYM@WAS, and PIT@DET, where current starter projections were comparatively accurate or where reweighting could disturb useful results.
Net evidence: 4 helped versus 5 hurt. Do not change this lever from the slate. Cross-slate persistence also does not support it.

Lever tension 3: recent/season run-mean weighting

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

Helps: ARI@SD, CLE@KC, COL@CHW, HOU@OAK, NYM@WAS, TB@PHI.
Hurts: ATL@MIA, LAA@SEA, PIT@DET.
Net evidence: 6 helped versus 3 hurt, but the benefit is based primarily on large score errors and near-zero validated run-model signal. This supports diagnostic investigation, not immediate aggressive reweighting.

Tempting one-game fixes that must not be made

Do not reverse the CLE@KC side solely because the 0.12 edge lost. The same magnitude of edge correctly selected COL@CHW, LAA@SEA, and HOU@OAK.
Do not widen every score because HOU@OAK ended 9–0. The register expressly warns against learning from one blowout.
Do not alter pitcher weights because of CIN@TOR or HOU@OAK. The cross-game count is 4 helped versus 5 hurt.
Do not disable season/recent form because it failed in TB@PHI; the same directional structure helped multiple other games.
Do not blame CHC@BOS’s correct winner or its model accuracy on the malformed artifact.
Do not infer that the simulator was bypassed solely from zero winner flips; the persistent code-level evidence requires a diagnostic audit first.
Do not lower all confidence caps based on the current slate’s moderate bucket results alone; persistent high-raw clustering is not present on this slate.

6. Data and contract defects

Audit findings by severity

Critical

TB@PHI: Published probability differs materially from the calibrated/governed value without a recorded transformation. The lineage records:
Raw model pick probability: 0.60695
Calibrated probability: 0.6056
Governor output: 0.6056
Score-coherence probability: 0.588988
Published pick probability: 0.57564
Recency guardrail: false
Transformation: no recorded explanation

This is a probability-contract failure. Promotion should remain blocked until reconciled.

Errors

CHC@BOS: Structural audit error: BOS/Tanner Houck has an implausible 1.0-inning starter outing in both current and simulation artifacts.
CHC@BOS: Same pitching-realism issue appears twice, once for current and once for simulation output.

These are artifact/pipeline representation errors, not proof of a wrong model winner.

Warning

NYM@WAS: Sean Manaea pitch-count row differs from the pitching box score: batters faced 26 in the summary versus 24 in the box score.

Informational and coverage defects

Structural audit summary reports 104 informational findings.
Nine games have no matched player-level actual feed.
Coverage reports 15 structural audits for an expected 14-game slate.
The integrity-revision audit is missing.
The simulation produced zero flip candidates and zero published flips.
No market cross-check was available.

Prediction-model defects versus artifact bugs

Prediction-model/process defects

Near-zero validated run-model signal: total R² 0.002415.
Persistent margin compression.
Persistent simulation inertness.
Inconsistent or unexplained probability publication for TB@PHI.
Under-dispersed score paths and poor hitter-level separation:
H projection correlation 0.054, spread ratio 0.323.
AB correlation -0.164.
R correlation -0.070.
HR correlation -0.142.
Several metrics are beaten by publishing the slate mean for every hitter.

Presentation/artifact defects

CHC@BOS implausible starter workload representation.
NYM@WAS pitch-count mismatch.
Missing integrity-revision audit.
15-versus-14 coverage and settled-game count mismatch.

These artifact defects must not be used to infer corresponding accuracy failures.

Complete probability-lineage audit

The pick-side lineage was checked for all 14 games:

ARI@SD: raw 0.5176 → calibrated 0.5127 → governor 0.5127 → score coherence/recency → published 0.510815; transformation recorded, no material unexplained delta.
ATL@MIA: raw 0.5071 → calibrated/governed/published 0.505; recorded.
CHC@BOS: raw 0.5000/governor input 0.50075 → calibrated/governed 0.5071 → published 0.505149; recorded.
CIN@TOR: raw 0.5332 → calibrated 0.5212 → score/recency → published 0.511591; recorded.
CLE@KC: raw 0.6363 → calibrated 0.6785 → governor output 0.6785 → published 0.560; transformation recorded as score coherence/recency.
COL@CHW: raw 0.6596 → calibrated 0.6745 → governor output 0.6745 → published 0.560; recorded.
HOU@OAK: raw 0.6674 → calibrated 0.6476 → governor output 0.6476 → published 0.560; recorded.
LAA@SEA: raw 0.6485 → calibrated 0.5973 → published 0.560; recorded.
LAD@SF: raw 0.7304 → calibrated 0.6405 → published 0.580; recorded.
NYM@WAS: raw 0.6402 → calibrated 0.6055 → published 0.530557; recorded.
PIT@DET: raw 0.5182 → calibrated 0.5131 → published 0.505; recorded.
STL@MIL: raw 0.7139 → calibrated 0.6314 → published 0.580; recorded.
TB@PHI: raw 0.60695 → calibrated 0.6056 → governor 0.6056 → score coherence 0.588988 → published 0.57564; material transformation not recorded.
TEX@MIN: raw 0.5247 → calibrated 0.5166 → published 0.51197; recorded.

The only declared material bypass/unexplained delta is TB@PHI. The other large raw-to-published changes have recorded transformations and should not be labeled bypasses.

Coverage reconciliation

The slate cannot be called clean because:

Expected games: 14.
Game records: 14.
Structural audit count: 15.
Canonical artifacts loaded: 15.
Continuous-learning games: 15.
Missing integrity-revision audit.
Settled count is 14 in one place and games count is 15 in another.

The per-game result set is complete for the 14 listed games, but the audit-count contract requires correction.

7. Cross-slate persistence

Raw structural evidence

Margin distribution

Current slate evidence shows:

Actual margins ≥4: 7 of 14.
Predicted margins ≤2: 11 of 14.
Mean predicted margin: 1.79.
Mean actual margin: 3.36.
Current margin-bias series endpoint: -1.571.
Trend direction: improving, but still materially negative.

The improvement does not remove the two-slate persistence of compression.

Pipeline-stage transitions

Across the current slate:

Variance application improved run errors slightly.
Enhanced prediction improved winner accuracy but worsened team-run and margin MAE and regressed TB@PHI.
Savant and fundamental review did not improve winner accuracy after enhanced prediction.
Capped target, simulation result, and published final were inert relative to the preceding stages.

Calibration monitoring

Weighted absolute calibration error: 0.1031.
Current metric-trend direction: improving.
The current slate is better than several recent slates, but the level remains materially nonzero.
High-raw cluster: 4 games, 4 hits, but the cluster is not persistent enough to support a cap change now.

High-raw cluster

Four games exceeded the raw cluster threshold and all hit:

COL@CHW, HOU@OAK, LAD@SF, STL@MIL.

The cluster is useful diagnostic evidence but not a persistent failure on this slate. It should not trigger a one-slate confidence-cap adjustment.

Simulation control

Observed games: 15 in the simulation-control block.
Flip candidates: 0.
Published flips: 0.
Status: diagnostic-only without independent support.
The cross-slate escalation confirms this has persisted 11 consecutive slates.

Finding streaks

Every finding_streaks entry:

high_raw_probability_cluster: consecutive-slate count 0; incidence 2/11; not present currently as a persistent finding.
pipeline_winner_instability: consecutive-slate count 0; incidence 6/11; not present currently.
score_margin_compression: consecutive-slate count 2; present currently.
simulation_inert_on_slate: consecutive-slate count 11; present currently.
single_slate_calibration_instability: consecutive-slate count 0; incidence 6/11; not present currently.
weak_run_model_signal: consecutive-slate count 11; present currently.

Escalated findings

These are established failures, not single-slate noise.

Simulation inertness

Streak: 11 consecutive slates; incidence 11.
Target lever: simulation.allow_score, although the register marks it non-parameter-actionable.
Metric harmed: team_run_mae, total_run_mae.
Direction of fix: Confirm the simulator is actually running and independently estimate the winner distribution. The supplied self-assessment identifies a shared-draw, early-exit problem: the simulation uses the same target-realization draws for score construction, may stop after roughly two selected draws, and therefore cannot produce a meaningful win rate.

Weak run-model signal

Streak: 11 consecutive slates; incidence 11.
Target levers: training.xgb.n_estimators, training.xgb.max_depth, training.xgb.learning_rate, but the register marks the finding non-parameter-actionable.
Metric harmed: team_run_mae, total_run_mae.
Direction of fix: Isolated feature/target backtest. Add or validate market no-vig probability where source-labelled, park/weather environment, bullpen rest/fatigue, and platoon-adjusted lineup quality. Bump model schema when semantics change.

Top-confidence-pick loss streak

top_confidence_pick.loss_streak: 0.
The 11-slate sequence was: hit, hit, miss, hit, miss, miss, hit, hit, hit, miss, hit.
The current slate’s top confidence pick was a hit.
Therefore there is no current consecutive top-pick loss streak.

Metric-trend directions

All supplied metric trends are explicitly improving:

Margin bias: improving.
Mean team-run error on losses: improving.
Weighted calibration error: improving.
Winner accuracy: improving.

“Improving” does not mean solved. In particular, the run model remains near-zero-signal and margin compression remains present.

8. Adjustments for today

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

1. Simulation inertness — highest evidence strength

Action type: Code/diagnostic fix, not a parameter promotion.
Registered parameter: simulation.allow_score is the registered lever, but the finding explicitly says it is not parameter-actionable.
Bounded change: Do not change the parameter blindly. Run a shadow implementation with an independent unsteered winner-distribution batch of approximately 200–500 draws, or a closed-form Poisson/Skellam difference using expected runs and variance. Remove the early-exit dependency from the probability batch.
Expected metric: Improve the diagnostic validity of winner probability; potentially improve team_run_mae and total_run_mae.
Helps: All games, especially games where the simulation currently produces no independent winner intervention.
Hurts/risk: Additional compute and possible score/winner disagreements; rollback if independent simulation worsens rolling Brier or winner accuracy.
Why first: Persistent 11-slate code-level failure and direct self-assessment evidence.

2. Weak run-model signal — persistent but operator-only

Action type: Isolated training/backtest code and feature-contract work.
Registered parameters: training.xgb.n_estimators, training.xgb.max_depth, and training.xgb.learning_rate are listed, but no bounded nightly parameter move is recommended because the finding is explicitly non-parameter-actionable.
Bounded change: Do not tune these parameters in production today. Add the proposed feature-contract candidates in an isolated experiment, retrain with the same temporal fingerprint, and compare holdout R² before and after.
Expected metric: Increase away/home/total run R² from approximately 0.031/0.002/0.002; reduce team_run_mae and total_run_mae.
Helps: Broadly the full slate, particularly score-error games ARI@SD, CLE@KC, COL@CHW, HOU@OAK, NYM@WAS, and TB@PHI.
Hurts/risk: Feature leakage, unstable schema changes, or deterioration in exact-run and winner metrics. Require schema-version increment and rollback if holdout performance does not improve.
Why second: Persistent 11-slate failure and near-zero holdout signal.

3. Margin compression — strongest current single-slate driver with persistence

Registered parameter: score.variance_multiplier.
Direction: Increase modestly.
Magnitude: Shadow increase of approximately 5% relative to the current value, with no simultaneous change to score.variance_floor.
Expected metric: Reduce margin compression, improve tail coverage, and reduce margin MAE; potentially improve score calibration. The effect on winner Brier is uncertain.
Helps: CLE@KC and TB@PHI are explicitly identified by lever tension; also protects against under-dispersed outcomes in ARI@SD, CIN@TOR, HOU@OAK, and STL@MIL.
Hurts/risk: Could worsen already close-score games such as ATL@MIA, LAA@SEA, and PIT@DET, and can increase total-run MAE if applied indiscriminately.
Gate: Keep as a shadow experiment with rollback if rolling total-run MAE increases materially or low-scoring-game calibration degrades.
Why third: Present for two consecutive slates, directly supported by current margin distribution, and has a registered actionable lever.

4. Confidence caps — do not change today

Registered levers: confidence.base_cap, confidence.moderate_cap, and confidence.empirical_max_weight.
Reason for abstention: The current high-raw cluster went 4-for-4, the top-confidence pick hit, and the persistent high-raw finding is not present on this slate. Although the self-assessment proposes lower caps, the cross-game and cross-slate evidence does not isolate a current cap failure.
Potential benefit: Lower Brier and calibration error if overconfidence persists.
Potential harm: Unnecessarily suppressing useful moderate edges such as COL@CHW, LAA@SEA, LAD@SF, and STL@MIL.
Status: Monitor, do not promote a change today.

5. Pitcher weighting — abstain

Registered lever: player.pitcher.v4 form weights / player.pitcher_prior_games.
Reason: 4 games helped and 5 hurt; cross-slate starter ERA bias is not separable from noise.
Affected risk: A change could disturb accurate starter projections in ARI@SD, CHC@BOS, LAA@SEA, NYM@WAS, and PIT@DET.
Action: No parameter move. Continue monitoring starter ER MAE and outs MAE.

9. Bottom line

Known

The system correctly selected 9 of 14 winners, outperforming the supplied baseline.
Winner accuracy was materially better than score accuracy.
Score margins were compressed: predicted margins were too narrow and run-model holdout R² was near zero.
Two failures are persistent across 11 slates: inert simulation behavior and weak run-model signal.
Margin compression is present for two consecutive slates and supports a bounded variance shadow experiment.
TB@PHI has a critical probability-lineage defect requiring reconciliation.
CHC@BOS has artifact errors, and NYM@WAS has a pitch-count warning.
The slate is not artifact-clean because of the 15-versus-14 coverage contradiction and the missing integrity-revision audit.

Uncertain

Whether the TB@PHI publication delta reflects a governor bypass, an undocumented score-coherence transformation, or a reporting omission. The supplied guardrail correctly prohibits assuming bypass before tracing the math.
Whether the observed starter ERA sign reflects a real bias or slate-to-slate noise.
Whether recent/season run-mean weighting is genuinely beneficial, because the model’s underlying run features have almost no validated signal.
The exact causal contribution of the simulation to score or winner probabilities, because its current batch is not independently informative.
Player-level causes for the nine games without matched player actuals.

Single safest next action

Block promotion and run the operator-level probability-lineage and simulation contract audit first: reconcile TB@PHI’s raw-to-published probability transformation and replace the inert shared-draw simulation probability with an independent, non-early-exiting winner-distribution shadow. This improves artifact trust and diagnostic validity without making an unsupported model-parameter change.

Team	Starter	Projected	Actual	ER err
COL	Kyle Freeland	5 ER / 13 outs	4 ER / 12 outs	-1
CHW	Anthony Kay	3 ER / 14 outs	1 ER / 15 outs	-2

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Munetaka Murakami (1B)	CHW	4 / 4	1 / 1	1 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0 / 1	1 / 1	-0.95	0.351 elite	bad day
Sam Antonacci (2B)	CHW	3 / 4	2 / 1	1 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	0.78	0.348 elite	projection too low
Colson Montgomery (3B)	CHW	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	2 / 1	-0.47	0.294 below average	—
Braden Montgomery (RF)	CHW	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	-0.33	0.325 above average	—

Signal	Home-away edge	Meaning
Season quality	1.494	positive favors home; negative favors away
Recent form	1.02	positive favors home; negative favors away
Starting-pitcher matchup	0.048	positive favors home; negative favors away
Composite	1.722	positive favors home; negative favors away

Stage	Away-home runs	Confidence	What changed
historical_ml	4.19-4.97	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-8	0.56	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	5-8	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	5-8	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	5-8	n/a	Capped at max 12 realistic runs
simulation_result	5-8	n/a	Final inning-by-inning simulation result

Stage	Before	After	Toward final pick	What changed
variance_applied	4.19-4.97	4-5	0.219	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	5-8	2	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	5-8	5-8	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	5-8	5-8	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	5-8	5-8	0	Capped at max 12 realistic runs
simulation_result	5-8	5-8	0	Final inning-by-inning simulation result

Team	Player	Role	Net effect	Largest change
CHW	Randal Grichuk	batter / LF	better	predicted_RBI 2.58
CHW	Munetaka Murakami	batter / 1B	better	predicted_RBI 2.389
COL	Jake McCarthy	batter / LF	worse	predicted_SO 2.321
COL	Connor Norby	batter / DH	worse	predicted_AB 1.825
CHW	Drew Romo	batter / C	better	predicted_RBI 1.76
COL	Ezequiel Tovar	batter / SS	deboosted	predicted_RBI 1.69
COL	Troy Johnston	batter / 1B	boosted	predicted_RBI 1.598
CHW	Tommy Pham	batter / DH	better	predicted_AB 1.5

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	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_FINALIZED_MATCH_BOX	CHW model-prediction batting totals differ from final simulation box score (AB: current=31, sim=34; H: current=8, sim=9; 2B: current=1, sim=2; HR: current=1, sim=3; RBI: current=4, sim=8; R: current=4, sim=8; BB: current=3, sim=4; SO: current=8, sim=9).	CHW
info	CURRENT_PLAYER_MATCH_BOX	CHW current finalized hitter lines are model predictions and differ from the simulation box score: Randal Grichuk: AB current=4 box=5, RBI current=1 box=3, R current=0 box=1, SO current=1 box=3; Tommy Pham: AB current=3 box=4, H current=1 box=0, R current=0 box=1; Munetaka Murakami: AB current=4 box=5, H current=1 box=3, HR current=1 box=2, RBI current=1 box=3; Sam Antonacci: AB current=4 box=3, H current=1 box=0, BB current=0 box=1, SO current=1 box=2; Braden Montgomery: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Colson Montgomery: RBI current=1 box=0, R current=0 box=1, BB current=0 box=1; Luisangel Acuña: R current=0 box=1, BB current=0 box=1; Drew Romo: AB current=3 box=4, H current=0 box=2, 2B current=0 box=1, HR current=0 box=1	CHW
info	CURRENT_FINALIZED_MATCH_BOX	COL model pitcher projections differ from final simulation pitching box score (ER: current=7, sim=8).	COL
info	CURRENT_FINALIZED_MATCH_BOX	COL model-prediction batting totals differ from final simulation box score (AB: current=33, sim=36; H: current=8, sim=9; 2B: current=2, sim=1; RBI: current=4, sim=5; R: current=4, sim=5; BB: current=3, sim=2; SO: current=8, sim=7).	COL
info	CURRENT_PLAYER_MATCH_BOX	COL current finalized hitter lines are model predictions and differ from the simulation box score: Connor Norby: AB current=4 box=5, 2B current=1 box=0, BB current=1 box=0, SO current=1 box=0; Cole Carrigg: 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0, SO current=1 box=0; Adael Amador: RBI current=1 box=0, R current=1 box=0, BB current=1 box=0, SO current=0 box=1; Troy Johnston: 2B current=0 box=1, RBI current=0 box=2, R current=0 box=1, BB current=0 box=1; Jordan Beck: AB current=3 box=4, SO current=1 box=0; Kyle Karros: AB current=4 box=3, R current=0 box=1; Jake McCarthy: AB current=4 box=5, RBI current=1 box=0, SO current=1 box=3; Ezequiel Tovar: AB current=3 box=4, H current=0 box=1, RBI current=0 box=2, R current=0 box=1	COL
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

Team	Starter	Projected	Actual	ER err
LAA	Yusei Kikuchi	4 ER / 15 outs	3 ER / 15 outs	-1
SEA	Logan Gilbert	0 ER / 17 outs	0 ER / 12 outs	0

Signal	Home-away edge	Meaning
Season quality	0.316	positive favors home; negative favors away
Recent form	0.25	positive favors home; negative favors away
Starting-pitcher matchup	0.097	positive favors home; negative favors away
Composite	0.809	positive favors home; negative favors away

Stage	Away-home runs	Confidence	What changed
historical_ml	3.62-5.22	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.56	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_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

Stage	Before	After	Toward final pick	What changed
variance_applied	3.62-5.22	4-5	-0.602	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	4-5	0	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_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

Team	Player	Role	Net effect	Largest change
SEA	Randy Arozarena	batter / LF	better	predicted_RBI 3.5
LAA	Oswald Peraza	batter / SS	better	predicted_H 2.296
LAA	Denzer Guzman	batter / 3B	worse	predicted_SO 2.026
SEA	Cole Young	batter / 2B	worse	predicted_AB -1.96
SEA	Cal Raleigh	batter / DH	deboosted	predicted_2B 1.879
LAA	Moisés Ballesteros	batter / C	worse	predicted_RBI 1.667
LAA	Mitch Farris	pitcher / Closer	better	predicted_IP 1.667
SEA	Josh Naylor	batter / 1B	worse	predicted_SO 1.477

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	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	LAA model-prediction batting totals differ from final simulation box score (AB: current=32, sim=36; H: current=7, sim=9; 2B: current=1, sim=2; 3B: current=0, sim=1; RBI: current=4, sim=3; BB: current=2, sim=0; 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: Zach Neto: AB current=4 box=5, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Wade Meckler: R current=1 box=0, BB current=1 box=0, SO current=1 box=2; Vaughn Grissom: RBI current=1 box=0, SO current=1 box=0; Moisés Ballesteros: HR current=0 box=1, RBI current=0 box=2, R current=0 box=1; Josh Lowe: H current=1 box=0, SO current=1 box=2, HBP current=0 box=1; Oswald Peraza: AB current=3 box=4, H current=0 box=3, 2B current=0 box=1, R current=0 box=1; Bryce Teodosio: AB current=3 box=4; Christian Moore: AB current=3 box=4, H current=1 box=2, 3B current=0 box=1, R current=0 box=1	LAA
info	CURRENT_FINALIZED_MATCH_BOX	SEA model-prediction batting totals differ from final simulation box score (AB: current=34, sim=32; 2B: current=1, sim=2; 3B: current=0, sim=1; HR: current=1, sim=2; RBI: current=4, sim=5; R: current=4, sim=5; BB: current=4, sim=1).	SEA
info	CURRENT_PLAYER_MATCH_BOX	SEA current finalized hitter lines are model predictions and differ from the simulation box score: Cal Raleigh: H current=1 box=2, 2B current=0 box=2, RBI current=1 box=0, R current=0 box=1; Dominic Canzone: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Randy Arozarena: H current=1 box=3, 3B current=0 box=1, HR current=1 box=2, RBI current=1 box=4; Julio Rodríguez: H current=1 box=0, RBI current=1 box=0, SO current=1 box=2; J.P. Crawford: H current=1 box=2, BB current=1 box=0, SO current=1 box=2; Josh Naylor: RBI current=0 box=1, R current=0 box=1, SO current=1 box=2; Jhonny Pereda: H current=1 box=0, SO current=1 box=0; Cole Young: AB current=4 box=2, H current=1 box=0, BB current=0 box=1	SEA
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

Team	Starter	Projected	Actual	ER err
LAD	Jack Dreyer	1 ER / 15 outs	1 ER / 6 outs	0
SF	Carson Seymour	4 ER / 12 outs	0 ER / 3 outs	-4

Pitcher	Team	Outs	ERA	ER	H	HR	BB	exp BB	SO	exp SO	Context (as-of)
LAD	Jack Dreyer (Starter)	6 / 15	4.5 / 1.8	1 / 1	2 / 3	0 / 1	0 / 2	2.46	2 / 7	8.84	venue away: 0 vs 9 ERA (ok) · opp run factor 1 (weak) · 66 prior appearances this season
LAD	Alex Vesia (Middle Reliever (High Leverage))	3 / 3	0 / 18 (-18.00)	0 / 2	0 / 2	0 / 0	0 / 1	0.52	1 / 0	1.26

Batter	Team	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
Shohei Ohtani (DH)	LAD	3 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	0 / 1	-1.26	0.397 elite	bad day
Andy Pages (CF)	LAD	3 / 3	1 / 1	0 / 0	0 / 0	1 / 0	4 / 1	1 / 1	1 / 0	1 / 1	1.26	0.343 elite	projection too low
Max Muncy (3B)	LAD	3 / 4	2 / 1	1 / 0	0 / 0	0 / 0	0 / 1	1 / 0	2 / 1	0 / 1	1.11	0.349 elite	projection too low
Freddie Freeman (1B)	LAD	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	-1.11	0.363 elite	bad day
Tommy Edman (2B)	LAD	3 / 3	2 / 1	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	1.11	0.303 average	no signal
Teoscar Hernández (LF)	LAD	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	-0.47	0.332 above average	—
Mookie Betts (SS)	LAD	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	2 / 0	-0.47	0.339 above average	—
Kyle Tucker (RF)	LAD	5 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0 / 0	-0.33	0.331 above average	—
Hunter Feduccia (C)	LAD	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	0.33	0.312 average	—

Signal	Home-away edge	Meaning
Season quality	-1.631	positive favors home; negative favors away
Recent form	-1.146	positive favors home; negative favors away
Starting-pitcher matchup	0.003	positive favors home; negative favors away
Composite	-1.717	positive favors home; negative favors away

Stage	Away-home runs	Confidence	What changed
historical_ml	6.2-2.25	n/a	Base ML historical model prediction
variance_applied	6-2	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	0.58	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	6-3	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-3	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-3	n/a	Capped at max 12 realistic runs
simulation_result	6-3	n/a	Final inning-by-inning simulation result

Stage	Before	After	Toward final pick	What changed
variance_applied	6.2-2.25	6-2	0.058	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-2	5-4	-3	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	5-4	6-3	2	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-3	6-3	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-3	6-3	-0	Capped at max 12 realistic runs
simulation_result	6-3	6-3	-0	Final inning-by-inning simulation result

Team	Player	Role	Net effect	Largest change
SF	Bo Davidson	batter / RF	worse	predicted_SO 2
SF	Dylan Smith	pitcher / Closer	better	predicted_IP 2
SF	Christian Koss	batter / SS	better	predicted_RBI 1.83
LAD	Teoscar Hernández	batter / LF	better	predicted_RBI 1.61
LAD	Mookie Betts	batter / SS	better	predicted_SO 1.55
SF	Scott Bandura	batter / LF	worse	predicted_AB 1.5
LAD	Hunter Feduccia	batter / C	worse	predicted_AB 1.464
LAD	Max Muncy	batter / 3B	better	predicted_RBI 1.44

Deep slate report — 20260927

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 (14)

ARI@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

ATL@MIAHIT

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@BOSHIT

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

CIN@TORHIT

Player-level breakdown unavailable

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

CLE@KCMISS

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

COL@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

HOU@OAKHIT

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@SEAHIT

Player-level breakdown unavailable

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

LAD@SFHIT

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

NYM@WASMISS

Player-level breakdown unavailable

Fundamental evidence entering the pick

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	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	LAD model pitcher projections differ from final simulation pitching box score (HR: current=1, sim=2).	LAD
info	CURRENT_FINALIZED_MATCH_BOX	LAD model-prediction batting totals differ from final simulation box score (AB: current=33, sim=35; 2B: current=1, sim=3; HR: current=1, sim=3; RBI: current=4, sim=6; R: current=5, sim=6; BB: current=4, sim=1; SO: current=7, sim=9).	LAD
info	CURRENT_PLAYER_MATCH_BOX	LAD current finalized hitter lines are model predictions and differ from the simulation box score: Shohei Ohtani: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, BB current=1 box=0; Teoscar Hernández: H current=1 box=2, HR current=0 box=1, RBI current=0 box=2; Kyle Tucker: 2B current=0 box=1, BB current=1 box=0, SO current=0 box=1; Mookie Betts: H current=1 box=2, 2B current=0 box=1, HR current=0 box=1, SO current=0 box=2; Max Muncy: AB current=4 box=3, H current=1 box=2, HR current=0 box=1, RBI current=1 box=2; Freddie Freeman: H current=1 box=0, 2B current=1 box=0, BB current=1 box=0, SO current=1 box=2; Hunter Feduccia: AB current=3 box=4, SO current=1 box=2; Tommy Edman: AB current=3 box=4, H current=1 box=0, SO current=1 box=0	LAD
info	CURRENT_FINALIZED_MATCH_BOX	SF model-prediction batting totals differ from final simulation box score (AB: current=30, sim=32; H: current=7, sim=5; 2B: current=1, sim=2; HR: current=1, sim=2; RBI: current=4, sim=3; R: current=4, sim=3; BB: current=2, sim=3; SO: current=10, sim=8).	SF
info	CURRENT_PLAYER_MATCH_BOX	SF current finalized hitter lines are model predictions and differ from the simulation box score: Shay Whitcomb: H current=1 box=0, RBI current=1 box=0, BB current=0 box=1, SO current=1 box=0; Jonah Cox: AB current=4 box=3, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Victor Bericoto: H current=1 box=2, 2B current=0 box=1, HR current=1 box=0, RBI current=1 box=0; Bo Davidson: HR current=0 box=1, SO current=1 box=3; Marcelo Mayer: AB current=3 box=4, H current=1 box=0, SO current=1 box=2; Drew Cavanaugh: H current=1 box=0, SO current=1 box=0; Scott Bandura: AB current=3 box=4, H current=0 box=1, 2B current=0 box=1; Christian Koss: AB current=3 box=4, H current=0 box=1, HR current=0 box=1, RBI current=0 box=2	SF
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

Team	Starter	Projected	Actual	ER err
NYM	Sean Manaea	4 ER / 17 outs	5 ER / 18 outs	1
WAS	DJ Herz	4 ER / 11 outs	4 ER / 9 outs	0

Signal	Home-away edge	Meaning
Season quality	0.155	positive favors home; negative favors away
Recent form	-0.194	positive favors home; negative favors away
Starting-pitcher matchup	0.059	positive favors home; negative favors away
Composite	-0.193	positive favors home; negative favors away

Stage	Away-home runs	Confidence	What changed
historical_ml	6.31-4	n/a	Base ML historical model prediction
variance_applied	6-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-5	0.531	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

Stage	Before	After	Toward final pick	What changed
variance_applied	6.31-4	6-4	-0.317	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-4	6-5	-1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
embedded_savant_run_environment	6-5	7-5	1	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

Team	Player	Role	Net effect	Largest change
WAS	Brady House	batter / 3B	better	predicted_RBI 3.6
NYM	Francisco Lindor	batter / SS	boosted	predicted_RBI 2.59
NYM	Luis Torrens	batter / C	worse	predicted_SO 2.325
NYM	Carson Benge	batter / RF	better	predicted_H 1.934
NYM	Bo Bichette	batter / 3B	worse	predicted_AB -1.91
NYM	Marcus Semien	batter / 2B	worse	predicted_RBI 1.7
NYM	Juan Soto	batter / LF	worse	predicted_H 1.56
WAS	Andrés Chaparro	batter / 1B	worse	predicted_AB 1.413

Severity	Rule	Finding	Subject
warning	PITCH_COUNT_PITCHER_MATCH	NYM pitch_count row for Sean Manaea differs from pitching box score: batters_faced/BF: summary=26, box=24.	NYM Sean Manaea
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	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	NYM model pitcher projections differ from final simulation pitching box score (H: current=8, sim=7).	NYM
info	CURRENT_FINALIZED_MATCH_BOX	NYM model-prediction batting totals differ from final simulation box score (AB: current=32, sim=40; H: current=9, sim=13; 2B: current=1, sim=2; RBI: current=5, sim=7; R: current=4, sim=7; BB: current=3, sim=2; SO: current=8, sim=11; HBP: current=0, sim=1).	NYM
info	CURRENT_PLAYER_MATCH_BOX	NYM current finalized hitter lines are model predictions and differ from the simulation box score: Carson Benge: AB current=4 box=5, H current=1 box=3, 2B current=0 box=1; Francisco Lindor: AB current=4 box=5, H current=1 box=2, 2B current=0 box=1, HR current=0 box=1; Marcus Semien: AB current=4 box=5, H current=1 box=2, RBI current=0 box=2, R current=0 box=1; Brett Baty: AB current=3 box=4, H current=1 box=2, RBI current=0 box=1, R current=0 box=1; A.J. Ewing: AB current=3 box=4, H current=1 box=0, SO current=1 box=2; Ronny Mauricio: AB current=3 box=4, H current=1 box=0, SO current=1 box=0; Juan Soto: AB current=4 box=5, H current=1 box=3, HR current=1 box=0, RBI current=1 box=0; Bo Bichette: AB current=4 box=3, 2B current=1 box=0, RBI current=1 box=0, BB current=0 box=1	NYM
info	CURRENT_FINALIZED_MATCH_BOX	WAS model pitcher projections differ from final simulation pitching box score (ER: current=6, sim=7).	WAS
info	CURRENT_FINALIZED_MATCH_BOX	WAS model-prediction batting totals differ from final simulation box score (AB: current=33, sim=31; H: current=8, sim=7; 2B: current=2, sim=1; R: current=6, sim=5; BB: current=3, sim=2; SO: current=8, sim=9; HBP: 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, 2B 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; Brady House: AB current=4 box=3, H current=1 box=2, HR current=0 box=1, RBI current=1 box=4; Dylan Crews: H current=1 box=0, R current=1 box=0, SO current=1 box=2; Andrés Chaparro: H current=1 box=2, HR current=1 box=0, BB current=1 box=0; Nasim Nuñez: AB current=3 box=2, BB current=0 box=1; Jacob Young: AB current=3 box=4, H current=1 box=2, R current=0 box=1; Harry Ford: H current=1 box=0, BB current=1 box=0	WAS
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

Team	Starter	Projected	Actual	ER err
PIT	Jared Jones	1 ER / 17 outs	0 ER / 12 outs	-1
DET	River Ryan	1 ER / 9 outs	2 ER / 4 outs	1

Signal	Home-away edge	Meaning
Season quality	0.036	positive favors home; negative favors away
Recent form	-0.282	positive favors home; negative favors away
Starting-pitcher matchup	0.006	positive favors home; negative favors away
Composite	-0.218	positive favors home; negative favors away

Stage	Away-home runs	Confidence	What changed
historical_ml	4.88-3.91	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-3	0.505	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	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