Deep slate report 20260831

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_20260831.json	n/a
forensic deep audit	loaded	probability, provenance, starter and decision contracts	/app/output/reports/PREDICTION_SYSTEM_DEEP_AUDIT_20260831.json	n/a
continuous learning	loaded	pipeline-stage quality, calibration, probability lineage	/app/output/learning/continuous_learning_daily_20260831.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 (12)

Severity	Domain	Finding and evidence	Affected	Action & guardrail
HIGH	pipeline_stages	An intermediate stage materially degraded winner direction historical_ml → variance_applied changed winner accuracy by -0.1667 (-2 hits)	BAL@COL, NYY@LAA	Do thisList the exact games whose winner flipped at this stage, and keep both the before-stage and after-stage picks in the trace. Don't concludeEven if a later stage recovers the pick, this unstable step is still a risk; do not treat it as harmless.
HIGH	score_distribution	Published margins are compressed relative to realized margins 11/12 predictions were ≤2 runs; 6/12 actual margins were ≥4; mean predicted margin 1.42 vs actual 4.92	DET@MIN, MIL@CHC, NYY@LAA, OAK@TEX, SD@CIN, SF@ATL	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	calibration	Published confidence buckets were poorly aligned on this slate weighted absolute calibration error 0.1661 across 12 games	BAL@COL, CHW@HOU, DET@MIN, MIA@WAS, MIL@CHC, NYM@TB, NYY@LAA, OAK@TEX, PHI@ARI, SD@CIN, SEA@BOS, SF@ATL	Do thisWatch the same confidence buckets over the full rolling window before changing any calibration weight. Don't concludeA single slate has far too few games to establish a stable calibration curve.
HIGH	model_validation	Validated run models explain very little holdout variance home/away/total R² = 0.0378/0.0275/0.0272	BAL@COL, CHW@HOU, DET@MIN, MIA@WAS, MIL@CHC, NYM@TB, NYY@LAA, OAK@TEX, PHI@ARI, SD@CIN, SEA@BOS, SF@ATL	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 12/12 games structurally audited	BAL@COL, CHW@HOU, DET@MIN, MIA@WAS, MIL@CHC, NYM@TB, NYY@LAA, OAK@TEX, PHI@ARI, SD@CIN, SEA@BOS, SF@ATL	Do thisKeep the per-game audit receipt in every report. Don't concludeFull audit coverage does not cancel the findings raised inside the audited games.
INFO	probability_lineage	Material calibrated-to-published movements have recorded lineage 8 material movement(s), all explained by named score-coherence and/or recency transformations	CHW@HOU, DET@MIN, MIA@WAS, MIL@CHC, NYY@LAA, OAK@TEX, SEA@BOS, SF@ATL	Do thisKeep showing every probability layer and the named transformation between them. Don't concludeA published probability above the calibrated one is not a governor bypass when a recorded transformation explains it.
INFO	simulation	Simulation produced no winner-direction intervention on this slate flip candidates 0; published flips 0	BAL@COL, CHW@HOU, DET@MIN, MIA@WAS, MIL@CHC, NYM@TB, NYY@LAA, OAK@TEX, PHI@ARI, SD@CIN, SEA@BOS, SF@ATL	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	4 consecutive · 4 in window	simulation.allow_score	confirm the simulation module is running, not silently no-oping	simulation_inert_on_slate has persisted 4 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	4 consecutive · 4 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 4 consecutive slates; this is no longer single-slate noise. Treat it as an actionable pattern and raise the code/diagnostic action named; do not force a parameter move.

Finding	Consecutive slates	Incidence	Present now	Fix type
high_raw_probability_cluster	0	1/4	no	param
pipeline_winner_instability	2	2/4	yes	code/diagnostic
score_margin_compression	1	1/4	yes	param
simulation_inert_on_slate	4	4/4	yes	code/diagnostic
single_slate_calibration_instability	1	3/4	yes	param
weak_run_model_signal	4	4/4	yes	code/diagnostic

Metric	Direction	Series (old → new)
winner_accuracy	flat	[0.4667, 0.9412, 0.5, 0.6667]
weighted_calibration_error	improving	[0.2179, 0.3961, 0.0693, 0.1661]
margin_bias	regressing	[-1.333, -1.882, -2.786, -3.5]
team_run_error_on_losses	regressing	[5.0, 3.0, 7.0, 8.25]

Stage	Winner hits	Accuracy	Team-run MAE	Total-run MAE	Margin MAE
historical_ml	8/12	66.7%	3.128	3.991	4.702
variance_applied	6/12	50%	3.125	3.917	4.75
enhanced_prediction	8/12	66.7%	3.458	4.25	4.75
embedded_savant_run_environment	8/12	66.7%	3.417	4.333	4.667
fundamental_edge_target_review	8/12	66.7%	3.417	4.333	4.667
capped_target	8/12	66.7%	3.417	4.333	4.667
simulation_result	8/12	66.7%	3.417	4.333	4.667
published_final	8/12	66.7%	3.417	4.333	4.667

Transition	Δ accuracy	Δ hits	Δ team-run MAE	Δ margin MAE	Games whose winner/tie changed
historical_ml → variance_applied	-0.1667	-2	-0.0033	0.0477	BAL@COL: BAL → TIE; NYY@LAA: LAA → TIE; PHI@ARI: ARI → TIE
variance_applied → enhanced_prediction	0.1667	2	0.3333	0	BAL@COL: TIE → BAL; NYY@LAA: TIE → NYY; PHI@ARI: TIE → PHI
enhanced_prediction → embedded_savant_run_environment	0	0	-0.0416	-0.0833
embedded_savant_run_environment → fundamental_edge_target_review	0	0	0	0
fundamental_edge_target_review → capped_target	0	0	0	0
capped_target → simulation_result	0	0	0	0
simulation_result → published_final	0	0	0	0

Game	Outcome	Raw model	Calibrated	Governor	Score coherence	Published	Explanation
BAL@COL	hit	0.7585	0.7435	0.7435	0.7492	0.7492	score_coherence aligned probability to the published score margin
CHW@HOU	hit	0.5085	0.5082	0.5082	0.5466	0.53	score coherence followed by matchup-recency guardrail
DET@MIN	hit	0.5085	0.5082	0.5082	0.5466	0.53	score coherence followed by matchup-recency guardrail
MIA@WAS	hit	0.5085	0.5082	0.5082	0.5466	0.53	score coherence followed by matchup-recency guardrail
MIL@CHC	miss	0.5085	0.5082	0.5082	0.5466	0.53	score coherence followed by matchup-recency guardrail
NYM@TB	miss	0.67	0.661	0.661	0.6655	0.6655	score_coherence aligned probability to the published score margin
NYY@LAA	miss	0.5943	0.591	0.591	0.588	0.5509	score coherence followed by matchup-recency guardrail
OAK@TEX	hit	0.5085	0.5082	0.5082	0.5466	0.53	score coherence followed by matchup-recency guardrail
PHI@ARI	hit	0.5085	0.5082	0.5082	0.5466	0.5269	score coherence followed by matchup-recency guardrail
SD@CIN	hit	0.6947	0.677	0.677	0.6735	0.6735	score_coherence aligned probability to the published score margin
SEA@BOS	hit	0.585	0.5816	0.5816	0.5833	0.5602	score coherence followed by matchup-recency guardrail
SF@ATL	miss	0.5085	0.5082	0.5082	0.5466	0.53	score coherence followed by matchup-recency guardrail

Sides analyzed	Broke on pitching	Broke on offense	Both	Variance	Mean starter ERA error	Mean lineup runs vs proj
24	2	7	6	9	0.83	0.08

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

Game	Pred margin	Actual margin	Pred − actual
BAL@COL	3	1	2
CHW@HOU	1	3	-2
DET@MIN	1	10	-9
MIA@WAS	1	3	-2
MIL@CHC	1	14	-13
NYM@TB	2	1	1
NYY@LAA	1	9	-8
OAK@TEX	2	7	-5
PHI@ARI	1	1	0
SD@CIN	2	5	-3
SEA@BOS	1	1	0
SF@ATL	1	4	-3

BAL@COLHIT

Predicted: BAL 7-4 (conf -0.50, strong)Actual: BAL 2-1Brier 0.063total bias +8

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.50 (strong); this is not the half-scale distance from 50%
team-run absolute error 8; total bias +8
pregame home-minus-away evidence: season -1.044, recent -0.976, pitcher -0.025, composite -1.412
confidence pipeline reached 0.744 before later publication/recency controls; published home probability was 0.251
score-volatility actions: supported_blowout_margin_expansion
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 8 (11 vs 3)
BAL offense over-projected by 5
COL offense over-projected by 3

Team	Starter	Projected	Actual	ER err
BAL	Cade Povich	2 ER / 14 outs	1 ER / 15 outs	-1
COL	Tanner Gordon	2 ER / 15 outs	1 ER / 15 outs	-1

Decision internals: run factors home×1.013 / away×1.038; 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:

Pete Alonso (BAL 1B): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Gunnar Henderson (BAL SS): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.26]
Samuel Basallo (BAL C): 1-for-3, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.64]

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

Cade Povich (BAL) threw 15 outs, 1 ER = 1.8 ERA against a projected 3.86 ERA (-2.06 ERA error) only 6 prior appearances in the covered season - the projection had little of his own record to stand on

Bullpen behind him: 0 ER over 5 outs.

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

Who did the damage to him:

Troy Johnston: 1-for-3, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.329, above average - a good day, not a model error
TJ Rumfield: 1-for-2, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.334, above average - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

HOME pitching

Tanner Gordon (COL) threw 15 outs, 1 ER = 1.8 ERA against a projected 3.6 ERA (-1.8 ERA error)

Bullpen behind him: 0 ER over 6 outs.

The AWAY lineup he faced produced 2 runs on 4 hits and 0 HR, against a projected 5 runs / 8 hits (-3 runs vs projection).

Who did the damage to him:

Samuel Basallo: 1-for-3, 0 HR, 1 RBI [+0.64 runs vs projection] season xwOBA 0.322, above average - real projection error
Coby Mayo: 1-for-3, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.319, average - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
BAL	Cade Povich (Starter)	15 / 14	1.8 / 3.86	1 / 2	5 / 1	0 / 0	3 / 3	4 / 6	venue away: 9.9 vs 0.71 ERA (ok) · opp run factor 1.01 (average) · 6 prior appearances this season
BAL	Andrew Kittredge (Closer)	3 / 4	0 / 0	0 / 0	1 / 0	0 / 0	0 / 2	1 / 2
BAL	Rico Garcia (Middle Reliever (High Leverage))	2 / 4	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 0
COL	Tanner Gordon (Starter)	15 / 15	1.8 / 3.6	1 / 2	3 / 2	0 / 1	1 / 4	3 / 6	venue home: 9.33 vs 1.59 ERA (worse here) · opp run factor 1.04 (average) · 19 prior appearances this season
COL	Brennan Bernardino (Setup Man)	3 / 3	0 / 18 (-18.00)	0 / 2	1 / 3	0 / 1	0 / 1	0 / 0
COL	Jordan Romano (Closer)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
BAL	Pete Alonso (1B)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	1 / 1	-1.4	0.392 elite	bad day
BAL	Gunnar Henderson (SS)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	0 / 1	-1.26	0.316 average	no signal
BAL	Samuel Basallo (C)	3 / 4	1 / 1	1 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 0	1 / 1	0.64	0.322 above average	projection too low
BAL	Colton Cowser (CF)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.321 above average	—
BAL	Coby Mayo (DH)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	1 / 0	0 / 1	0.33	0.319 average	—
BAL	Dylan Beavers (LF)	3 / 3	1 / 1	1 / 1	0 / 0	0 / 0	0 / 1	1 / 0	0 / 1	0 / 1	-0.33	0.304 below average	—
COL	Willi Castro (3B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	3 / 1	-0.33	0.309 average	—
COL	Troy Johnston (DH)	3 / 3	1 / 1	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	0.33	0.329 above average	—
COL	TJ Rumfield (1B)	2 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0 / 1	0.33	0.334 above average	—
COL	Connor Norby (2B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	0 / 1	-0.33	0.301 below average	—
COL	Jake McCarthy (CF)	5 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	0 / 1	-0.31	0.319 average	—
COL	Hunter Goodman (C)	3 / 5	1 / 1	1 / 0	0 / 0	0 / 1	0 / 1	1 / 1	1 / 0	1 / 1	-0.29	0.324 above average	—
BAL	Leody Taveras (RF)	1 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	2 / 0	0 / 1	0.19	0.273 weak	—
BAL	Blaze Alexander (3B)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	0 / 1	-0.14	0.356 elite	—
BAL	Jackson Holliday (2B)	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	0	0.303 below average	—
COL	Mickey Moniak (RF)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	0 / 1	0	0.281 weak	—
COL	Jordan Beck (LF)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	1 / 1	0	0.243 weak	—
COL	Ezequiel Tovar (SS)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.278 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-1.044	positive favors home; negative favors away
Recent form	-0.976	positive favors home; negative favors away
Starting-pitcher matchup	-0.025	positive favors home; negative favors away
Composite	-1.412	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	5.03-4.67	n/a	Base ML historical model prediction
variance_applied	5-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	7-4	0.749	Enhanced ML with recent form, H2H, low-score analysis
embedded_savant_run_environment	7-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	7-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	7-4	n/a	Capped at max 12 realistic runs
simulation_result	7-4	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	5.03-4.67	5-5	-0.354	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-5	7-4	3	Enhanced ML with recent form, H2H, low-score analysis
embedded_savant_run_environment	7-4	7-4	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	7-4	7-4	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	7-4	7-4	-0	Capped at max 12 realistic runs
simulation_result	7-4	7-4	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
COL	Connor Norby	batter / 2B	better	predicted_RBI 2.52
BAL	Dylan Beavers	batter / LF	worse	predicted_AB 1.948
BAL	Jackson Holliday	batter / 2B	boosted	predicted_AB 1.8
COL	Hunter Goodman	batter / C	worse	predicted_AB -1.73
COL	Willi Castro	batter / 3B	worse	predicted_BB 1.653
BAL	Gunnar Henderson	batter / SS	worse	predicted_BB 1.576
BAL	Leody Taveras	batter / RF	better	predicted_RBI 1.544
BAL	Coby Mayo	batter / DH	better	predicted_RBI 1.466

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	BAL pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	BAL
info	CURRENT_PITCHER_DUPLICATION	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	BAL model-prediction batting totals differ from final simulation box score (AB: current=31, sim=37; H: current=8, sim=10; 3B: current=0, sim=1; HR: current=2, sim=4; RBI: current=5, sim=7; R: current=5, sim=7; BB: current=3, sim=6; SO: current=9, sim=6).	BAL
info	CURRENT_PLAYER_MATCH_BOX	BAL current finalized hitter lines are model predictions and differ from the simulation box score: Coby Mayo: AB current=3 box=4, H current=1 box=2, HR current=0 box=1, RBI current=1 box=2; Dylan Beavers: AB current=3 box=5, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Gunnar Henderson: AB current=5 box=3, HR current=1 box=0, RBI current=1 box=0, BB current=1 box=2; Samuel Basallo: HR current=0 box=1, RBI current=0 box=1, R current=0 box=1, BB current=0 box=1; Jackson Holliday: AB current=3 box=5, H current=0 box=1, 3B current=0 box=1, RBI current=0 box=1; Pete Alonso: SO current=1 box=0; Leody Taveras: AB current=3 box=4, HR current=0 box=1, RBI current=1 box=2, R current=0 box=1; Blaze Alexander: AB current=3 box=4, H current=1 box=2, 2B current=0 box=1, BB current=0 box=1	BAL
info	CURRENT_FINALIZED_MATCH_BOX	COL model-prediction batting totals differ from final simulation box score (AB: current=34, sim=32; H: current=8, sim=5; HR: current=1, sim=0; BB: current=3, sim=5).	COL
info	CURRENT_PLAYER_MATCH_BOX	COL current finalized hitter lines are model predictions and differ from the simulation box score: Jake McCarthy: AB current=4 box=5, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Willi Castro: AB current=4 box=2, H current=1 box=0, R current=1 box=0, BB current=1 box=2; Mickey Moniak: H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Hunter Goodman: AB current=5 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Troy Johnston: AB current=3 box=4, H current=1 box=2, RBI current=0 box=1, R current=0 box=1; TJ Rumfield: AB current=4 box=3, R current=0 box=1; Connor Norby: AB current=4 box=3, 2B current=0 box=1, RBI current=1 box=3, R current=0 box=1; Jordan Beck: AB current=3 box=4, R current=0 box=1, SO current=1 box=2	COL
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

CHW@HOUHIT

Predicted: HOU 4-5 (conf +0.06, lean)Actual: HOU 3-6Brier 0.221total bias +0

Root cause: Solid correct call

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 2; total bias +0
pregame home-minus-away evidence: season -0.516, recent -0.103, pitcher +0.059, composite -0.001
confidence pipeline reached 0.508 before later publication/recency controls; published home probability was 0.530
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

No large single-factor errors.

Team	Starter	Projected	Actual	ER err
CHW	Anthony Kay	3 ER / 16 outs	5 ER / 14 outs	2
HOU	Peter Lambert	1 ER / 16 outs	2 ER / 14 outs	1

Decision internals: run factors home×1.04 / away×0.982; 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:

Colson Montgomery (CHW SS): 2-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 2.37]
Cam Smith (HOU RF): 2-for-4, 0 HR, 2 RBI (projected 0.0 H, 0.0 HR) [run impact 1.87]
Jeremy Peña (HOU SS): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.93]
Nelson Velázquez (HOU LF): 1-for-3, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Braden Montgomery (CHW RF): 0-for-5, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]

Starter miss: Anthony Kay (CHW) posted 9.64 ERA vs projected 5.06 (5 ER in 14 outs); he is materially worse pitching away (4.89 ERA there vs 3.44 on the other side), a venue-role weakness the model under-weighted.

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

Pitcher versus batter: what actually broke this game

AWAY pitching

Anthony Kay (CHW) threw 14 outs, 5 ER = 9.64 ERA against a projected 5.06 ERA (4.58 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 6 runs on 8 hits and 1 HR, against a projected 4 runs / 7 hits (2 runs vs projection).

Who did the damage to him:

Cam Smith: 2-for-4, 0 HR, 2 RBI [+1.87 runs vs projection] season xwOBA 0.319, average - a good day, not a model error
Nelson Velázquez: 1-for-3, 1 HR, 2 RBI [+0.93 runs vs projection] season xwOBA 0.348, elite - real projection error
Christian Vázquez: 1-for-3, 0 HR, 0 RBI [+0.8 runs vs projection] season xwOBA 0.238, weak - a good day, not a model error

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

HOME pitching

Peter Lambert (HOU) threw 14 outs, 2 ER = 3.86 ERA against a projected 1.69 ERA (2.17 ERA error)

Bullpen behind him: 0 ER over 6 outs.

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

Who did the damage to him:

Colson Montgomery: 2-for-3, 1 HR, 1 RBI [+2.37 runs vs projection] season xwOBA 0.307, average - a good day, not a model error
Chase Meidroth: 1-for-3, 0 HR, 1 RBI [+0.35 runs vs projection] season xwOBA 0.292, below average - a good day, not a model error
Sam Antonacci: 1-for-4, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.342, elite - a good day, not a model error

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

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
CHW	Anthony Kay (Starter)	14 / 16	9.64 / 5.06 (+4.58)	5 / 3	6 / 4	1 / 0	2 / 0	5 / 4	venue away: 4.89 vs 3.44 ERA (worse here) · opp run factor 1.04 (average) · 27 prior appearances this season
HOU	Peter Lambert (Starter)	14 / 16	3.86 / 1.69	2 / 1	3 / 3	0 / 1	5 / 3	6 / 7	venue home: 4.05 vs 3 ERA (worse here) · opp run factor 0.98 (average) · 23 prior appearances this season
HOU	Bryan Abreu (Middle Reliever (High Leverage))	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	2 / 1
HOU	Steven Okert (Setup Man)	3 / 3	0 / 27 (-27.00)	0 / 3	1 / 4	0 / 1	0 / 1	2 / 1

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
CHW	Colson Montgomery (SS)	3 / 4	2 / 1	1 / 0	0 / 0	1 / 0	1 / 1	2 / 0	2 / 0	1 / 1	2.37	0.307 average	no signal
HOU	Cam Smith (RF)	4 / 3	2 / 0	1 / 0	1 / 0	0 / 0	2 / 0	1 / 0	0 / 0	1 / 1	1.87	0.319 average	no signal
HOU	Jeremy Peña (SS)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	1 / 1	-0.93	0.326 above average	bad day
HOU	Nelson Velázquez (LF)	3 / 3	1 / 1	0 / 0	0 / 0	1 / 0	2 / 0	1 / 0	0 / 0	2 / 1	0.93	0.348 elite	projection too low
CHW	Braden Montgomery (RF)	5 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	-0.8	0.329 above average	bad day
HOU	Christian Vázquez (C)	3 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	1 / 1	0.8	0.238 weak	good day
CHW	Tristan Peters (CF)	2 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	-0.78	0.319 average	no signal
HOU	Christian Walker (1B)	3 / 4	0 / 1	0 / 1	0 / 0	0 / 0	1 / 1	0 / 1	1 / 1	0 / 1	-0.78	0.3 below average	projection too high
CHW	Munetaka Murakami (1B)	2 / 4	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	3 / 1	1 / 2	-0.74	0.35 elite	bad day
CHW	Andrew Benintendi (DH)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	2 / 1	-0.47	0.31 average	—
HOU	Daulton Varsho (CF)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.292 below average	—
CHW	Chase Meidroth (2B)	3 / 4	1 / 1	0 / 1	0 / 0	0 / 0	1 / 0	1 / 1	2 / 0	0 / 1	0.35	0.292 below average	—
CHW	Sam Antonacci (LF)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	1 / 0	0.33	0.342 elite	—
HOU	Isaac Paredes (3B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	0 / 1	0 / 0	-0.33	0.32 average	—
CHW	Miguel Vargas (3B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	0	0.386 elite	—
HOU	Yordan Alvarez (DH)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	1 / 1	1 / 1	1 / 1	0 / 1	0	0.45 elite	—
HOU	Nick Allen (2B)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	0	0.252 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.516	positive favors home; negative favors away
Recent form	-0.103	positive favors home; negative favors away
Starting-pitcher matchup	0.059	positive favors home; negative favors away
Composite	-0.001	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.13-5.15	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.53	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	n/a	Capped at max 12 realistic runs
simulation_result	4-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.13-5.15	4-5	-0.022	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 projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	4-5	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	4-5	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	4-5	0	Capped at max 12 realistic runs
simulation_result	4-5	4-5	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
CHW	Colson Montgomery	batter / SS	worse	predicted_AB -2.46
HOU	Cam Smith	batter / RF	boosted	predicted_RBI 1.74
CHW	Sam Antonacci	batter / LF	better	predicted_AB 1.539
CHW	Braden Montgomery	batter / RF	better	predicted_RBI 1.485
CHW	Munetaka Murakami	batter / 1B	worse	predicted_SO -1.475
HOU	Christian Walker	batter / 1B	better	predicted_RBI 1.47
HOU	Isaac Paredes	batter / 3B	worse	predicted_SO 1.434
HOU	Yordan Alvarez	batter / DH	worse	predicted_SO 1.246

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	CHW pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	CHW
info	CURRENT_PITCHER_DUPLICATION	HOU pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	HOU
info	CURRENT_FINALIZED_MATCH_BOX	CHW model pitcher projections differ from final simulation pitching box score (H: current=6, sim=7; BB: current=0, sim=1; HR: current=1, sim=2).	CHW
info	CURRENT_FINALIZED_MATCH_BOX	CHW model-prediction batting totals differ from final simulation box score (AB: current=33, sim=34; H: current=8, sim=7; HR: current=1, sim=2; RBI: current=5, sim=4; R: current=5, sim=4; BB: current=4, sim=6).	CHW
info	CURRENT_PLAYER_MATCH_BOX	CHW current finalized hitter lines are model predictions and differ from the simulation box score: Sam Antonacci: AB current=3 box=5, H current=1 box=2, 2B current=0 box=1, HR current=0 box=1; Munetaka Murakami: AB current=4 box=5, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Braden Montgomery: AB current=4 box=3, HR current=0 box=1, RBI current=1 box=2, SO current=1 box=2; Andrew Benintendi: AB current=3 box=4, H current=1 box=2, 2B current=0 box=1, R current=0 box=1; Miguel Vargas: AB current=4 box=5, R current=1 box=0, BB current=1 box=0, SO current=1 box=2; Chase Meidroth: H current=1 box=0, 2B current=1 box=0, R current=1 box=0, SO current=1 box=0; Colson Montgomery: AB current=4 box=2, H current=1 box=0, RBI current=1 box=0, BB current=0 box=2; Tristan Peters: AB current=4 box=3, 2B current=1 box=0, RBI current=1 box=0, R current=0 box=1	CHW
info	CURRENT_FINALIZED_MATCH_BOX	HOU model-prediction batting totals differ from final simulation box score (AB: current=33, sim=32; HR: current=1, sim=2; RBI: current=4, sim=5; R: current=4, sim=5; BB: current=3, sim=1; SO: current=8, sim=7).	HOU
info	CURRENT_PLAYER_MATCH_BOX	HOU current finalized hitter lines are model predictions and differ from the simulation box score: Isaac Paredes: 2B current=0 box=1, RBI current=1 box=0, BB current=1 box=0, SO current=0 box=2; Jeremy Peña: AB current=5 box=4, HR current=1 box=0, RBI current=1 box=0, SO current=1 box=0; Yordan Alvarez: H current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Christian Walker: AB current=4 box=3, H current=1 box=2, 2B current=1 box=0, RBI current=1 box=2; Nick Allen: H current=1 box=0, SO current=1 box=0; Christian Vázquez: H current=0 box=1, HR current=0 box=1, RBI current=0 box=1, R current=0 box=1; Cam Smith: AB current=3 box=4, H current=0 box=1, HR current=0 box=1, RBI current=0 box=2; Daulton Varsho: H current=1 box=0, SO current=1 box=0	HOU
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

DET@MINHIT

Predicted: MIN 4-5 (conf +0.06, lean)Actual: MIN 1-11Brier 0.221total bias -3

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 9; total bias -3
pregame home-minus-away evidence: season -0.484, recent +0.302, pitcher -0.006, composite -0.025
confidence pipeline reached 0.508 before later publication/recency controls; published home probability was 0.530
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs under-projected by 3 (9 vs 12)
DET offense over-projected by 3
MIN offense under-projected by 6
Taj Bradley (MIN) allowed 3 fewer ER than projected (1 vs 4)

Team	Starter	Projected	Actual	ER err
DET	Jackson Jobe	3 ER / 14 outs	4 ER / 15 outs	1
MIN	Taj Bradley	4 ER / 16 outs	1 ER / 21 outs	-3

Decision internals: run factors home×1.026 / away×1.032; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

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

Top deviations that drove the game:

Ryan Jeffers (MIN C): 2-for-3, 2 HR, 4 RBI (projected 1.0 H, 0.0 HR) [run impact 2.33]
Dillon Dingler (DET C): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Hao-Yu Lee (DET 3B): 2-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Kody Clemens (MIN 2B): 1-for-2, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 1.26]
Walker Jenkins (MIN CF): 1-for-3, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 0.95]

Starter miss: Taj Bradley (MIN) posted 1.29 ERA vs projected 6.75 (1 ER in 21 outs); no credible venue, opponent-strength or experience signal explains it — an outlier start (variance).

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

Pitcher versus batter: what actually broke this game

AWAY pitching

Jackson Jobe (DET) threw 15 outs, 4 ER = 7.2 ERA against a projected 5.79 ERA (1.41 ERA error) only 4 prior appearances in the covered season - the projection had little of his own record to stand on

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 11 runs on 9 hits and 4 HR, against a projected 4 runs / 9 hits (7 runs vs projection).

Who did the damage to him:

Ryan Jeffers: 2-for-3, 2 HR, 4 RBI [+2.33 runs vs projection] season xwOBA 0.356, elite - real projection error
Kody Clemens: 1-for-2, 1 HR, 2 RBI [+1.26 runs vs projection] season xwOBA 0.341, elite - real projection error
Walker Jenkins: 1-for-3, 1 HR, 2 RBI [+0.95 runs vs projection] season xwOBA 0.253, weak - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

HOME pitching

Taj Bradley (MIN) threw 21 outs, 1 ER = 1.29 ERA against a projected 6.75 ERA (-5.46 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The AWAY lineup he faced produced 1 runs on 5 hits and 1 HR, against a projected 3 runs / 7 hits (-2 runs vs projection).

Who did the damage to him:

Hao-Yu Lee: 2-for-3, 1 HR, 1 RBI [+1.4 runs vs projection] season xwOBA 0.301, below average - a good day, not a model error
Zach McKinstry: 1-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.282, weak - a good day, not a model error
Brett Callahan: 1-for-3, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.342, elite - a good day, not a model error

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

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
DET	Jackson Jobe (Starter)	15 / 14	7.2 / 5.79	4 / 3	6 / 3	2 / 1	2 / 3	4 / 6	venue away: 0.87 vs 7.88 ERA (ok) · opp run factor 1.03 (average) · thin sample: 4 prior appearances this season (high uncertainty)
MIN	Taj Bradley (Starter)	21 / 16	1.29 / 6.75 (-5.46)	1 / 4	4 / 4	1 / 1	3 / 3	9 / 6	venue home: 3.51 vs 4.35 ERA (ok) · opp run factor 1.03 (average) · 26 prior appearances this season

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
MIN	Ryan Jeffers (C)	3 / 4	2 / 1	0 / 0	0 / 0	2 / 0	4 / 0	2 / 1	1 / 1	1 / 0	2.33	0.356 elite	projection too low
DET	Dillon Dingler (C)	4 / 5	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	1 / 1	-1.4	0.344 elite	bad day
DET	Hao-Yu Lee (3B)	3 / 3	2 / 1	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 0	1 / 1	1.4	0.301 below average	good day
MIN	Kody Clemens (2B)	2 / 4	1 / 1	0 / 0	0 / 0	1 / 0	2 / 1	2 / 0	1 / 0	1 / 1	1.26	0.341 elite	projection too low
MIN	Walker Jenkins (CF)	3 / 3	1 / 1	0 / 1	0 / 0	1 / 0	2 / 0	2 / 0	1 / 0	0 / 0	0.95	0.253 weak	good day
MIN	Brooks Lee (3B)	5 / 5	1 / 1	0 / 0	0 / 0	0 / 1	3 / 1	0 / 1	0 / 0	0 / 1	-0.93	0.274 weak	projection too high
DET	Kevin McGonigle (SS)	4 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	-0.8	0.35 elite	bad day
MIN	Luke Keaschall (RF)	3 / 4	2 / 1	1 / 0	0 / 0	0 / 0	0 / 0	3 / 1	1 / 1	0 / 1	0.78	0.315 average	no signal
DET	Gleyber Torres (2B)	4 / 5	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	-0.64	0.327 above average	bad day
MIN	Trevor Larnach (LF)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	2 / 1	-0.64	0.329 above average	bad day
MIN	Kaelen Culpepper (SS)	3 / 4	1 / 1	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	0 / 1	0.64	0.32 average	no signal
DET	Zach McKinstry (RF)	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	0.47	0.282 weak	—
DET	Max Clark (CF)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	-0.47	0.321 above average	—
MIN	Josh Bell (DH)	5 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	0 / 1	-0.47	0.333 above average	—
DET	Brett Callahan (LF)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	2 / 1	0.33	0.342 elite	—
DET	Colt Keith (DH)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	2 / 1	-0.14	0.335 above average	—
MIN	Royce Lewis (1B)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	1 / 1	-0.14	0.316 average	—
DET	Spencer Torkelson (1B)	1 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0	0.31 average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.484	positive favors home; negative favors away
Recent form	0.302	positive favors home; negative favors away
Starting-pitcher matchup	-0.006	positive favors home; negative favors away
Composite	-0.025	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.93-4.9	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.53	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	n/a	Capped at max 12 realistic runs
simulation_result	4-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.93-4.9	4-5	0.027	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 projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	4-5	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	4-5	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	4-5	0	Capped at max 12 realistic runs
simulation_result	4-5	4-5	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
MIN	Royce Lewis	batter / 1B	better	predicted_RBI 4.63
DET	Kevin McGonigle	batter / SS	deboosted	predicted_AB -2.7
DET	Gleyber Torres	batter / 2B	worse	predicted_AB -1.63
DET	Dillon Dingler	batter / C	worse	predicted_RBI 1.52
DET	Brett Callahan	batter / LF	worse	predicted_AB 1.5
MIN	Trevor Larnach	batter / LF	better	predicted_AB -1.5
DET	Zach McKinstry	batter / RF	worse	predicted_SO 1.443
MIN	Brooks Lee	batter / 3B	worse	predicted_SO 1.328

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	DET pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	DET
info	CURRENT_PITCHER_DUPLICATION	MIN pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	MIN
info	CURRENT_FINALIZED_MATCH_BOX	DET model-prediction batting totals differ from final simulation box score (AB: current=34, sim=30; H: current=7, sim=4; 2B: current=1, sim=2; R: current=3, sim=4; SO: current=8, sim=9; SF: current=0, sim=1; HBP: current=0, sim=1).	DET
info	CURRENT_PLAYER_MATCH_BOX	DET current finalized hitter lines are model predictions and differ from the simulation box score: Colt Keith: AB current=3 box=4, H current=1 box=0, SO current=1 box=0; Dillon Dingler: AB current=5 box=4, RBI current=1 box=2, SO current=1 box=2; Brett Callahan: AB current=3 box=4, H current=1 box=0; Gleyber Torres: AB current=5 box=3, RBI current=1 box=0; Kevin McGonigle: AB current=5 box=2, 2B current=0 box=1, RBI current=1 box=2, BB current=1 box=0; Hao-Yu Lee: AB current=3 box=2, H current=1 box=0, BB current=0 box=1; Zach McKinstry: AB current=3 box=4, SO current=0 box=2; Spencer Torkelson: H current=0 box=1, R current=0 box=1, SO current=1 box=0	DET
info	CURRENT_FINALIZED_MATCH_BOX	MIN model-prediction batting totals differ from final simulation box score (AB: current=36, sim=29; H: current=9, sim=5; 2B: current=2, sim=1; HR: current=1, sim=2; RBI: current=4, sim=5; R: current=4, sim=5; SO: current=7, sim=6).	MIN
info	CURRENT_PLAYER_MATCH_BOX	MIN current finalized hitter lines are model predictions and differ from the simulation box score: Luke Keaschall: H current=1 box=0, R current=1 box=0, BB current=1 box=0; Kody Clemens: H current=1 box=0, RBI current=1 box=0; Trevor Larnach: AB current=4 box=2, RBI current=1 box=0, R current=0 box=1, SO current=1 box=0; Ryan Jeffers: AB current=4 box=3, H current=1 box=0, SO current=0 box=1; Brooks Lee: AB current=5 box=4, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Josh Bell: RBI current=1 box=0, SO current=1 box=0; Kaelen Culpepper: AB current=4 box=3, H current=1 box=0; Royce Lewis: AB current=4 box=3, H current=1 box=3, HR current=0 box=2, RBI current=0 box=5	MIN
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

MIA@WASHIT

Predicted: WAS 3-4 (conf +0.06, lean)Actual: WAS 3-6Brier 0.221total bias -2

Root cause: Solid correct call

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 2; total bias -2
pregame home-minus-away evidence: season -0.050, recent -0.122, pitcher -0.062, composite -0.015
confidence pipeline reached 0.508 before later publication/recency controls; published home probability was 0.530
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

Ryan Gusto (MIA) allowed 3 fewer ER than projected (1 vs 4)

Team	Starter	Projected	Actual	ER err
MIA	Ryan Gusto	4 ER / 14 outs	1 ER / 13 outs	-3
WAS	Will Dion	1 ER / 15 outs	0 ER / 9 outs	-1

Decision internals: run factors home×0.938 / 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:

Daylen Lile (WAS LF): 3-for-3, 2 HR, 4 RBI (projected 1.0 H, 1.0 HR) [run impact 1.87]
Jorbit Vivas (WAS 2B): 0-for-2, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Heriberto Hernández (MIA LF): 0-for-2, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.74]

Starter miss: Ryan Gusto (MIA) posted 2.08 ERA vs projected 7.71 (1 ER in 13 outs); he is materially worse pitching away (4.57 ERA there vs 3.62 on the other side), a venue-role weakness the model under-weighted.

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

Pitcher versus batter: what actually broke this game

AWAY pitching

Ryan Gusto (MIA) threw 13 outs, 1 ER = 2.08 ERA against a projected 7.71 ERA (-5.63 ERA error)

Bullpen behind him: 1 ER over 4 outs.

The HOME lineup he faced produced 6 runs on 11 hits and 2 HR, against a projected 5 runs / 8 hits (1 runs vs projection).

Who did the damage to him:

Daylen Lile: 3-for-3, 2 HR, 4 RBI [+1.87 runs vs projection] season xwOBA 0.329, above average - real projection error
Nasim Nuñez: 1-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.267, weak - a good day, not a model error
Andrew Pinckney: 2-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.338, above average - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

HOME pitching

Will Dion (WAS) threw 9 outs, 0 ER = 0 ERA against a projected 1.8 ERA (-1.8 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The AWAY lineup he faced produced 3 runs on 10 hits and 0 HR, against a projected 5 runs / 7 hits (-2 runs vs projection).

Who did the damage to him:

Javier Sanoja: 2-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.287, below average - a good day, not a model error
Esteury Ruiz: 1-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.292, below average - a good day, not a model error
Brian Navarreto: 1-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.167, weak - a good day, not a model error

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

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
MIA	Ryan Gusto (Starter)	13 / 14	2.08 / 7.71 (-5.63)	1 / 4	4 / 2	0 / 2	2 / 2	4 / 3	venue away: 4.57 vs 3.62 ERA (worse here) · opp run factor 0.94 (strong) · 18 prior appearances this season
MIA	Tyler Zuber (Setup Man)	4 / 4	6.75 / 0 (+6.75)	1 / 0	1 / 0	0 / 0	0 / 0	3 / 0
WAS	Will Dion (Starter)	9 / 15	0 / 1.8	0 / 1	2 / 0	0 / 0	1 / 2	1 / 7	opp run factor 1 (average) · 18 prior appearances this season

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
WAS	Daylen Lile (LF)	3 / 5	3 / 1	0 / 0	0 / 0	2 / 1	4 / 1	2 / 1	1 / 1	0 / 1	1.87	0.329 above average	projection too low
WAS	Jorbit Vivas (2B)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	-0.8	0.298 below average	projection too high
MIA	Heriberto Hernández (LF)	2 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	2 / 0	0 / 1	-0.74	0.353 elite	bad day
MIA	Javier Sanoja (2B)	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0.47	0.287 below average	—
MIA	Leo Jiménez (3B)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.314 average	—
MIA	Esteury Ruiz (RF)	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0.47	0.292 below average	—
MIA	Brian Navarreto (C)	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0.47	0.167 weak	—
WAS	Nasim Nuñez (SS)	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	0.47	0.267 weak	—
WAS	Andrew Pinckney (RF)	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	0.47	0.338 above average	—
MIA	Jakob Marsee (CF)	4 / 3	2 / 1	1 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0.45	0.305 average	—
MIA	Agustín Ramírez (DH)	2 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0 / 1	-0.33	0.275 weak	—
WAS	Andrés Chaparro (1B)	4 / 3	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 0	3 / 1	-0.31	0.354 elite	—
MIA	Kyle Stowers (1B)	4 / 4	2 / 1	0 / 1	0 / 0	0 / 0	2 / 1	0 / 1	0 / 1	1 / 2	-0.17	0.335 above average	—
WAS	Keibert Ruiz (C)	4 / 5	2 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 0	0.16	0.286 below average	—
WAS	Abimelec Ortiz (DH)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	1 / 1	-0.14	0.294 below average	—
WAS	Brady House (3B)	4 / 5	2 / 1	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0 / 1	1 / 1	0.14	0.336 above average	—
WAS	Jacob Young (CF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	1 / 1	-0.14	0.319 average	—
MIA	Otto Lopez (SS)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 0	1 / 1	0	0.324 above average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.05	positive favors home; negative favors away
Recent form	-0.122	positive favors home; negative favors away
Starting-pitcher matchup	-0.062	positive favors home; negative favors away
Composite	-0.015	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.16-5.84	n/a	Base ML historical model prediction
variance_applied	4-6	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	3-4	0.53	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	3-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-4	n/a	Capped at max 12 realistic runs
simulation_result	3-4	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.16-5.84	4-6	0.322	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-6	3-4	-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	3-4	3-4	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-4	3-4	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-4	3-4	0	Capped at max 12 realistic runs
simulation_result	3-4	3-4	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
MIA	Esteury Ruiz	batter / RF	worse	predicted_SO 3.04
WAS	Daylen Lile	batter / LF	better	predicted_RBI 2.33
MIA	Kyle Stowers	batter / 1B	better	predicted_RBI 1.75
MIA	Javier Sanoja	batter / 2B	better	predicted_SO 1.704
WAS	Andrés Chaparro	batter / 1B	worse	predicted_AB 1.5
MIA	Otto Lopez	batter / SS	worse	predicted_H -1.38
WAS	Jacob Young	batter / CF	worse	predicted_AB -1.342
WAS	Nasim Nuñez	batter / SS	worse	predicted_SO 1.24

Structural audit findings

Severity	Rule	Finding	Subject
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_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	MIA model-prediction batting totals differ from final simulation box score (H: current=7, sim=4; 2B: current=1, sim=2; R: current=5, sim=3; SO: current=9, sim=13; HBP: current=0, sim=1).	MIA
info	CURRENT_PLAYER_MATCH_BOX	MIA current finalized hitter lines are model predictions and differ from the simulation box score: Agustín Ramírez: AB current=3 box=4, H current=1 box=0, R current=1 box=0, BB current=1 box=0; Javier Sanoja: AB current=3 box=4, RBI current=0 box=1, R current=0 box=1, SO current=0 box=2; Leo Jiménez: 2B current=0 box=1, R current=0 box=1, BB current=0 box=1, SO current=1 box=0; Kyle Stowers: AB current=4 box=3, H current=1 box=2, HR current=0 box=1, RBI current=1 box=2; Heriberto Hernández: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Otto Lopez: AB current=5 box=4, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Esteury Ruiz: AB current=3 box=4, SO current=1 box=4; Jakob Marsee: H current=1 box=0, R current=1 box=0	MIA
info	CURRENT_FINALIZED_MATCH_BOX	WAS model pitcher projections differ from final simulation pitching box score (H: current=5, sim=4).	WAS
info	CURRENT_FINALIZED_MATCH_BOX	WAS model-prediction batting totals differ from final simulation box score (AB: current=33, sim=26; H: current=8, sim=2; 2B: current=2, sim=0; HR: current=1, sim=2; R: current=5, sim=4; SO: current=7, sim=4; 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: Abimelec Ortiz: AB current=3 box=4, H current=1 box=0, SO current=1 box=0; Andrés Chaparro: AB current=3 box=4, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Jorbit Vivas: AB current=3 box=2, H current=1 box=0, R current=0 box=1; Keibert Ruiz: AB current=5 box=3, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Nasim Nuñez: SO current=1 box=2; Andrew Pinckney: AB current=3 box=2, H current=1 box=0, R current=0 box=1, BB current=0 box=1; Daylen Lile: AB current=5 box=3, RBI current=1 box=3, SO current=1 box=0; Brady House: AB current=5 box=3, HR current=0 box=1, BB current=1 box=0, SO current=1 box=0	WAS
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

MIL@CHCMISS

Predicted: MIL 6-5 (conf -0.06, lean)Actual: CHC 3-17Brier 0.281total bias -9

Root cause: Starting-pitcher projection was wrong

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 15; total bias -9
pregame home-minus-away evidence: season -0.392, recent -0.350, pitcher -0.113, composite -0.731
confidence pipeline reached 0.508 before later publication/recency controls; published home probability was 0.470
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs under-projected by 9 (11 vs 20)
MIL offense over-projected by 3
CHC offense under-projected by 12
Kyle Harrison (MIL) allowed 6 more ER than projected (9 vs 3)

Team	Starter	Projected	Actual	ER err
MIL	Kyle Harrison	3 ER / 17 outs	9 ER / 11 outs	6
CHC	Clay Holmes	2 ER / 18 outs	0 ER / 18 outs	-2

Decision internals: run factors home×0.937 / away×1.05; 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 3B): 3-for-6, 3 HR, 4 RBI (projected 1.0 H, 1.0 HR) [run impact 2.47]
Carson Kelly (CHC C): 3-for-5, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 2.2]
Tyrone Taylor (CHC LF): 4-for-6, 1 HR, 3 RBI (projected 1.0 H, 0.0 HR) [run impact 2.03]
Seiya Suzuki (CHC RF): 3-for-5, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 1.56]
Cooper Pratt (MIL SS): 2-for-4, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]

Starter miss: Kyle Harrison (MIL) posted 22.09 ERA vs projected 4.76 (9 ER in 11 outs); he is materially worse pitching away (4.66 ERA there vs 1.26 on the other side), a venue-role weakness the model under-weighted.

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

Pitcher versus batter: what actually broke this game

AWAY pitching

Kyle Harrison (MIL) threw 11 outs, 9 ER = 22.09 ERA against a projected 4.76 ERA (17.33 ERA error)

Bullpen behind him: 0 ER over 4 outs.

The HOME lineup he faced produced 16 runs on 18 hits and 8 HR, against a projected 5 runs / 8 hits (11 runs vs projection).

Who did the damage to him:

Alex Bregman: 3-for-6, 3 HR, 4 RBI [+2.47 runs vs projection] season xwOBA 0.318, average - a good day, not a model error
Carson Kelly: 3-for-5, 1 HR, 2 RBI [+2.2 runs vs projection] season xwOBA 0.311, average - a good day, not a model error
Tyrone Taylor: 4-for-6, 1 HR, 3 RBI [+2.03 runs vs projection] season xwOBA 0.319, average - a good day, not a model error

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

HOME pitching

Clay Holmes (CHC) threw 18 outs, 0 ER = 0 ERA against a projected 3 ERA (-3 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The AWAY lineup he faced produced 2 runs on 7 hits and 1 HR, against a projected 5 runs / 9 hits (-3 runs vs projection).

Who did the damage to him:

Cooper Pratt: 2-for-4, 1 HR, 2 RBI [+1.4 runs vs projection] season xwOBA 0.313, average - a good day, not a model error
William Contreras: 1-for-2, 0 HR, 0 RBI [+0.31 runs vs projection] season xwOBA 0.333, above average - a good day, not a model error

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

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
MIL	Kyle Harrison (Starter)	11 / 17	22.09 / 4.76 (+17.33)	9 / 3	10 / 4	5 / 1	2 / 1	3 / 6	venue away: 4.66 vs 1.26 ERA (worse here) · opp run factor 0.94 (strong) · 21 prior appearances this season
MIL	Chad Patrick (Middle Reliever (High Leverage))	4 / 3	0 / 9 (-9.00)	0 / 1	0 / 2	0 / 0	2 / 0	0 / 0
CHC	Clay Holmes (Starter)	18 / 18	0 / 3 (-3.00)	0 / 2	5 / 1	0 / 0	2 / 0	8 / 5	venue home: 2.08 vs 2.42 ERA (ok) · opp run factor 1.05 (strong) · 13 prior appearances this season

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
CHC	Alex Bregman (3B)	6 / 5	3 / 1	0 / 0	0 / 0	3 / 1	4 / 1	3 / 1	0 / 1	0 / 1	2.47	0.318 average	no signal
CHC	Carson Kelly (C)	5 / 4	3 / 1	0 / 0	0 / 0	1 / 0	2 / 1	1 / 1	1 / 0	0 / 1	2.2	0.311 average	no signal
CHC	Tyrone Taylor (LF)	6 / 3	4 / 1	0 / 1	0 / 0	1 / 0	3 / 0	2 / 0	0 / 0	0 / 0	2.03	0.319 average	no signal
CHC	Seiya Suzuki (RF)	5 / 5	3 / 1	0 / 1	0 / 0	1 / 0	1 / 1	3 / 1	1 / 1	0 / 1	1.56	0.339 above average	projection too low
MIL	Cooper Pratt (SS)	4 / 3	2 / 1	0 / 0	0 / 0	1 / 0	2 / 0	1 / 0	0 / 0	0 / 1	1.4	0.313 average	no signal
CHC	Nico Hoerner (SS)	3 / 4	1 / 1	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	1 / 0	0 / 0	1.26	0.324 above average	projection too low
CHC	Matt Shaw (2B)	4 / 3	1 / 1	0 / 0	0 / 0	1 / 0	2 / 0	2 / 0	1 / 0	0 / 1	1.26	0.307 average	no signal
MIL	Brice Turang (2B)	4 / 5	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	2 / 1	-1.11	0.335 above average	bad day
MIL	Christian Yelich (DH)	3 / 5	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	2 / 1	-1.11	0.314 average	no signal
CHC	Pete Crow-Armstrong (CF)	5 / 6	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0 / 0	1 / 1	-0.93	0.369 elite	bad day
CHC	Miguel Amaya (DH)	3 / 3	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	2 / 0	1 / 0	1 / 1	0.8	0.289 below average	good day
MIL	Jackson Chourio (LF)	4 / 5	1 / 1	1 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0 / 0	2 / 1	-0.62	0.334 above average	bad day
MIL	Sal Frelick (RF)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0	-0.47	0.287 below average	—
MIL	William Contreras (C)	2 / 5	1 / 1	1 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0 / 1	0.31	0.333 above average	—
CHC	Michael Busch (1B)	5 / 4	1 / 1	1 / 0	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	1 / 1	0.31	0.328 above average	—
MIL	Jake Bauers (1B)	2 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	0 / 1	0	0.373 elite	—
MIL	David Hamilton (3B)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.275 weak	—
MIL	Garrett Mitchell (CF)	2 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.346 elite	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.392	positive favors home; negative favors away
Recent form	-0.35	positive favors home; negative favors away
Starting-pitcher matchup	-0.113	positive favors home; negative favors away
Composite	-0.731	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	5.08-3.94	n/a	Base ML historical model prediction
variance_applied	5-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-5	0.53	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	6-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-5	n/a	Capped at max 12 realistic runs
simulation_result	6-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	5.08-3.94	5-4	-0.137	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	6-5	-0	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	6-5	6-5	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-5	6-5	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-5	6-5	-0	Capped at max 12 realistic runs
simulation_result	6-5	6-5	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
CHC	Pete Crow-Armstrong	batter / CF	worse	predicted_AB -3.29
CHC	Seiya Suzuki	batter / RF	better	predicted_RBI 2.25
MIL	Christian Yelich	batter / DH	better	predicted_AB -1.68
MIL	Cooper Pratt	batter / SS	better	predicted_RBI 1.65
MIL	Sal Frelick	batter / RF	worse	predicted_SO 1.618
MIL	William Contreras	batter / C	worse	predicted_H -1.58
CHC	Tyrone Taylor	batter / LF	better	predicted_AB 1.471
CHC	Matt Shaw	batter / 2B	worse	predicted_AB 1.38

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	CHC pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	CHC
info	CURRENT_PITCHER_DUPLICATION	MIL pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	MIL
info	CURRENT_FINALIZED_MATCH_BOX	CHC model-prediction batting totals differ from final simulation box score (AB: current=37, sim=35; 2B: current=2, sim=3; HR: current=2, sim=1; BB: current=3, sim=2; HBP: current=0, sim=3).	CHC
info	CURRENT_PLAYER_MATCH_BOX	CHC current finalized hitter lines are model predictions and differ from the simulation box score: Tyrone Taylor: AB current=3 box=4, 2B current=1 box=0, HR current=0 box=1, RBI current=0 box=1; Nico Hoerner: AB current=4 box=3, H current=1 box=0, SO current=0 box=1, HBP current=0 box=1; Alex Bregman: H current=1 box=2, HR current=1 box=0, BB current=1 box=0; Pete Crow-Armstrong: AB current=6 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Miguel Amaya: AB current=3 box=4; Matt Shaw: AB current=3 box=4, H current=1 box=2, 2B current=0 box=1, R current=0 box=1; Seiya Suzuki: AB current=5 box=4, H current=1 box=2, RBI current=1 box=3; Carson Kelly: H current=1 box=0, RBI current=1 box=0, R current=1 box=0	CHC
info	CURRENT_FINALIZED_MATCH_BOX	MIL model-prediction batting totals differ from final simulation box score (AB: current=36, sim=31; H: current=9, sim=4; 2B: current=2, sim=1; RBI: current=5, sim=6; R: current=5, sim=6; BB: current=4, sim=2; SO: current=8, sim=10).	MIL
info	CURRENT_PLAYER_MATCH_BOX	MIL current finalized hitter lines are model predictions and differ from the simulation box score: Jackson Chourio: AB current=5 box=4, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Jake Bauers: AB current=3 box=4, RBI current=1 box=2, BB current=1 box=0, SO current=1 box=2; Brice Turang: AB current=5 box=4, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; William Contreras: AB current=5 box=4, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; David Hamilton: H current=1 box=0, SO current=1 box=0; Sal Frelick: H current=1 box=0, SO current=0 box=2; Christian Yelich: AB current=5 box=3, H current=1 box=2, 2B current=1 box=0, HR current=0 box=1; Garrett Mitchell: AB current=4 box=3, H current=1 box=0, R current=0 box=1, BB current=0 box=1	MIL
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

NYM@TBMISS

Predicted: TB 3-5 (conf +0.33, strong)Actual: NYM 3-2Brier 0.443total bias +3

Root cause: Starting-pitcher projection was wrong

Evidence-backed diagnosis

published system edge 0.33 (strong); this is not the half-scale distance from 50%
team-run absolute error 3; total bias +3
pregame home-minus-away evidence: season +0.796, recent +0.380, pitcher -0.046, composite +0.792
confidence pipeline reached 0.661 before later publication/recency controls; published home probability was 0.665
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 8 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 3 (8 vs 5)
TB offense over-projected by 3
Robert Stock (NYM) allowed 4 fewer ER than projected (0 vs 4)
Ian Seymour (TB) allowed 3 more ER than projected (3 vs 0)
committed to TB at 0.33 confidence and it lost

Team	Starter	Projected	Actual	ER err
NYM	Robert Stock	4 ER / 12 outs	0 ER / 16 outs	-4
TB	Ian Seymour	0 ER / 15 outs	3 ER / 19 outs	3

Decision internals: run factors home×0.959 / away×1.004; 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:

Francisco Lindor (NYM SS): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.73]
Junior Caminero (TB 3B): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.26]
Mark Vientos (NYM DH): 1-for-4, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Yandy Díaz (TB DH): 0-for-5, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]

Starter miss: Robert Stock (NYM) posted 0.00 ERA vs projected 9.00 (0 ER in 16 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

Robert Stock (NYM) threw 16 outs, 0 ER = 0 ERA against a projected 9 ERA (-9 ERA error) only 5 prior appearances in the covered season - the projection had little of his own record to stand on

Bullpen behind him: 2 ER over 7 outs.

The HOME lineup he faced produced 2 runs on 9 hits and 0 HR, against a projected 5 runs / 8 hits (-3 runs vs projection).

Who did the damage to him:

Liam Hicks: 2-for-4, 0 HR, 1 RBI [+0.47 runs vs projection] season xwOBA 0.316, average - a good day, not a model error
Chandler Simpson: 2-for-4, 0 HR, 1 RBI [+0.47 runs vs projection] season xwOBA 0.261, weak - a good day, not a model error
Richie Palacios: 1-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.305, average - a good day, not a model error

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

HOME pitching

Ian Seymour (TB) threw 19 outs, 3 ER = 4.26 ERA against a projected 0 ERA (4.26 ERA error)

Bullpen behind him: 0 ER over 3 outs.

The AWAY lineup he faced produced 2 runs on 9 hits and 1 HR, against a projected 3 runs / 8 hits (-1 runs vs projection).

Who did the damage to him:

Mark Vientos: 1-for-4, 1 HR, 1 RBI [+0.93 runs vs projection] season xwOBA 0.314, average - a good day, not a model error
Francisco Alvarez: 1-for-3, 0 HR, 1 RBI [+0.47 runs vs projection] season xwOBA 0.326, above average - a good day, not a model error
Jared Young: 1-for-4, 0 HR, 0 RBI [+0.31 runs vs projection] season xwOBA 0.309, average - a good day, not a model error

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

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
NYM	Robert Stock (Starter)	16 / 12	0 / 9 (-9.00)	0 / 4	5 / 5	0 / 1	0 / 3	4 / 5	venue away: 24 vs 4.58 ERA (ok) · opp run factor 0.96 (average) · 5 prior appearances this season
NYM	Daniel Duarte (Setup Man)	2 / 5	27 / 0 (+27.00)	2 / 0	3 / 0	0 / 0	1 / 0	0 / 0
NYM	Nathan Lavender (Middle Reliever (High Leverage))	2 / 3	0 / 0	0 / 0	1 / 1	0 / 0	0 / 1	0 / 1
NYM	Kodai Senga (Closer)	3 / 4	0 / 6.75 (-6.75)	0 / 1	0 / 1	0 / 0	1 / 0	1 / 1
TB	Ian Seymour (Starter)	19 / 15	4.26 / 0 (+4.26)	3 / 0	6 / 3	1 / 0	0 / 3	10 / 6	venue home: 5.58 vs 3 ERA (worse here) · opp run factor 1 (weak) · 41 prior appearances this season
TB	Garrett Cleavinger (Middle Reliever (High Leverage))	3 / 2	0 / 40.5 (-40.50)	0 / 3	0 / 4	0 / 1	0 / 0	1 / 0

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
NYM	Francisco Lindor (SS)	4 / 5	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	1 / 1	-1.73	0.338 above average	bad day
TB	Junior Caminero (3B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	1 / 1	-1.26	0.363 elite	bad day
NYM	Mark Vientos (DH)	4 / 3	1 / 1	0 / 0	0 / 0	1 / 0	1 / 1	1 / 0	0 / 0	1 / 1	0.93	0.314 average	no signal
TB	Yandy Díaz (DH)	5 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	3 / 1	-0.8	0.355 elite	bad day
NYM	Francisco Alvarez (C)	3 / 3	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	2 / 1	0.47	0.326 above average	—
TB	Jonathan Aranda (1B)	3 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	0 / 1	-0.47	0.35 elite	—
TB	Liam Hicks (C)	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0 / 0	0 / 0	0.47	0.316 average	—
TB	Chandler Simpson (LF)	4 / 4	2 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 0	0.47	0.261 weak	—
TB	Richie Palacios (2B)	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0.47	0.305 average	—
NYM	Marcus Semien (2B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	-0.33	0.305 average	—
NYM	Jared Young (1B)	4 / 3	1 / 1	1 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	0.31	0.309 average	—
NYM	Bo Bichette (3B)	4 / 4	1 / 1	1 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	1 / 1	0.31	0.325 above average	—
NYM	A.J. Ewing (CF)	4 / 4	2 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	-0.17	0.317 average	—
TB	Jonny DeLuca (RF)	3 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	1 / 0	2 / 1	0.02	0.265 weak	—
NYM	Carson Benge (RF)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	3 / 1	0	0.336 above average	—
NYM	Christopher Morel (LF)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.213 weak	—
TB	Cedric Mullins (CF)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0	0.267 weak	—
TB	Taylor Walls (SS)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 0	0 / 1	0	0.249 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.796	positive favors home; negative favors away
Recent form	0.38	positive favors home; negative favors away
Starting-pitcher matchup	-0.046	positive favors home; negative favors away
Composite	0.792	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.21-5.83	n/a	Base ML historical model prediction
variance_applied	3-6	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	3-5	0.665	Enhanced ML with recent form, H2H, low-score analysis
embedded_savant_run_environment	3-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-5	n/a	Capped at max 12 realistic runs
simulation_result	3-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.21-5.83	3-6	0.385	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	3-6	3-5	-1	Enhanced ML with recent form, H2H, low-score analysis
embedded_savant_run_environment	3-5	3-5	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	3-5	3-5	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	3-5	3-5	0	Capped at max 12 realistic runs
simulation_result	3-5	3-5	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
TB	Jonny DeLuca	batter / RF	better	predicted_SO 2.28
TB	Liam Hicks	batter / C	worse	predicted_RBI 1.78
NYM	A.J. Ewing	batter / CF	better	predicted_RBI 1.705
TB	Jonathan Aranda	batter / 1B	worse	predicted_AB -1.69
NYM	Francisco Alvarez	batter / C	worse	predicted_H 1.63
TB	Yandy Díaz	batter / DH	worse	predicted_H -1.38
NYM	Francisco Lindor	batter / SS	worse	predicted_H -1.21
NYM	Bo Bichette	batter / 3B	worse	predicted_SO 1.206

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	NYM pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	NYM
info	CURRENT_PITCHER_DUPLICATION	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	NYM model pitcher projections differ from final simulation pitching box score (H: current=7, sim=6).	NYM
info	CURRENT_FINALIZED_MATCH_BOX	NYM model-prediction batting totals differ from final simulation box score (AB: current=33, sim=36; H: current=8, sim=9; SO: current=9, sim=7; 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: Mark Vientos: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0, HBP current=0 box=1; Carson Benge: AB current=4 box=5, H current=1 box=0, R current=1 box=0, SO current=1 box=2; Christopher Morel: AB current=3 box=4, H current=1 box=0; Francisco Lindor: AB current=5 box=4, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; A.J. Ewing: 2B current=1 box=0, HR current=0 box=1, RBI current=0 box=2, R current=0 box=1; Francisco Alvarez: H current=0 box=2, BB current=0 box=1, SO current=1 box=0; Marcus Semien: AB current=4 box=3, H current=1 box=2, R current=0 box=1; Jared Young: AB current=3 box=4, H current=1 box=2, RBI current=1 box=0, SO current=1 box=0	NYM
info	CURRENT_FINALIZED_MATCH_BOX	TB model-prediction batting totals differ from final simulation box score (AB: current=34, sim=29; H: current=8, sim=6; RBI: current=4, sim=5; BB: current=3, sim=4; SO: current=6, sim=7; SF: current=0, sim=1).	TB
info	CURRENT_PLAYER_MATCH_BOX	TB current finalized hitter lines are model predictions and differ from the simulation box score: Yandy Díaz: AB current=5 box=4, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Jonathan Aranda: AB current=5 box=3, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Liam Hicks: H current=1 box=2, HR current=0 box=1, RBI current=0 box=2, BB current=0 box=1; Junior Caminero: AB current=4 box=3, HR current=1 box=0, SO current=1 box=0; Jonny DeLuca: R current=0 box=1, SO current=1 box=3; Cedric Mullins: R current=0 box=1, BB current=0 box=1; Taylor Walls: AB current=3 box=2, H current=1 box=0, RBI current=0 box=1, SF current=0 box=1	TB
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

NYY@LAAMISS

Predicted: NYY 5-4 (conf -0.10, moderate)Actual: LAA 1-10Brier 0.303total bias -2

Root cause: Backed the wrong side with real conviction

Evidence-backed diagnosis

published system edge 0.10 (moderate); this is not the half-scale distance from 50%
team-run absolute error 10; total bias -2
pregame home-minus-away evidence: season -1.373, recent -0.696, pitcher +0.046, composite -1.009
confidence pipeline reached 0.591 before later publication/recency controls; published home probability was 0.449
score-volatility actions: supported_blowout_margin_expansion
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

NYY offense over-projected by 4
LAA offense under-projected by 6
committed to NYY at 0.10 confidence and it lost

Team	Starter	Projected	Actual	ER err
NYY	Elmer Rodríguez	2 ER / 14 outs	4 ER / 11 outs	2
LAA	Walbert Ureña	2 ER / 17 outs	1 ER / 21 outs	-1

Decision internals: run factors home×0.998 / away×0.952; 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:

Vaughn Grissom (LAA 1B): 3-for-5, 0 HR, 4 RBI (projected 1.0 H, 0.0 HR) [run impact 1.56]
Luis García Jr. (NYY 1B): 0-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Wade Meckler (LAA LF): 2-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 1.13]
Spencer Jones (NYY RF): 1-for-3, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Zach Neto (LAA SS): 1-for-4, 0 HR, 1 RBI (projected 1.0 H, 1.0 HR) [run impact -0.93]

Starter miss: Elmer Rodríguez (NYY) posted 9.82 ERA vs projected 3.86 (4 ER in 11 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

Elmer Rodríguez (NYY) threw 11 outs, 4 ER = 9.82 ERA against a projected 3.86 ERA (5.96 ERA error) only 5 prior appearances in the covered season - the projection had little of his own record to stand on

Bullpen behind him: 3 ER over 3 outs.

The HOME lineup he faced produced 9 runs on 12 hits and 0 HR, against a projected 3 runs / 7 hits (6 runs vs projection).

Who did the damage to him:

Vaughn Grissom: 3-for-5, 0 HR, 4 RBI [+1.56 runs vs projection] season xwOBA 0.316, average - a good day, not a model error
Wade Meckler: 2-for-3, 0 HR, 0 RBI [+1.13 runs vs projection] season xwOBA 0.312, average - a good day, not a model error
Tyler Heineman: 1-for-3, 0 HR, 0 RBI [+0.78 runs vs projection] season xwOBA 0.243, weak - a good day, not a model error

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

HOME pitching

Walbert Ureña (LAA) threw 21 outs, 1 ER = 1.29 ERA against a projected 3.18 ERA (-1.89 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The AWAY lineup he faced produced 1 runs on 6 hits and 1 HR, against a projected 5 runs / 8 hits (-4 runs vs projection).

Who did the damage to him:

Spencer Jones: 1-for-3, 1 HR, 1 RBI [+0.93 runs vs projection] season xwOBA 0.329, above average - real projection error
Amed Rosario: 2-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.312, average - a good day, not a model error

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

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
NYY	Elmer Rodríguez (Starter)	11 / 14	9.82 / 3.86 (+5.96)	4 / 2	4 / 5	0 / 1	3 / 2	5 / 4	venue away: 3.24 vs 5.52 ERA (ok) · opp run factor 1 (weak) · 5 prior appearances this season
NYY	David Bednar (Closer)	3 / 3	27 / 0 (+27.00)	3 / 0	4 / 0	0 / 0	0 / 0	0 / 1
LAA	Walbert Ureña (Starter)	21 / 17	1.29 / 3.18	1 / 2	5 / 5	1 / 2	0 / 2	8 / 8	venue home: 2.54 vs 3.71 ERA (ok) · opp run factor 0.95 (strong) · 24 prior appearances this season

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
LAA	Vaughn Grissom (1B)	5 / 4	3 / 1	2 / 0	0 / 0	0 / 0	4 / 1	2 / 0	0 / 0	0 / 1	1.56	0.316 average	no signal
NYY	Luis García Jr. (1B)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	2 / 0	-1.4	0.348 elite	bad day
LAA	Wade Meckler (LF)	3 / 3	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	2 / 0	1 / 1	1.13	0.312 average	no signal
NYY	Spencer Jones (RF)	3 / 3	1 / 1	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 0	1 / 1	0.93	0.329 above average	projection too low
LAA	Zach Neto (SS)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 1	1 / 1	2 / 1	1 / 1	2 / 1	-0.93	0.305 average	no signal
NYY	Ben Rice (DH)	4 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	-0.8	0.378 elite	bad day
NYY	Trent Grisham (CF)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	2 / 1	-0.8	0.337 above average	bad day
LAA	Tyler Heineman (C)	3 / 3	1 / 0	1 / 0	0 / 0	0 / 0	0 / 0	2 / 0	0 / 0	0 / 1	0.78	0.243 weak	good day
NYY	Amed Rosario (3B)	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0.47	0.312 average	—
NYY	José Caballero (SS)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	2 / 1	-0.47	0.28 weak	—
LAA	Moisés Ballesteros (DH)	5 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	3 / 1	-0.47	0.291 below average	—
LAA	Christian Moore (2B)	5 / 4	2 / 1	1 / 0	0 / 0	0 / 0	2 / 0	1 / 0	0 / 1	1 / 1	0.45	0.238 weak	—
LAA	Josh Lowe (RF)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	0.33	0.29 below average	—
NYY	Cody Bellinger (LF)	4 / 5	2 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	-0.17	0.342 elite	—
LAA	Mike Trout (CF)	5 / 5	2 / 1	0 / 1	0 / 0	0 / 0	2 / 1	1 / 1	0 / 1	1 / 1	-0.17	0.376 elite	—
NYY	Jazz Chisholm Jr. (2B)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	0	0.294 below average	—
NYY	Austin Wells (C)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 1	0	0.291 below average	—
LAA	Oswald Peraza (3B)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	0	0.271 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-1.373	positive favors home; negative favors away
Recent form	-0.696	positive favors home; negative favors away
Starting-pitcher matchup	0.046	positive favors home; negative favors away
Composite	-1.009	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.98-4.44	n/a	Base ML historical model prediction
variance_applied	4-4	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	0.551	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	5-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	5-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	5-4	n/a	Capped at max 12 realistic runs
simulation_result	5-4	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.98-4.44	4-4	0.454	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 winner_starter_is_below_average, projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	5-4	5-4	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	5-4	5-4	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	5-4	5-4	-0	Capped at max 12 realistic runs
simulation_result	5-4	5-4	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
NYY	Spencer Jones	batter / RF	better	predicted_H 2.311
NYY	Amed Rosario	batter / 3B	better	predicted_AB 2.301
LAA	Mike Trout	batter / CF	worse	predicted_SO 1.864
LAA	Christian Moore	batter / 2B	better	predicted_RBI 1.64
NYY	Ben Rice	batter / DH	worse	predicted_H 1.54
NYY	Luis García Jr.	batter / 1B	worse	predicted_SO 1.394
NYY	Austin Wells	batter / C	worse	predicted_RBI 1.39
LAA	Tyler Heineman	batter / C	worse	predicted_AB 1.345

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	LAA pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	LAA
info	CURRENT_PITCHER_DUPLICATION	NYY pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	NYY
info	CURRENT_FINALIZED_MATCH_BOX	LAA model-prediction batting totals differ from final simulation box score (AB: current=34, sim=35; H: current=7, sim=8; 2B: current=1, sim=2; RBI: current=3, sim=4; R: current=3, sim=4; BB: current=3, sim=2).	LAA
info	CURRENT_PLAYER_MATCH_BOX	LAA current finalized hitter lines are model predictions and differ from the simulation box score: Wade Meckler: H current=1 box=0, R current=1 box=0, BB current=0 box=1; Vaughn Grissom: 2B current=0 box=1, RBI current=1 box=0, SO current=1 box=0; Christian Moore: 2B current=0 box=1, RBI current=0 box=2, R current=0 box=1, BB current=1 box=0; Josh Lowe: HR current=0 box=1, RBI current=0 box=1, R current=0 box=1, BB current=0 box=1; Oswald Peraza: AB current=3 box=4, H current=0 box=1, RBI current=0 box=1, R current=0 box=1; Moisés Ballesteros: H current=1 box=2, R current=0 box=1, SO current=1 box=2; Zach Neto: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Mike Trout: AB current=5 box=4, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0	LAA
info	CURRENT_FINALIZED_MATCH_BOX	NYY model-prediction batting totals differ from final simulation box score (AB: current=33, sim=39; H: current=8, sim=12; 2B: current=1, sim=2; HR: current=1, sim=2; RBI: current=6, sim=5; BB: current=3, sim=2; SO: current=8, sim=12).	NYY
info	CURRENT_PLAYER_MATCH_BOX	NYY current finalized hitter lines are model predictions and differ from the simulation box score: Cody Bellinger: AB current=5 box=4, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Amed Rosario: AB current=3 box=5, HR current=0 box=1, RBI current=0 box=2, R current=0 box=1; Ben Rice: H current=1 box=3, RBI current=1 box=0, BB current=1 box=0; Spencer Jones: AB current=3 box=4, H current=1 box=3, 2B current=0 box=1, R current=0 box=1; Luis García Jr.: AB current=4 box=5, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Jazz Chisholm Jr.: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=1, SO current=1 box=2; Trent Grisham: AB current=4 box=5, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Austin Wells: AB current=3 box=4, H current=0 box=1, 2B current=0 box=1, RBI current=1 box=2	NYY
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

OAK@TEXHIT

Predicted: TEX 3-5 (conf +0.06, lean)Actual: TEX 1-8Brier 0.221total bias -1

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 5; total bias -1
pregame home-minus-away evidence: season +0.830, recent +0.118, pitcher +0.006, composite +0.858
confidence pipeline reached 0.508 before later publication/recency controls; published home probability was 0.530
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

TEX offense under-projected by 3
Gage Jump (OAK) allowed 5 more ER than projected (7 vs 2)

Team	Starter	Projected	Actual	ER err
OAK	Gage Jump	2 ER / 15 outs	7 ER / 13 outs	5
TEX	Jacob deGrom	1 ER / 15 outs	0 ER / 21 outs	-1

Decision internals: run factors home×0.969 / away×0.963; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

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

Top deviations that drove the game:

Lawrence Butler (OAK RF): 1-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.26]
Justin Foscue (TEX 2B): 1-for-2, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Zack Gelof (OAK 3B): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Darell Hernaiz (OAK SS): 0-for-2, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Jonah Heim (OAK C): 1-for-3, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 0.78]

Starter miss: Gage Jump (OAK) posted 14.54 ERA vs projected 3.60 (7 ER in 13 outs); no credible venue, opponent-strength or experience signal explains it — an outlier start (variance).

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

Pitcher versus batter: what actually broke this game

AWAY pitching

Gage Jump (OAK) threw 13 outs, 7 ER = 14.54 ERA against a projected 3.6 ERA (10.94 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 7 runs on 8 hits and 1 HR, against a projected 5 runs / 8 hits (2 runs vs projection).

Who did the damage to him:

Justin Foscue: 1-for-2, 1 HR, 1 RBI [+0.93 runs vs projection] season xwOBA 0.331, above average - real projection error
Cody Freeman: 2-for-4, 0 HR, 2 RBI [+0.47 runs vs projection] season xwOBA 0.231, weak - a good day, not a model error
Danny Jansen: 0-for-3, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.25, weak - a good day, not a model error

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

HOME pitching

Jacob deGrom (TEX) threw 21 outs, 0 ER = 0 ERA against a projected 1.8 ERA (-1.8 ERA error)

Bullpen behind him: 0 ER over 0 outs.

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

Who did the damage to him:

Jonah Heim: 1-for-3, 0 HR, 0 RBI [+0.78 runs vs projection] season xwOBA 0.292, below average - a good day, not a model error

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

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
OAK	Gage Jump (Starter)	13 / 15	14.54 / 3.6 (+10.94)	7 / 2	6 / 6	1 / 1	3 / 2	2 / 8	venue away: 3.49 vs 5.62 ERA (ok) · opp run factor 0.97 (weak) · 17 prior appearances this season
TEX	Jacob deGrom (Starter)	21 / 15	0 / 1.8	0 / 1	1 / 3	0 / 1	0 / 2	8 / 8	venue home: 2.67 vs 5.86 ERA (ok) · opp run factor 0.96 (weak) · 25 prior appearances this season

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
OAK	Lawrence Butler (RF)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 1	2 / 1	-1.26	0.324 above average	bad day
TEX	Justin Foscue (2B)	2 / 3	1 / 1	0 / 0	0 / 0	1 / 0	1 / 1	1 / 0	0 / 0	0 / 1	0.93	0.331 above average	projection too low
OAK	Zack Gelof (3B)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	-0.8	0.316 average	no signal
OAK	Darell Hernaiz (SS)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 1	2 / 1	-0.8	0.262 weak	projection too high
OAK	Jonah Heim (C)	3 / 3	1 / 0	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0.78	0.292 below average	good day
OAK	Jeff McNeil (2B)	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	2 / 0	-0.78	0.307 average	no signal
OAK	Henry Bolte (CF)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	-0.47	0.311 average	—
OAK	Tommy White (1B)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	-0.47	0.298 below average	—
OAK	Max Muncy (DH)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	-0.47	0.293 below average	—
TEX	Elias Díaz (DH)	2 / 3	0 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 0	0 / 1	-0.47	0.278 weak	—
TEX	Wyatt Langford (LF)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	1 / 1	1 / 1	-0.47	0.313 average	—
TEX	Cody Freeman (3B)	4 / 3	2 / 1	0 / 0	0 / 0	0 / 0	2 / 0	0 / 0	0 / 0	0 / 0	0.47	0.231 weak	—
TEX	Danny Jansen (C)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	0.33	0.25 weak	—
TEX	Brandon Nimmo (RF)	3 / 5	2 / 1	0 / 1	0 / 0	0 / 0	2 / 1	1 / 1	0 / 1	0 / 1	-0.17	0.353 elite	—
TEX	Jake Burger (1B)	4 / 4	2 / 1	1 / 0	0 / 0	0 / 1	1 / 1	1 / 1	0 / 0	0 / 1	-0.15	0.312 average	—
TEX	Corey Seager (SS)	2 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	2 / 1	0 / 1	-0.14	0.351 elite	—
TEX	Ezequiel Duran (CF)	4 / 4	1 / 1	0 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 0	0 / 1	0.02	0.306 average	—
OAK	Carlos Cortes (LF)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	1 / 0	0 / 0	0 / 1	0	0.306 average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.83	positive favors home; negative favors away
Recent form	0.118	positive favors home; negative favors away
Starting-pitcher matchup	0.006	positive favors home; negative favors away
Composite	0.858	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.77-5.03	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.53	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

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.77-5.03	4-5	-0.258	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

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
TEX	Wyatt Langford	batter / LF	better	predicted_RBI 2.67
TEX	Danny Jansen	batter / C	better	predicted_H 2.54
OAK	Max Muncy	batter / DH	worse	predicted_SO 1.855
OAK	Henry Bolte	batter / CF	better	predicted_BB 1.733
TEX	Brandon Nimmo	batter / RF	worse	predicted_AB -1.55
TEX	Cody Freeman	batter / 3B	better	predicted_SO 1.524
TEX	Justin Foscue	batter / 2B	worse	predicted_AB 1.5
OAK	Carlos Cortes	batter / LF	worse	predicted_SO 1.475

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	OAK pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	OAK
info	CURRENT_PITCHER_DUPLICATION	TEX pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	TEX
info	CURRENT_FINALIZED_MATCH_BOX	OAK model-prediction batting totals differ from final simulation box score (AB: current=32, sim=31; H: current=8, sim=4; 2B: current=1, sim=0; HR: current=1, sim=2; R: current=4, sim=3; BB: current=3, sim=4; SO: current=8, sim=11).	OAK
info	CURRENT_PLAYER_MATCH_BOX	OAK current finalized hitter lines are model predictions and differ from the simulation box score: Henry Bolte: AB current=4 box=2, HR current=0 box=1, RBI current=0 box=1, BB current=0 box=2; Tommy White: AB current=4 box=3, H current=1 box=0, R current=1 box=0, BB current=0 box=1; Max Muncy: AB current=3 box=4, H current=1 box=0, SO current=1 box=3; Carlos Cortes: AB current=3 box=4, H current=1 box=0, RBI current=1 box=0, SO current=1 box=2; Zack Gelof: AB current=4 box=3, H current=1 box=0, R current=1 box=0; Lawrence Butler: HR current=1 box=0, BB current=1 box=0, SO current=1 box=2; Jonah Heim: AB current=3 box=4; Jeff McNeil: H current=1 box=2, 2B current=1 box=0, HR current=0 box=1, R current=0 box=1	OAK
info	CURRENT_FINALIZED_MATCH_BOX	TEX model-prediction batting totals differ from final simulation box score (H: current=8, sim=9; 2B: current=2, sim=1; HR: current=1, sim=4; RBI: current=4, sim=5; BB: current=3, sim=2; SO: current=8, sim=10).	TEX
info	CURRENT_PLAYER_MATCH_BOX	TEX current finalized hitter lines are model predictions and differ from the simulation box score: Ezequiel Duran: H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0; Corey Seager: 2B current=0 box=1, BB current=1 box=0; Jake Burger: AB current=4 box=3, H current=1 box=0, HR current=1 box=0, RBI current=1 box=0; Brandon Nimmo: AB current=5 box=3, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Elias Díaz: AB current=3 box=4, SO current=1 box=0; Wyatt Langford: H current=1 box=2, HR current=0 box=2, RBI current=0 box=3, BB current=1 box=0; Cody Freeman: AB current=3 box=4, HR current=0 box=1, RBI current=0 box=1, R current=0 box=1; Justin Foscue: AB current=3 box=4, RBI current=1 box=0, R current=0 box=1, SO current=1 box=0	TEX
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

PHI@ARIHIT

Predicted: PHI 5-4 (conf -0.05, lean)Actual: PHI 2-1Brier 0.224total bias +6

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.05 (lean); this is not the half-scale distance from 50%
team-run absolute error 6; total bias +6
pregame home-minus-away evidence: season -0.301, recent -0.311, pitcher -0.110, composite -0.665
confidence pipeline reached 0.508 before later publication/recency controls; published home probability was 0.473
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 6 (9 vs 3)
PHI offense over-projected by 3
ARI offense over-projected by 3
Aaron Nola (PHI) allowed 3 fewer ER than projected (1 vs 4)

Team	Starter	Projected	Actual	ER err
PHI	Aaron Nola	4 ER / 16 outs	1 ER / 15 outs	-3
ARI	Brandon Pfaadt	1 ER / 19 outs	2 ER / 18 outs	1

Decision internals: run factors home×0.928 / away×1.037; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

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

Top deviations that drove the game:

Kyle Schwarber (PHI DH): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -1.4]
Ryan Waldschmidt (ARI CF): 2-for-4, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 0.94]
Lars Nootbaar (ARI LF): 0-for-2, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.78]
Ketel Marte (ARI DH): 1-for-3, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.6]

Starter miss: Aaron Nola (PHI) posted 1.80 ERA vs projected 6.75 (1 ER in 15 outs); no credible venue, opponent-strength or experience signal explains it — an outlier start (variance).

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

Pitcher versus batter: what actually broke this game

AWAY pitching

Aaron Nola (PHI) threw 15 outs, 1 ER = 1.8 ERA against a projected 6.75 ERA (-4.95 ERA error)

Bullpen behind him: 0 ER over 5 outs.

The HOME lineup he faced produced 1 runs on 9 hits and 0 HR, against a projected 4 runs / 8 hits (-3 runs vs projection).

Who did the damage to him:

Ryan Waldschmidt: 2-for-4, 0 HR, 0 RBI [+0.94 runs vs projection] season xwOBA 0.304, below average - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

HOME pitching

Brandon Pfaadt (ARI) threw 18 outs, 2 ER = 3 ERA against a projected 1.42 ERA (1.58 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The AWAY lineup he faced produced 2 runs on 6 hits and 1 HR, against a projected 4 runs / 8 hits (-2 runs vs projection).

Who did the damage to him:

Garrett Stubbs: 1-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.221, weak - a good day, not a model error

Verdict: offense: the lineup scored well below its projection

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
PHI	Aaron Nola (Starter)	15 / 16	1.8 / 6.75 (-4.95)	1 / 4	5 / 7	0 / 0	2 / 2	4 / 2	venue away: 4.42 vs 5.72 ERA (ok) · opp run factor 0.93 (average) · 27 prior appearances this season
PHI	Jhoan Duran (Closer)	3 / 4	0 / 0	0 / 0	2 / 0	0 / 0	0 / 0	2 / 1
PHI	Orion Kerkering (Middle Reliever (High Leverage))	2 / 4	0 / 0	0 / 0	1 / 1	0 / 0	1 / 0	1 / 0
ARI	Brandon Pfaadt (Starter)	18 / 19	3 / 1.42	2 / 1	5 / 3	1 / 0	1 / 0	1 / 4	venue home: 3.24 vs 2.57 ERA (worse here) · opp run factor 1.04 (average) · 24 prior appearances this season

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
PHI	Kyle Schwarber (DH)	3 / 5	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 1	0 / 2	-1.4	0.365 elite	bad day
ARI	Ryan Waldschmidt (CF)	4 / 3	2 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0.94	0.304 below average	good day
ARI	Lars Nootbaar (LF)	2 / 3	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	0 / 0	-0.78	0.346 elite	bad day
ARI	Ketel Marte (DH)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 0	0 / 1	-0.6	0.349 elite	bad day
PHI	Brandon Marsh (LF)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	-0.47	0.306 average	—
PHI	Garrett Stubbs (C)	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 1	0.47	0.221 weak	—
ARI	Tim Tawa (1B)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	3 / 1	-0.47	0.304 below average	—
PHI	Bryson Stott (3B)	4 / 4	1 / 1	0 / 1	0 / 0	1 / 0	1 / 0	1 / 0	0 / 1	0 / 1	0.29	0.327 above average	—
PHI	Bryce Harper (RF)	4 / 4	2 / 1	1 / 1	0 / 0	0 / 0	1 / 1	0 / 1	0 / 1	0 / 1	0.14	0.393 elite	—
PHI	Alec Bohm (1B)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 0	0 / 0	-0.14	0.323 above average	—
PHI	Trea Turner (SS)	3 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 0	1 / 1	-0.14	0.294 below average	—
ARI	Gabriel Moreno (C)	5 / 5	2 / 1	1 / 1	0 / 0	0 / 0	1 / 1	0 / 1	0 / 1	1 / 1	0.14	0.366 elite	—
PHI	Luis Arraez (2B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0	0.307 average	—
PHI	Justin Crawford (CF)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0	0.281 weak	—
ARI	Nolan Arenado (3B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	1 / 1	0	0.282 weak	—
ARI	Ildemaro Vargas (2B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 0	0	0.299 below average	—
ARI	Corbin Carroll (RF)	5 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	0	0.342 elite	—
ARI	Geraldo Perdomo (SS)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	2 / 0	0	0.327 above average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.301	positive favors home; negative favors away
Recent form	-0.311	positive favors home; negative favors away
Starting-pitcher matchup	-0.11	positive favors home; negative favors away
Composite	-0.665	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.6-4.85	n/a	Base ML historical model prediction
variance_applied	5-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-4	0.527	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	5-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	5-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	5-4	n/a	Capped at max 12 realistic runs
simulation_result	5-4	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.6-4.85	5-5	0.251	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-5	5-4	1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by winner_starter_is_below_average, projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	5-4	5-4	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	5-4	5-4	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	5-4	5-4	-0	Capped at max 12 realistic runs
simulation_result	5-4	5-4	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
ARI	Lars Nootbaar	batter / LF	better	predicted_H 2.25
PHI	Brandon Marsh	batter / LF	worse	predicted_SO 2.039
ARI	Gabriel Moreno	batter / C	worse	predicted_AB -1.6
PHI	Trea Turner	batter / SS	worse	predicted_AB -1.15
PHI	Bryson Stott	batter / 3B	worse	predicted_H 1.1
ARI	Geraldo Perdomo	batter / SS	better	predicted_H -1.06
ARI	Ryan Waldschmidt	batter / CF	deboosted	predicted_SO -1.029
PHI	Justin Crawford	batter / CF	better	predicted_AB 1.025

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	ARI pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	ARI
info	CURRENT_PITCHER_DUPLICATION	PHI pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	PHI
info	CURRENT_FINALIZED_MATCH_BOX	ARI model-prediction batting totals differ from final simulation box score (2B: current=2, sim=3; 3B: current=0, sim=1; HR: current=1, sim=0; SO: current=6, sim=3; SF: current=0, sim=1).	ARI
info	CURRENT_PLAYER_MATCH_BOX	ARI current finalized hitter lines are model predictions and differ from the simulation box score: Ketel Marte: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Lars Nootbaar: H current=1 box=3, 2B current=1 box=2, R current=0 box=1; Gabriel Moreno: AB current=5 box=3, 2B current=1 box=0, RBI current=1 box=2, BB current=1 box=0; Nolan Arenado: H current=1 box=0, R current=1 box=0, SO current=1 box=0; Ildemaro Vargas: RBI current=1 box=0; Corbin Carroll: AB current=4 box=5, 3B current=0 box=1, R current=1 box=0; Geraldo Perdomo: H current=1 box=0, RBI current=0 box=1, R current=0 box=1; Tim Tawa: AB current=3 box=4	ARI
info	CURRENT_FINALIZED_MATCH_BOX	PHI model-prediction batting totals differ from final simulation box score (AB: current=36, sim=35; 2B: current=2, sim=3; RBI: current=4, sim=5; R: current=4, sim=5; BB: current=3, sim=0; SO: current=8, sim=7).	PHI
info	CURRENT_PLAYER_MATCH_BOX	PHI current finalized hitter lines are model predictions and differ from the simulation box score: Bryce Harper: 2B current=1 box=0, HR current=0 box=1, BB current=1 box=0; Alec Bohm: 2B current=0 box=1, R current=0 box=1; Bryson Stott: H current=1 box=2, 2B current=1 box=0, RBI current=0 box=1, R current=0 box=1; Kyle Schwarber: AB current=5 box=4, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Luis Arraez: SO current=1 box=0; Brandon Marsh: H current=1 box=0, R current=1 box=0, SO current=1 box=3; Trea Turner: AB current=5 box=4, 2B current=0 box=1; Garrett Stubbs: SO current=1 box=0	PHI
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

SD@CINHIT

Predicted: SD 6-4 (conf -0.35, strong)Actual: SD 5-0Brier 0.107total bias +5

Root cause: Right winner, wrong score

Evidence-backed diagnosis

published system edge 0.35 (strong); this is not the half-scale distance from 50%
team-run absolute error 5; total bias +5
pregame home-minus-away evidence: season -0.695, recent -0.361, pitcher -0.005, composite -0.844
confidence pipeline reached 0.677 before later publication/recency controls; published home probability was 0.326
score-volatility actions: supported_blowout_margin_expansion
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

total runs over-projected by 5 (10 vs 5)
CIN offense over-projected by 4
Michael King (SD) allowed 3 fewer ER than projected (0 vs 3)

Team	Starter	Projected	Actual	ER err
SD	Michael King	3 ER / 18 outs	0 ER / 18 outs	-3
CIN	Brady Singer	4 ER / 15 outs	4 ER / 13 outs	0

Decision internals: run factors home×1.002 / away×1.007; 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:

Manny Machado (SD 3B): 4-for-5, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 1.39]
Ty France (SD 1B): 3-for-5, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.94]
Jake Cronenworth (SD 2B): 0-for-5, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Eugenio Suárez (CIN 3B): 0-for-3, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact -0.8]
Sal Stewart (CIN 1B): 1-for-4, 0 HR, 0 RBI (projected 1.0 H, 1.0 HR) [run impact -0.62]

Starter miss: Michael King (SD) posted 0.00 ERA vs projected 4.50 (0 ER in 18 outs); he is materially worse pitching away (3.56 ERA there vs 2.94 on the other side), a venue-role weakness the model under-weighted.

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

Pitcher versus batter: what actually broke this game

AWAY pitching

Michael King (SD) threw 18 outs, 0 ER = 0 ERA against a projected 4.5 ERA (-4.5 ERA error)

Bullpen behind him: 0 ER over 0 outs.

The HOME lineup he faced produced 0 runs on 4 hits and 0 HR, against a projected 5 runs / 8 hits (-5 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

Brady Singer (CIN) threw 13 outs, 4 ER = 8.31 ERA against a projected 7.2 ERA (1.11 ERA error)

Bullpen behind him: 0 ER over 3 outs.

The AWAY lineup he faced produced 5 runs on 15 hits and 0 HR, against a projected 4 runs / 7 hits (1 runs vs projection).

Who did the damage to him:

Manny Machado: 4-for-5, 0 HR, 0 RBI [+1.39 runs vs projection] season xwOBA 0.333, above average - real projection error
Ty France: 3-for-5, 0 HR, 1 RBI [+0.94 runs vs projection] season xwOBA 0.331, above average - real projection error
Luis Campusano: 2-for-4, 0 HR, 0 RBI [+0.61 runs vs projection] season xwOBA 0.335, 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)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
SD	Michael King (Starter)	18 / 18	0 / 4.5 (-4.50)	0 / 3	4 / 5	0 / 0	3 / 0	6 / 6	venue away: 3.56 vs 2.94 ERA (worse here) · opp run factor 1 (weak) · 27 prior appearances this season
CIN	Brady Singer (Starter)	13 / 15	8.31 / 7.2	4 / 4	10 / 8	0 / 1	0 / 1	5 / 6	venue home: 3.82 vs 5.67 ERA (ok) · opp run factor 1.01 (average) · 26 prior appearances this season
CIN	Tejay Antone (Setup Man)	3 / 3	0 / 18 (-18.00)	0 / 2	1 / 3	0 / 0	0 / 0	2 / 1

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
SD	Manny Machado (3B)	5 / 4	4 / 1	1 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0 / 1	1.39	0.333 above average	projection too low
SD	Ty France (1B)	5 / 3	3 / 1	0 / 0	0 / 0	0 / 0	1 / 1	1 / 0	0 / 0	0 / 1	0.94	0.331 above average	projection too low
SD	Jake Cronenworth (2B)	5 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	2 / 1	-0.8	0.272 weak	projection too high
CIN	Eugenio Suárez (3B)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	2 / 1	-0.8	0.29 below average	projection too high
CIN	Sal Stewart (1B)	4 / 5	1 / 1	1 / 0	0 / 0	0 / 1	0 / 1	0 / 1	0 / 0	0 / 1	-0.62	0.351 elite	bad day
SD	Luis Campusano (C)	4 / 3	2 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	1 / 1	0.61	0.335 above average	projection too low
CIN	Tyler Stephenson (C)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 1	0 / 1	0 / 1	1 / 0	2 / 1	-0.6	0.335 above average	bad day
SD	Dustin Harris (LF)	5 / 3	2 / 1	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	0 / 0	1 / 0	0.47	0.327 above average	—
CIN	Edwin Arroyo (2B)	4 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	4 / 1	-0.47	0.291 below average	—
CIN	JJ Bleday (LF)	3 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0 / 1	-0.47	0.349 elite	—
SD	Jackson Merrill (CF)	4 / 5	2 / 1	0 / 0	0 / 0	0 / 1	1 / 1	1 / 1	0 / 0	2 / 1	-0.46	0.336 above average	—
CIN	Elly De La Cruz (DH)	2 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 1	0 / 1	2 / 1	1 / 1	-0.45	0.364 elite	—
SD	Xander Bogaerts (DH)	5 / 3	1 / 1	1 / 0	0 / 0	0 / 0	2 / 0	1 / 0	0 / 0	3 / 0	0.31	0.321 above average	—
SD	Fernando Tatis Jr. (RF)	5 / 5	1 / 1	1 / 1	0 / 0	0 / 0	0 / 1	1 / 1	0 / 0	0 / 1	0	0.374 elite	—
SD	Sung-Mun Song (SS)	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.256 weak	—
CIN	Héctor Rodríguez (RF)	5 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	2 / 1	0	0.275 weak	—
CIN	Matt McLain (SS)	3 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	0	0.307 average	—
CIN	TJ Friedl (CF)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0	0.248 weak	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	-0.695	positive favors home; negative favors away
Recent form	-0.361	positive favors home; negative favors away
Starting-pitcher matchup	-0.005	positive favors home; negative favors away
Composite	-0.844	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	4.79-3.34	n/a	Base ML historical model prediction
variance_applied	5-3	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	6-4	0.674	Enhanced ML with recent form, H2H, low-score analysis
embedded_savant_run_environment	6-4	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-4	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-4	n/a	Capped at max 12 realistic runs
simulation_result	6-4	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	4.79-3.34	5-3	0.547	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	5-3	6-4	-0	Enhanced ML with recent form, H2H, low-score analysis
embedded_savant_run_environment	6-4	6-4	-0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	6-4	6-4	-0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	6-4	6-4	-0	Capped at max 12 realistic runs
simulation_result	6-4	6-4	-0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
SD	Sung-Mun Song	batter / SS	worse	predicted_SO 2.457
SD	Dustin Harris	batter / LF	worse	predicted_H 2.25
CIN	Sal Stewart	batter / 1B	worse	predicted_H 1.67
CIN	Héctor Rodríguez	batter / RF	better	predicted_RBI 1.526
SD	Jackson Merrill	batter / CF	better	predicted_AB -1.49
SD	Ty France	batter / 1B	better	predicted_RBI 1.39
SD	Jake Cronenworth	batter / 2B	worse	predicted_SO 1.083
SD	Fernando Tatis Jr.	batter / RF	better	predicted_AB -1.07

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	CIN pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	CIN
info	CURRENT_PITCHER_DUPLICATION	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	CIN model-prediction batting totals differ from final simulation box score (AB: current=34, sim=35; 2B: current=1, sim=0; HR: current=2, sim=1; R: current=5, sim=4; BB: current=3, sim=1; SO: current=9, sim=7).	CIN
info	CURRENT_PLAYER_MATCH_BOX	CIN current finalized hitter lines are model predictions and differ from the simulation box score: Héctor Rodríguez: RBI current=0 box=2, BB current=0 box=1, SO current=1 box=0; Elly De La Cruz: 2B current=1 box=0, RBI current=1 box=0, BB current=1 box=0; Eugenio Suárez: H current=1 box=0, RBI current=1 box=0, BB current=1 box=0; Matt McLain: H current=1 box=0, R current=1 box=0; Tyler Stephenson: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Edwin Arroyo: AB current=3 box=4; TJ Friedl: AB current=3 box=4, H current=0 box=1, HR current=0 box=1, RBI current=0 box=1; JJ Bleday: BB current=1 box=0	CIN
info	CURRENT_FINALIZED_MATCH_BOX	SD model-prediction batting totals differ from final simulation box score (AB: current=34, sim=37; H: current=7, sim=11; 2B: current=1, sim=3; RBI: current=4, sim=6; R: current=4, sim=6; BB: current=3, sim=2; SO: current=7, sim=9; SF: current=0, sim=1).	SD
info	CURRENT_PLAYER_MATCH_BOX	SD current finalized hitter lines are model predictions and differ from the simulation box score: Fernando Tatis Jr.: AB current=5 box=4, H current=1 box=2, RBI current=1 box=0, BB current=0 box=1; Manny Machado: AB current=4 box=5, 2B current=0 box=1, BB current=1 box=0; Xander Bogaerts: AB current=3 box=4, 2B current=0 box=1, SO current=0 box=1; Ty France: AB current=3 box=4, RBI current=1 box=2, R current=0 box=1, SO current=1 box=0; Jackson Merrill: AB current=5 box=4, H current=1 box=2, RBI current=1 box=2; Sung-Mun Song: AB current=3 box=4, SO current=1 box=3; Jake Cronenworth: RBI current=1 box=0, R current=0 box=1, BB current=1 box=0, SO current=1 box=2; Dustin Harris: AB current=3 box=4, H current=1 box=3, RBI current=0 box=2	SD
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

SEA@BOSHIT

Predicted: BOS 2-3 (conf +0.12, moderate)Actual: BOS 8-9Brier 0.193total bias -12

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 12; total bias -12
pregame home-minus-away evidence: season +0.957, recent +0.533, pitcher +0.016, composite +0.874
confidence pipeline reached 0.582 before later publication/recency controls; published home probability was 0.560
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 12 (5 vs 17)
SEA offense under-projected by 6
BOS offense under-projected by 6
George Kirby (SEA) allowed 3 more ER than projected (5 vs 2)
Payton Tolle (BOS) allowed 4 more ER than projected (4 vs 0)

Team	Starter	Projected	Actual	ER err
SEA	George Kirby	2 ER / 17 outs	5 ER / 15 outs	3
BOS	Payton Tolle	0 ER / 18 outs	4 ER / 17 outs	4

Decision internals: run factors home×0.975 / away×0.959; savant lineup adjustment applied

Player-level breakdown (projected vs actual)

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

Top deviations that drove the game:

Randy Arozarena (SEA LF): 4-for-5, 1 HR, 4 RBI (projected 1.0 H, 0.0 HR) [run impact 2.01]
Dominic Canzone (SEA DH): 2-for-5, 0 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Roman Anthony (BOS DH): 2-for-4, 0 HR, 0 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Jarren Duran (BOS LF): 1-for-4, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 1.26]
Mickey Gasper (BOS 1B): 1-for-4, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]

Starter miss: Payton Tolle (BOS) posted 6.35 ERA vs projected 0.00 (4 ER in 17 outs); he is materially worse pitching home (3.64 ERA there vs 2.59 on the other side), a venue-role weakness the model under-weighted.

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

Pitcher versus batter: what actually broke this game

AWAY pitching

George Kirby (SEA) threw 15 outs, 5 ER = 9 ERA against a projected 3.18 ERA (5.82 ERA error)

Bullpen behind him: 1 ER over 4 outs.

The HOME lineup he faced produced 8 runs on 12 hits and 2 HR, against a projected 4 runs / 9 hits (4 runs vs projection).

Who did the damage to him:

Roman Anthony: 2-for-4, 0 HR, 0 RBI [+1.4 runs vs projection] season xwOBA 0.361, elite - real projection error
Jarren Duran: 1-for-4, 1 HR, 2 RBI [+1.26 runs vs projection] season xwOBA 0.296, below average - a good day, not a model error
Mickey Gasper: 1-for-4, 1 HR, 1 RBI [+0.93 runs vs projection] season xwOBA 0.372, elite - real projection error

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

HOME pitching

Payton Tolle (BOS) threw 17 outs, 4 ER = 6.35 ERA against a projected 0 ERA (6.35 ERA error)

Bullpen behind him: 4 ER over 7 outs.

The AWAY lineup he faced produced 8 runs on 11 hits and 2 HR, against a projected 5 runs / 7 hits (3 runs vs projection).

Who did the damage to him:

Randy Arozarena: 4-for-5, 1 HR, 4 RBI [+2.01 runs vs projection] season xwOBA 0.349, elite - real projection error
Dominic Canzone: 2-for-5, 0 HR, 2 RBI [+1.4 runs vs projection] season xwOBA 0.356, elite - real projection error
Weston Wilson: 1-for-4, 0 HR, 0 RBI [+0.8 runs vs projection] season xwOBA 0.26, weak - a good day, not a model error

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

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
SEA	George Kirby (Starter)	15 / 17	9 / 3.18 (+5.82)	5 / 2	7 / 5	2 / 1	2 / 0	2 / 5	venue away: 5.19 vs 3.21 ERA (worse here) · opp run factor 0.98 (average) · 25 prior appearances this season
SEA	Andrés Muñoz (Closer)	4 / 3	6.75 / 9	1 / 1	1 / 1	0 / 0	0 / 0	3 / 0
BOS	Payton Tolle (Starter)	17 / 18	6.35 / 0 (+6.35)	4 / 0	6 / 2	1 / 0	2 / 3	6 / 7	venue home: 3.64 vs 2.59 ERA (worse here) · opp run factor 0.96 (weak) · 22 prior appearances this season
BOS	Aroldis Chapman (Closer)	3 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	2 / 0
BOS	Erik Miller (Setup Man)	3 / 3	9 / 0 (+9.00)	1 / 0	1 / 1	1 / 0	0 / 1	2 / 0
BOS	Tyron Guerrero (Middle Reliever (High Leverage))	1 / 3	81 / 18 (+63.00)	3 / 2	3 / 2	0 / 1	1 / 0	0 / 1

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
SEA	Randy Arozarena (LF)	5 / 4	4 / 1	1 / 1	0 / 0	1 / 0	4 / 0	3 / 1	0 / 1	0 / 1	2.01	0.349 elite	projection too low
SEA	Dominic Canzone (DH)	5 / 4	2 / 1	1 / 0	1 / 0	0 / 0	2 / 1	0 / 0	0 / 0	1 / 1	1.4	0.356 elite	projection too low
BOS	Roman Anthony (DH)	4 / 4	2 / 1	1 / 0	1 / 0	0 / 0	0 / 0	2 / 0	1 / 1	0 / 1	1.4	0.361 elite	projection too low
BOS	Jarren Duran (LF)	4 / 4	1 / 1	0 / 0	0 / 0	1 / 0	2 / 1	1 / 0	1 / 0	0 / 1	1.26	0.296 below average	good day
BOS	Mickey Gasper (1B)	4 / 3	1 / 1	0 / 0	0 / 0	1 / 0	1 / 1	1 / 0	1 / 1	1 / 0	0.93	0.372 elite	projection too low
SEA	Weston Wilson (3B)	4 / 3	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	3 / 1	0.8	0.26 weak	good day
BOS	Wilyer Abreu (RF)	5 / 4	0 / 1	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	2 / 1	-0.78	0.347 elite	bad day
BOS	Caleb Durbin (3B)	2 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	2 / 0	0 / 1	0.66	0.288 below average	good day
SEA	Cole Young (2B)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	2 / 1	0 / 0	1 / 1	-0.47	0.316 average	—
SEA	J.P. Crawford (SS)	2 / 4	1 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	1 / 1	0.47	0.327 above average	—
BOS	Ceddanne Rafaela (CF)	4 / 6	2 / 1	0 / 0	0 / 0	0 / 0	3 / 1	0 / 0	0 / 0	1 / 1	0.47	0.282 weak	—
SEA	Cal Raleigh (C)	5 / 4	1 / 1	0 / 0	0 / 0	1 / 1	1 / 1	1 / 1	0 / 1	1 / 1	-0.33	0.315 average	—
SEA	Taylor Ward (RF)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	-0.33	0.333 above average	—
SEA	Julio Rodríguez (CF)	5 / 5	1 / 1	1 / 0	0 / 0	0 / 0	1 / 1	0 / 1	0 / 0	2 / 1	0.31	0.336 above average	—
BOS	Trevor Story (SS)	5 / 4	2 / 1	1 / 0	0 / 0	0 / 1	3 / 0	0 / 1	0 / 0	1 / 1	-0.15	0.251 weak	—
SEA	Josh Naylor (1B)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 1	-0.14	0.313 average	—
BOS	Nick Sogard (2B)	4 / 4	2 / 1	1 / 1	0 / 0	0 / 0	0 / 1	2 / 1	0 / 1	0 / 1	0.14	0.327 above average	—
BOS	Connor Wong (C)	4 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	1 / 1	0	0.286 below average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	0.957	positive favors home; negative favors away
Recent form	0.533	positive favors home; negative favors away
Starting-pitcher matchup	0.016	positive favors home; negative favors away
Composite	0.874	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.66-5.52	n/a	Base ML historical model prediction
variance_applied	4-6	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	2-3	0.56	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	2-3	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	2-3	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	2-3	n/a	Capped at max 12 realistic runs
simulation_result	2-3	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.66-5.52	4-6	0.144	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-6	2-3	-1	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	2-3	2-3	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	2-3	2-3	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	2-3	2-3	0	Capped at max 12 realistic runs
simulation_result	2-3	2-3	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
BOS	Ceddanne Rafaela	batter / CF	deboosted	predicted_AB -2.73
BOS	Trevor Story	batter / SS	worse	predicted_AB -2.16
SEA	Julio Rodríguez	batter / CF	worse	predicted_AB -2.01
BOS	Payton Tolle	pitcher / Starter	better	predicted_SO 2
SEA	J.P. Crawford	batter / SS	worse	predicted_AB -1.65
SEA	Cole Young	batter / 2B	worse	predicted_RBI 1.61
BOS	Wilyer Abreu	batter / RF	worse	predicted_RBI 1.58
SEA	Weston Wilson	batter / 3B	worse	predicted_AB 1.5

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	BOS pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	BOS
info	CURRENT_PITCHER_DUPLICATION	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	BOS model-prediction batting totals differ from final simulation box score (AB: current=36, sim=29; H: current=9, sim=6; 2B: current=2, sim=1; 3B: current=0, sim=1; RBI: current=4, sim=3; R: current=4, sim=3; BB: current=3, sim=1; SO: current=8, sim=5).	BOS
info	CURRENT_PLAYER_MATCH_BOX	BOS current finalized hitter lines are model predictions and differ from the simulation box score: Roman Anthony: H current=1 box=0, BB current=1 box=0, SO current=1 box=0; Mickey Gasper: AB current=3 box=4, 2B current=0 box=1, RBI current=1 box=0, BB current=1 box=0; Nick Sogard: AB current=4 box=3, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Caleb Durbin: AB current=4 box=3, H current=1 box=0, R current=1 box=0, SO current=1 box=0; Ceddanne Rafaela: AB current=6 box=3, H current=1 box=2, RBI current=1 box=0, R current=0 box=1; Wilyer Abreu: 2B current=1 box=0, HR current=0 box=1, RBI current=0 box=2, SO current=1 box=0; Connor Wong: SO current=1 box=0; Jarren Duran: AB current=4 box=3, 3B current=0 box=1, RBI current=1 box=0, SO current=1 box=2	BOS
info	CURRENT_FINALIZED_MATCH_BOX	SEA model-prediction batting totals differ from final simulation box score (AB: current=36, sim=32; H: current=7, sim=5; 2B: current=1, sim=2; RBI: current=3, sim=2; R: current=5, sim=2; SO: current=9, sim=8).	SEA
info	CURRENT_PLAYER_MATCH_BOX	SEA current finalized hitter lines are model predictions and differ from the simulation box score: Julio Rodríguez: AB current=5 box=3, H current=1 box=0, RBI current=1 box=0, R current=1 box=0; Cal Raleigh: H current=1 box=0, HR current=1 box=0, RBI current=1 box=0, R current=1 box=0; Taylor Ward: H current=1 box=0, R current=1 box=0, BB current=1 box=0, SO current=1 box=2; Randy Arozarena: 2B current=1 box=0, R current=1 box=0, BB current=1 box=0, SO current=1 box=0; Dominic Canzone: AB current=4 box=3, 2B current=0 box=1, RBI current=1 box=0, R current=0 box=1; Josh Naylor: 2B current=0 box=1; Cole Young: HR current=0 box=1, RBI current=0 box=2; Weston Wilson: AB current=3 box=4, H current=0 box=1	SEA
info	RECONCILIATION_PROPAGATION	No score_target_reconciliation step was used for this game.

SF@ATLMISS

Predicted: ATL 4-5 (conf +0.06, lean)Actual: SF 7-3Brier 0.281total bias -1

Root cause: Starting-pitcher projection was wrong

Evidence-backed diagnosis

published system edge 0.06 (lean); this is not the half-scale distance from 50%
team-run absolute error 5; total bias -1
pregame home-minus-away evidence: season +1.228, recent +0.182, pitcher -0.083, composite +0.718
confidence pipeline reached 0.508 before later publication/recency controls; published home probability was 0.530
score-volatility actions: no_change
forecast-file integrity: clean (0 defects, 7 informational note(s)); the saved artifact is not why the pick missed

Quantified contributing factors

SF offense under-projected by 3
Bryce Elder (ATL) allowed 4 more ER than projected (5 vs 1)

Team	Starter	Projected	Actual	ER err
SF	Anthony Molina	4 ER / 9 outs	3 ER / 15 outs	-1
ATL	Bryce Elder	1 ER / 18 outs	5 ER / 16 outs	4

Decision internals: run factors home×1 / away×1.083; 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:

Jonah Cox (SF CF): 2-for-5, 1 HR, 2 RBI (projected 1.0 H, 0.0 HR) [run impact 1.4]
Andrew Knizner (SF C): 2-for-4, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 1.25]
Matt Olson (ATL 1B): 0-for-4, 0 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact -1.11]
Osleivis Basabe (SF 3B): 1-for-4, 1 HR, 1 RBI (projected 1.0 H, 0.0 HR) [run impact 0.93]
Nate Furman (SF 2B): 2-for-3, 0 HR, 0 RBI (projected 0.0 H, 0.0 HR) [run impact 0.61]

Starter miss: Bryce Elder (ATL) posted 8.44 ERA vs projected 1.50 (5 ER in 16 outs); he faced a strong offense (opponent run factor 1.08, 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

Anthony Molina (SF) threw 15 outs, 3 ER = 5.4 ERA against a projected 12 ERA (-6.6 ERA error) only 4 prior appearances in the covered season - the projection had little of his own record to stand on

Bullpen behind him: 0 ER over 7 outs.

The HOME lineup he faced produced 3 runs on 9 hits and 1 HR, against a projected 5 runs / 9 hits (-2 runs vs projection).

Who did the damage to him:

Austin Riley: 2-for-4, 0 HR, 0 RBI [+0.47 runs vs projection] season xwOBA 0.298, below average - a good day, not a model error
Drake Baldwin: 2-for-4, 0 HR, 1 RBI [+0.45 runs vs projection] season xwOBA 0.367, elite - a good day, not a model error
Sean Murphy: 1-for-3, 0 HR, 0 RBI [+0.33 runs vs projection] season xwOBA 0.275, weak - a good day, not a model error

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

HOME pitching

Bryce Elder (ATL) threw 16 outs, 5 ER = 8.44 ERA against a projected 1.5 ERA (6.94 ERA error)

Bullpen behind him: 1 ER over 1 outs.

The AWAY lineup he faced produced 6 runs on 10 hits and 3 HR, against a projected 3 runs / 7 hits (3 runs vs projection).

Who did the damage to him:

Jonah Cox: 2-for-5, 1 HR, 2 RBI [+1.4 runs vs projection] season xwOBA 0.274, weak - a good day, not a model error
Andrew Knizner: 2-for-4, 0 HR, 0 RBI [+1.25 runs vs projection] season xwOBA 0.271, weak - a good day, not a model error
Osleivis Basabe: 1-for-4, 1 HR, 1 RBI [+0.93 runs vs projection] season xwOBA 0.283, weak - a good day, not a model error

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

Starting pitchers & relievers: actual / projected + context (ERA is workload-aware: ER over 9 innings, so a short blow-up is not read like a quality start; ERA error in parentheses when material)

Team	Pitcher	Outs	ERA	ER	H	HR	BB	SO	Context (as-of)
SF	Anthony Molina (Starter)	15 / 9	5.4 / 12 (-6.60)	3 / 4	6 / 4	1 / 2	0 / 2	6 / 2	opp run factor 1 (strong) · thin sample: 4 prior appearances this season (high uncertainty)
SF	Dylan Smith (Closer)	3 / 3	0 / 0	0 / 0	1 / 0	0 / 0	1 / 0	1 / 2
SF	Jason Foley (Setup Man)	4 / 3	0 / 0	0 / 0	2 / 0	0 / 0	0 / 0	1 / 2
ATL	Bryce Elder (Starter)	16 / 18	8.44 / 1.5 (+6.94)	5 / 1	8 / 1	2 / 0	3 / 4	3 / 6	venue home: 3.49 vs 4.93 ERA (ok) · opp run factor 1.08 (weak) · 25 prior appearances this season
ATL	Dylan Lee (Setup Man)	1 / 3	27 / 0 (+27.00)	1 / 0	2 / 0	0 / 0	0 / 0	0 / 2

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

Team	Batter	AB	H	2B	3B	HR	RBI	R	BB	SO	Run impact	Season xwOBA	Signal
SF	Jonah Cox (CF)	5 / 3	2 / 1	0 / 0	0 / 0	1 / 0	2 / 0	1 / 0	0 / 0	1 / 1	1.4	0.274 weak	good day
SF	Andrew Knizner (C)	4 / 3	2 / 0	1 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 0	0 / 1	1.25	0.271 weak	good day
ATL	Matt Olson (1B)	4 / 4	0 / 1	0 / 1	0 / 0	0 / 0	1 / 1	0 / 1	0 / 1	1 / 1	-1.11	0.355 elite	bad day
SF	Osleivis Basabe (3B)	4 / 3	1 / 1	0 / 0	0 / 0	1 / 0	1 / 0	1 / 0	0 / 0	1 / 1	0.93	0.283 weak	good day
SF	Nate Furman (2B)	3 / 3	2 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	0 / 1	1 / 1	0.61	0.213 weak	good day
SF	Drew Gilbert (LF)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.328 above average	—
SF	Christian Koss (SS)	3 / 3	0 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	-0.47	0.24 weak	—
ATL	Mauricio Dubón (SS)	4 / 4	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 0	1 / 0	-0.47	0.302 below average	—
ATL	Austin Riley (3B)	4 / 4	2 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0.47	0.298 below average	—
ATL	Drake Baldwin (DH)	4 / 4	2 / 1	1 / 0	0 / 0	0 / 0	1 / 1	1 / 1	0 / 1	2 / 1	0.45	0.367 elite	—
SF	Rafael Devers (DH)	3 / 4	1 / 1	0 / 0	0 / 0	1 / 1	1 / 1	2 / 1	2 / 1	1 / 1	0.33	0.34 elite	—
ATL	Ronald Acuña Jr. (RF)	4 / 5	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 1	-0.33	0.362 elite	—
ATL	Sean Murphy (C)	3 / 3	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	1 / 0	1 / 0	2 / 1	0.33	0.275 weak	—
ATL	Michael Harris II (CF)	4 / 5	2 / 2	0 / 1	0 / 0	1 / 1	1 / 1	1 / 1	0 / 0	2 / 1	-0.31	0.341 elite	—
SF	Bryce Eldridge (1B)	5 / 4	2 / 1	0 / 1	0 / 0	0 / 0	1 / 1	0 / 1	0 / 1	0 / 1	-0.17	0.357 elite	—
SF	Jung Hoo Lee (RF)	4 / 5	0 / 1	0 / 0	0 / 0	0 / 0	0 / 1	0 / 1	1 / 0	1 / 1	-0.14	0.315 average	—
ATL	Ozzie Albies (2B)	4 / 4	1 / 1	0 / 0	0 / 0	0 / 0	0 / 0	0 / 1	0 / 0	0 / 1	0	0.284 below average	—
ATL	Mike Yastrzemski (LF)	4 / 3	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	0 / 0	1 / 1	0	0.308 average	—

Full model trace, player evidence, and audit findings

Fundamental evidence entering the pick

Signal	Home-away edge	Meaning
Season quality	1.228	positive favors home; negative favors away
Recent form	0.182	positive favors home; negative favors away
Starting-pitcher matchup	-0.083	positive favors home; negative favors away
Composite	0.718	positive favors home; negative favors away

Score and confidence evolution

Stage	Away-home runs	Confidence	What changed
historical_ml	3.85-4.61	n/a	Base ML historical model prediction
variance_applied	4-5	n/a	Deterministic baseline target (variance factor 1.0)
enhanced_prediction	4-5	0.53	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	n/a	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	n/a	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	n/a	Capped at max 12 realistic runs
simulation_result	4-5	n/a	Final inning-by-inning simulation result

Official postgame stage attribution

Stage	Before	After	Toward final pick	What changed
variance_applied	3.85-4.61	4-5	0.24	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 projected_margin_is_one_run, matchup_recency_guardrail)
embedded_savant_run_environment	4-5	4-5	0	Target score adjusted by temporal Savant lineup quality versus opposing starter quality
fundamental_edge_target_review	4-5	4-5	0	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
capped_target	4-5	4-5	0	Capped at max 12 realistic runs
simulation_result	4-5	4-5	0	Final inning-by-inning simulation result

Largest player-level adjustments from the ultra-detailed trace

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

Team	Player	Role	Net effect	Largest change
SF	Osleivis Basabe	batter / 3B	worse	predicted_RBI 2.692
ATL	Drake Baldwin	batter / DH	better	predicted_RBI 2.407
ATL	Michael Harris II	batter / CF	worse	predicted_AB -1.72
ATL	Mauricio Dubón	batter / SS	better	predicted_AB -1.705
SF	Jung Hoo Lee	batter / RF	worse	predicted_SO 1.615
SF	Andrew Knizner	batter / C	worse	predicted_AB 1.5
ATL	Ozzie Albies	batter / 2B	worse	predicted_SO 1.382
SF	Anthony Molina	pitcher / Starter	worse	predicted_IP -1.333

Structural audit findings

Severity	Rule	Finding	Subject
info	CURRENT_PITCHER_DUPLICATION	ATL pitcher list appears in both predicted_performance.pitchers and player_predictions_by_position.pitchers; IDs match and were de-duplicated in this audit.	ATL
info	CURRENT_PITCHER_DUPLICATION	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	ATL model-prediction batting totals differ from final simulation box score (AB: current=36, sim=30; H: current=9, sim=7; 2B: current=2, sim=1; HR: current=1, sim=3; RBI: current=4, sim=5; BB: current=3, sim=4; SO: current=8, sim=6; SF: current=0, sim=1).	ATL
info	CURRENT_PLAYER_MATCH_BOX	ATL current finalized hitter lines are model predictions and differ from the simulation box score: Drake Baldwin: H current=1 box=2, HR current=0 box=2, RBI current=1 box=3, BB current=1 box=0; Mauricio Dubón: AB current=4 box=2, H current=1 box=0, R current=0 box=1, BB current=0 box=1; Ronald Acuña Jr.: AB current=5 box=4, H current=1 box=2, 2B current=0 box=1, BB current=1 box=0; Ozzie Albies: H current=1 box=0, R current=1 box=0, SO current=1 box=2; Michael Harris II: AB current=5 box=3, H current=2 box=0, 2B current=1 box=0, HR current=1 box=0; Matt Olson: 2B current=1 box=0, RBI current=1 box=0, R current=1 box=0, BB current=1 box=0; Sean Murphy: R current=0 box=1, BB current=0 box=1; Austin Riley: H current=1 box=0, SO current=1 box=0	ATL
info	CURRENT_FINALIZED_MATCH_BOX	SF model-prediction batting totals differ from final simulation box score (AB: current=31, sim=30; H: current=7, sim=3; 2B: current=1, sim=0; HR: current=1, sim=2; RBI: current=3, sim=4; R: current=3, sim=4; BB: current=3, sim=5; SO: current=9, sim=8).	SF
info	CURRENT_PLAYER_MATCH_BOX	SF current finalized hitter lines are model predictions and differ from the simulation box score: Rafael Devers: AB current=4 box=3, HR current=1 box=0; Drew Gilbert: AB current=3 box=4, H current=1 box=0, SO current=1 box=0; Osleivis Basabe: AB current=3 box=4, H current=1 box=2, HR current=0 box=2, RBI current=0 box=3; Bryce Eldridge: AB current=4 box=3, H current=1 box=0, 2B current=1 box=0, RBI current=1 box=0; Jonah Cox: H current=1 box=0, R current=0 box=1, BB current=0 box=1; Christian Koss: H current=1 box=0, SO current=1 box=0; Andrew Knizner: AB current=3 box=4; Jung Hoo Lee: AB current=5 box=3, H current=1 box=0, RBI current=1 box=0, R current=1 box=0	SF
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; 12/12 games; 7/7 causes; 6/6 sources

1. Evidence receipt and gaps

Source manifest

All seven manifest entries were checked against their recorded status:

Source	Path	Bytes	Status	Use
Prediction review	`/app/output/reports/prediction_review_20260831.json`	300,127	Loaded	Official outcomes and scoring errors
Forensic deep audit	`/app/output/reports/PREDICTION_SYSTEM_DEEP_AUDIT_20260831.json`	195,470	Loaded	Probability, provenance, starter, and decision contracts
Continuous learning	`/app/output/learning/continuous_learning_daily_20260831.json`	1,934,747	Loaded	Stage quality, calibration, lineage, margins, starter comparisons
Detailed report	`/app/output/reports/detailed_20260831.html`	284,841	Loaded	Game and simulation rendering
Ultra-detailed player audit	`/app/output/reports/20260831_ultra_detailed_player_audit.html`	4,202,742	Loaded	Player-level prediction trace
Structural audit	`/app/output/reports/audit_report_20260831.html`	257,116	Loaded	Invariant findings
Integrity revision audit	`/app/output/reports/audit_report_20260831_integrity_revision.html`	Not supplied	Missing	Post-remediation proof for immutable revision artifacts

Coverage

All 12 expected games were reviewed:

Canonical artifacts loaded: 12
Continuous-learning games: 12
Forensic deep-audit games: 12
Games with starter comparisons: 12
Structural audits completed: 12
Games structurally audited: 12/12
Settled games: 12
Winner hits/misses: 8/4

The data marks learning eligibility as true because all game artifacts passed critical/error structural gates.

Gaps and contradictions

Missing integrity revision audit. The manifest explicitly reports status: "missing" for the immutable-revision audit. This is an evidence gap and an artifact-integrity limitation, although the structural summary reports zero critical and zero error findings.
Per-game audit contradiction. Each game has structural_audit.findings: [] and status: "PASS", but the diagnostic observations state that the structural audit exposed 7 or 8 non-pass findings. The summaries count only info findings—87 across the slate—but the underlying individual findings are not present in this JSON. Their content therefore cannot be independently enumerated or evaluated.
BAL@COL probability-path naming inconsistency. The game-level trace records raw_pick_probability: 0.724282, while the probability-lineage row records raw_model_pick_probability: 0.7585 and governor_input_probability: 0.724282. The lineage is materially more complete; the game-level raw_pick_probability field appears to refer to a post-raw/pre-governor value rather than the true raw model output. This is a trace-labeling defect or schema ambiguity, not an unexplained publication bypass.
Confidence-edge convention. The listed confidence_home_minus_away is a signed home-minus-away edge, not a probability itself. For example, BAL@COL has -0.4985, corresponding to a published away probability of approximately 0.749 and home probability of approximately 0.251. The reports correctly warn that the “0.50” strong edge is not a half-scale distance from 50%.
No source data for actual player box-score detail beyond starter ER/outs and listed adjustment observations. No unprovided player performance should be inferred.

---

2. Executive diagnosis

Overall result

Winner accuracy: 8/12 = 66.67%
Misses: 4
Mean team-run error: not directly supplied as a slate aggregate, but the pipeline reports published team-run MAE of 3.4167
Published margin mean: 1.42
Actual margin mean: 4.92
Predicted margins ≤2 runs: 11/12
Actual margins ≥4 runs: 6/12
Published margin compression: material
Published weighted absolute calibration error: 0.1661
Validated run-model holdout R²:
Away runs: 0.027504
Home runs: 0.037781
Total runs: 0.02717
Simulation winner interventions: 0 flips in 12 games

A correct winner was often accompanied by a poor score prediction. The slate produced 8 winner hits, but only two games were classified as solid_correct_call—CHW@HOU and MIA@WAS. The other six winner hits were either materially wrong scores or, in one case, a correct winner despite substantial starter/run error.

Separate diagnoses

Winner accuracy

Winner accuracy was acceptable on this slate at 66.7%, but not uniformly reliable:

Historical ML: 8/12
Variance-applied: 6/12
Enhanced prediction and all later stages: 8/12

The key process concern is not the final accuracy alone. The historical_ml → variance_applied transition regressed winner accuracy by 2 hits, from 8 to 6, before the enhanced stage recovered those two games. The affected direction changes were:

BAL@COL: BAL → TIE
NYY@LAA: LAA → TIE
PHI@ARI: ARI → TIE

The later recovery does not erase the intermediate regression.

Calibration quality

Calibration was poor on this 12-game slate:

Weighted absolute calibration error: 0.1661
0.50–0.55 bucket: average confidence 0.5296, observed hit rate 0.7143, error −0.1847 — underconfident
0.55–0.60 bucket: confidence 0.5555, observed hit rate 0.50, error +0.0555 — overconfident
0.65–0.70 bucket: confidence 0.6695, observed hit rate 0.50, error +0.1695 — overconfident
0.70–0.80 bucket: confidence 0.7492, observed hit rate 1.00, error −0.2508 — underconfident

The bucket behavior is unstable rather than monotonic. The issue register appropriately warns not to infer a stable calibration curve from seven games or from this one slate. The high-raw cluster also deserves attention: it contained 3 games and 2 hits, while the associated empirical hit rates were only 0.5519, 0.5567, and 0.5766 despite calibrated probabilities from 0.661 to 0.7435.

Score accuracy

Score accuracy was the weakest prediction dimension:

Published margin MAE: 4.6667
Published team-run MAE: 3.4167
Published total-run MAE: 4.3333
Mean predicted margin: 1.42
Mean actual margin: 4.92
Actual margin ≥4: 6/12
Predicted margin ≤2: 11/12
Validated run-model R² was near zero for away, home, and total runs.

The issue register’s score_margin_compression finding is therefore material and high severity. It does not justify mechanically widening every score. The correct response is rolling out-of-sample evaluation of variance width, tail coverage, margin MAE, and total-run MAE.

Starter accuracy

Starter projections were a major contributor to the four losses:

starter_projection: 3 primary loss drivers
Starter-line identity matched in all 12 games.
The worst starter ER errors included:
Kyle Harrison: +6 ER error
Robert Stock: −4
Ian Seymour: +3
Bryce Elder: +4
Gage Jump: +5
Payton Tolle: +4

The starter lever tension reports that pitcher-form/prior weighting would have helped 9 games and hurt 1, but this is a small slate and does not establish causal superiority.

Artifact integrity

Structural coverage was complete and the aggregate structural audit reports:

Critical: 0
Error: 0
Warning: 0
Info: 87

However:

The integrity revision audit is missing.
Underlying per-game info findings are absent despite observations claiming 7–8 non-pass findings per game.
The BAL probability trace contains a field-label discrepancy.

Therefore, artifact integrity is broadly passing at the critical/error gate but not fully clean as a forensic evidence package.

Causal issue register reconciliation

ID	Severity	Evidence	Affected games	Action	Guardrail
`pipeline_winner_instability`	High	`historical_ml → variance_applied` reduced winner accuracy by 0.1667, or 2 hits	BAL@COL, NYY@LAA; PHI@ARI also changed direction but recovered	Identify exact boundary flips and preserve pre/post winners	Later recovery does not make the unstable intermediate transition harmless
`score_margin_compression`	High	11/12 predicted margins ≤2; 6/12 actual margins ≥4; predicted mean 1.42 vs actual 4.92	Listed six direct affected games: DET@MIN, MIL@CHC, NYY@LAA, OAK@TEX, SD@CIN, SF@ATL	Evaluate distribution width, tail coverage, and margin MAE on rolling OOS slates	One blowout does not justify widening every point estimate
`single_slate_calibration_instability`	High	Weighted absolute calibration error 0.1661	All 12	Monitor buckets over registered rolling window	This slate cannot establish a stable calibration curve
`weak_run_model_signal`	High	Run R² only 0.0275/0.0378/0.0272	All 12	Investigate features/target through isolated training backtest	Positive R² only beats the mean baseline; exact scores remain unreliable
`complete_slate_audit_coverage`	Info	12/12 structurally audited	All 12	Retain per-game receipt	Coverage does not erase internal findings
`probability_publication_delta_explained`	Info	8 material movements, all attributed to score coherence and/or recency	Eight listed games	Continue displaying all probability layers	A higher published value is not automatically a governor bypass
`simulation_inert_on_slate`	Info on this slate, escalated persistent pattern across slates	0 candidates and 0 published flips	All 12	Measure simulation contribution and verify execution	Zero flips is not itself a defect and does not prove no score work occurred

---

3. Every game

The pipeline score path for all games was:

historical_ml → variance_applied → enhanced_prediction → embedded_savant_run_environment → fundamental_edge_target_review → capped_target → simulation_result → published_final

Only the first two transitions changed winner direction on this slate. Simulation and later stages produced no winner flips.

BAL@COL — hit, correct winner/wrong score

Predicted: BAL 7, COL 4; winner BAL
Actual: BAL 2, COL 1; winner BAL
Published pick probability: BAL 0.749242; home COL probability approximately 0.250758
System edge: home-minus-away −0.4985, described as a strong away edge
Conviction: Strong
Brier: 0.06288
Team-run absolute error: 8
Total-run bias: +8; predicted total 11 vs actual 3
Starter projections:
Cade Povich: projected 2 ER/14 outs; actual 1 ER/15 outs; ER error −1
Tanner Gordon: projected 2 ER/15 outs; actual 1 ER/15 outs; ER error −1
Edges:
Season home-minus-away: −1.0438
Recent: −0.976
Pitcher: −0.0247
Composite: −1.4124
Controls:
Governor applied
Governed probability: 0.7435
Score coherence brought publication to 0.749242
No recency guardrail
Score action: supported_blowout_margin_expansion
Score path:
Historical ML initially selected BAL.
Variance application changed direction to TIE.
Enhanced prediction restored BAL.
Later stages preserved BAL.
Player/run adjustments:
Away runs factor: 1.0376
Home runs factor: 1.0129
Savant adjustment applied
Season quality moved score toward BAL by a margin adjustment of −0.4384
Audit: status PASS; 0 critical/error/warning; 101 good and 7 info in summary, but detailed findings are absent.
Primary cause: Score/run-environment miss, despite accurate starters and correct winner.
Secondary cause: Intermediate variance-stage winner instability; it temporarily moved the winner to a tie.

CHW@HOU — hit, solid correct call

Predicted: CHW 4, HOU 5; winner HOU
Actual: CHW 3, HOU 6; winner HOU
Published HOU probability: 0.530
System edge: home-minus-away +0.06
Conviction: Lean
Brier: 0.2209
Team-run absolute error: 2
Total-run bias: 0
Starter projections:
Anthony Kay: projected 3 ER/16 outs; actual 5 ER/14 outs; ER error +2
Peter Lambert: projected 1 ER/16 outs; actual 2 ER/14 outs; ER error +1
Edges:
Season: −0.5164
Recent: −0.1033
Pitcher: +0.0587
Composite: −0.0007
Controls:
Governor applied: 0.52085 input to 0.5082 governed probability
Score coherence plus matchup-recency guardrail
Published probability rose to 0.530
Score action: no_change
Score path: No reported final-stage winner change; winner remained HOU.
Player/run adjustments:
Away factor 0.9817
Home factor 1.0404
Savant adjustment applied
Audit: PASS; 0 critical/error/warning; 100 good, 8 info.
Primary cause: None material.
Secondary cause: Mild starter projection error, but it did not damage the winner or score substantially.

DET@MIN — hit, correct winner/wrong score

Predicted: DET 4, MIN 5; winner MIN
Actual: DET 1, MIN 11; winner MIN
Published MIN probability: 0.530
System edge: +0.06 home
Conviction: Lean
Brier: 0.2209
Team-run absolute error: 9
Total-run bias: −3; predicted total 9 vs actual 12
Starter projections:
Jackson Jobe: projected 3 ER/14 outs; actual 4 ER/15 outs; error +1
Taj Bradley: projected 4 ER/16 outs; actual 1 ER/21 outs; error −3
Edges:
Season: −0.4842
Recent: +0.302
Pitcher: −0.0063
Composite: −0.0247
Controls:
Governor: 0.52085 input → 0.5082 governed
Recency guardrail and score coherence
Published: 0.530
Score action: no_change
Score path: Winner remained MIN after historical ML and all later stages.
Player/run adjustments:
Away factor 1.0322
Home factor 1.0259
Bradley allowed 1 ER vs 4 projected
Audit: PASS; 0 critical/error/warning; 102 good, 7 info.
Primary cause: Run distribution/margin compression.
Secondary cause: Starter projection error, especially Bradley’s three-ER overprojection.

MIA@WAS — hit, solid correct call

Predicted: MIA 3, WAS 4; winner WAS
Actual: MIA 3, WAS 6; winner WAS
Published WAS probability: 0.530
System edge: +0.06 home
Conviction: Lean
Brier: 0.2209
Team-run error: 2
Total-run bias: −2
Starter projections:
Ryan Gusto: projected 4 ER/14 outs; actual 1 ER/13 outs; error −3
Will Dion: projected 1 ER/15 outs; actual 0 ER/9 outs; error −1
Edges:
Season: −0.0497
Recent: −0.122
Pitcher: −0.0621
Composite: −0.0145
Controls:
Governor: 0.49585 input → 0.5082 governed
Recency guardrail plus score coherence
Published: 0.530
Score action: no_change
Diagnostic flag: large_savant_run_environment_swing
Score path: Winner stayed WAS.
Player/run adjustments:
Away factor 1.0000
Home factor 0.9379
Gusto allowed one ER versus four projected.
Audit: PASS; 0 critical/error/warning; 101 good, 8 info.
Primary cause: None material.
Secondary cause: Starter ER overprojection, offset by a reasonably close team-run result.

MIL@CHC — miss

Predicted: MIL 6, CHC 5; winner MIL
Actual: MIL 3, CHC 17; winner CHC
Published pick probability: MIL 0.530, CHC 0.470
System edge: home-minus-away −0.06, a lean MIL edge
Conviction: Lean
Brier: 0.2809
Team-run absolute error: 15
Total-run bias: −9; predicted total 11 vs actual 20
Starter projections:
Kyle Harrison: projected 3 ER/17 outs; actual 9 ER/11 outs; error +6
Clay Holmes: projected 2 ER/18 outs; actual 0 ER/18 outs; error −2
Edges:
Season: −0.3922
Recent: −0.350
Pitcher: −0.1125
Composite: −0.7307
Controls:
Governor: 0.530 input → 0.5082 governed
Recency guardrail plus score coherence
Published pick remained MIL at 0.530
Score action: no_change
Diagnostic flag: large_savant_run_environment_swing
Score path: The final winner remained MIL; no later-stage winner recovery occurred.
Player/run adjustments:
Away factor 1.0498
Home factor 0.9373
Harrison’s six-ER miss was the largest starter error on the slate.
Audit: PASS; 0 critical/error/warning; 101 good, 7 info.
Primary cause: Starter projection failure, especially Harrison.
Secondary cause: Run-environment under-dispersion; CHC scored 12 more runs than projected and the total was nine runs too low.

NYM@TB — miss

Predicted: NYM 3, TB 5; winner TB
Actual: NYM 3, TB 2; winner NYM
Published TB probability: 0.6655
System edge: +0.331 home
Conviction: Strong
Brier: 0.44289, worst on the slate
Team-run absolute error: 3
Total-run bias: +3
Starter projections:
Robert Stock: projected 4 ER/12 outs; actual 0 ER/16 outs; error −4
Ian Seymour: projected 0 ER/15 outs; actual 3 ER/19 outs; error +3
Edges:
Season: +0.7963
Recent: +0.380
Pitcher: −0.0457
Composite: +0.7924
Controls:
Governor: 0.666292 input → 0.661 governed
No recency guardrail
Score coherence produced published 0.6655
Score action: no_change
Score path: Winner stayed TB through publication; simulation did not intervene.
Player/run adjustments:
Away factor 1.0042
Home factor 0.9585
Audit: PASS; 0 critical/error/warning; 100 good, 8 info.
Primary cause: Starter projection failure, with both starters materially misestimated.
Secondary cause: Wrong conviction side; the system backed TB with a genuine 0.33 edge and lost.
Additional classification: This was a high-confidence calibration failure, not merely a close-game miss.

NYY@LAA — miss

Predicted: NYY 5, LAA 4; winner NYY
Actual: NYY 1, LAA 10; winner LAA
Published NYY probability: 0.550872; LAA approximately 0.449128
System edge: home-minus-away −0.1017
Conviction: Moderate
Brier: 0.30346
Team-run absolute error: 10
Total-run bias: −2
Starter projections:
Elmer Rodríguez: projected 2 ER/14 outs; actual 4 ER/11 outs; error +2
Walbert Ureña: projected 2 ER/17 outs; actual 1 ER/21 outs; error −1
Edges:
Season: −1.3728
Recent: −0.696
Pitcher: +0.0456
Composite: −1.0087
Controls:
Governor: 0.570 input → 0.591 governed
Recency guardrail applied
Score coherence probability 0.588013
Published NYY probability reduced to 0.550872
Score action: supported_blowout_margin_expansion
Score path:
Historical ML: LAA → TIE
Variance stage was an intermediate regression.
Enhanced prediction: TIE → NYY
Later stages retained NYY.
Player/run adjustments:
Away factor 0.9521
Home factor 0.9977
Yankees offense was four runs overprojected; LAA offense six underprojected.
Audit: PASS; 0 critical/error/warning; 101 good, 7 info.
Primary cause: Wrong conviction side, though the edge was moderate rather than strong.
Secondary cause: Lean-side miss and score compression; actual margin was 9 while predicted margin was 1.
Required interpretation: a 0.12 home-away confidence edge is moderate, not a near coin flip; the system’s moderate NYY edge was nevertheless on the wrong side.

OAK@TEX — hit, correct winner/wrong score

Predicted: OAK 3, TEX 5; winner TEX
Actual: OAK 1, TEX 8; winner TEX
Published TEX probability: 0.530
System edge: +0.06 home
Conviction: Lean
Brier: 0.2209
Team-run absolute error: 5
Total-run bias: −1
Starter projections:
Gage Jump: projected 2 ER/15 outs; actual 7 ER/13 outs; error +5
Jacob deGrom: projected 1 ER/15 outs; actual 0 ER/21 outs; error −1
Edges:
Season: +0.8297
Recent: +0.118
Pitcher: +0.0056
Composite: +0.8582
Controls:
Governor: 0.52085 → 0.5082
Recency guardrail and score coherence
Published: 0.530
Score action: no_change
Diagnostic flag: large_savant_run_environment_swing
Score path: Winner remained TEX.
Player/run adjustments:
Away factor 0.9634
Home factor 0.9690
Jump’s five-ER error was material.
Audit: PASS; 0 critical/error/warning; 101 good, 7 info.
Primary cause: Starter projection error, although it did not reverse the winner.
Secondary cause: Margin compression.

PHI@ARI — hit, correct winner/wrong score

Predicted: PHI 5, ARI 4; winner PHI
Actual: PHI 2, ARI 1; winner PHI
Published PHI probability: 0.526943
System edge: −0.0539 home, favoring PHI
Conviction: Lean
Brier: 0.223783
Team-run absolute error: 6
Total-run bias: +6
Starter projections:
Aaron Nola: projected 4 ER/16 outs; actual 1 ER/15 outs; error −3
Brandon Pfaadt: projected 1 ER/19 outs; actual 2 ER/18 outs; error +1
Edges:
Season: −0.3008
Recent: −0.3113
Pitcher: −0.1097
Composite: −0.6646
Controls:
Governor: 0.530 → 0.5082
Recency guardrail and score coherence
Published: 0.526943
Score action: no_change
Score path:
Historical ML: ARI → TIE
Variance stage: remained tie.
Enhanced prediction: TIE → PHI
Later stages preserved PHI.
Player/run adjustments:
Away factor 1.0373
Home factor 0.9276
Nola’s three-ER overprojection contributed to the six-run total error.
Audit: PASS; 0 critical/error/warning; 101 good, 7 info.
Primary cause: Score/run-environment miss.
Secondary cause: Intermediate variance-stage direction change, later correctly recovered.

SD@CIN — hit, correct winner/wrong score

Predicted: SD 6, CIN 4; winner SD
Actual: SD 5, CIN 0; winner SD
Published SD probability: 0.673503
System edge: −0.347 home, favoring SD
Conviction: Strong
Brier: 0.1066
Team-run absolute error: 5
Total-run bias: +5
Starter projections:
Michael King: projected 3 ER/18 outs; actual 0 ER/18 outs; error −3
Brady Singer: projected 4 ER/15 outs; actual 4 ER/13 outs; error 0
Edges:
Season: −0.6951
Recent: −0.3607
Pitcher: −0.0053
Composite: −0.8441
Controls:
Governor: 0.691798 → 0.677
No recency guardrail
Score coherence led to published 0.673503
Score action: supported_blowout_margin_expansion
Score path: Winner remained SD.
Player/run adjustments:
Away factor 1.0072
Home factor 1.0019
King allowed zero ER rather than three.
Audit: PASS; 0 critical/error/warning; 101 good, 7 info.
Primary cause: Score/run-environment miss, especially CIN’s zero runs.
Secondary cause: Starter error from King’s three-ER overprojection.
This was a correct strong call with an incorrect exact score.

SEA@BOS — hit, correct winner/wrong score

Predicted: SEA 2, BOS 3; winner BOS
Actual: SEA 8, BOS 9; winner BOS
Published BOS probability: 0.560186
System edge: +0.1204 home
Conviction: Moderate
Brier: 0.193436
Team-run absolute error: 12
Total-run bias: −12; predicted total 5 vs actual 17
Starter projections:
George Kirby: projected 2 ER/17 outs; actual 5 ER/15 outs; error +3
Payton Tolle: projected 0 ER/18 outs; actual 4 ER/17 outs; error +4
Edges:
Season: +0.9569
Recent: +0.5327
Pitcher: +0.0158
Composite: +0.8742
Controls:
Governor: 0.58865 → 0.5816
Recency guardrail plus score coherence
Published: 0.560186
Score action: no_change
Score path: Winner remained BOS.
Player/run adjustments:
Away factor 0.9593
Home factor 0.9751
Both starters underperformed projections.
Audit: PASS; 0 critical/error/warning; 102 good, 7 info.
Primary cause: Run-environment miss/under-dispersion.
Secondary cause: Two-sided starter underestimation.
The 0.1204 edge was moderate, not a near coin flip, and the winner was correct despite an extremely poor score forecast.

SF@ATL — miss

Predicted: SF 4, ATL 5; winner ATL
Actual: SF 7, ATL 3; winner SF
Published ATL probability: 0.530
System edge: +0.06 home
Conviction: Lean
Brier: 0.2809
Team-run absolute error: 5
Total-run bias: −1
Starter projections:
Anthony Molina: projected 4 ER/9 outs; actual 3 ER/15 outs; error −1
Bryce Elder: projected 1 ER/18 outs; actual 5 ER/16 outs; error +4
Edges:
Season: +1.2284
Recent: +0.182
Pitcher: −0.0832
Composite: +0.7181
Controls:
Governor: 0.507202 → 0.5082
Recency guardrail and score coherence
Published: 0.530
Score action: no_change
Diagnostic flag: large_savant_run_environment_swing
Score path: Winner remained ATL; no recovery.
Player/run adjustments:
Away factor 1.0832
Home factor 1.0000
Elder allowed five ER rather than one.
Audit: PASS; 0 critical/error/warning; 102 good, 7 info.
Primary cause: Starter projection error, specifically Elder.
Secondary cause: Lean-side miss and margin compression.

Correct-winner/wrong-score games

The six explicit right_winner_wrong_score games were:

BAL@COL
DET@MIN
OAK@TEX
PHI@ARI
SD@CIN
SEA@BOS

CHW@HOU and MIA@WAS were classified as solid correct calls. MIL@CHC, NYM@TB, NYY@LAA, and SF@ATL were winner misses.

---

4. Miss reconstruction

MIL@CHC

Earliest wrong-direction stage

The prediction was already on the wrong side in the historical model and stayed wrong through publication. The data does not show a later stage moving the winner toward CHC. The earliest identifiable problem is therefore the base winner/run assessment, not a later transition.

Cause classification

Bad input: Not established. Starter identities matched and no contract error was reported.
Bad weighting: Possible, given the strong composite edge favoring MIL despite the outcome, but not proven.
Calibration: Secondary. Published probability was only 0.530, so the confidence was not extreme.
Starter error: Major. Harrison projected 3 ER/17 outs, actual 9 ER/11 outs.
Offense/bullpen miss: Material but not isolated. CHC scored 17 versus 5 projected; MIL scored 3 versus 6 projected.
Run-environment miss: Major. Total was underprojected by 9.
Irreducible variance: Present, especially the extreme 17–3 result, but the six-ER starter error is a modelable component.

NYM@TB

Earliest wrong-direction stage

The historical/base model already selected TB, and no later stage corrected it. The earliest evidence moving toward the wrong result was the high composite/season edge for TB:

Season edge: +0.7963 home
Recent edge: +0.380
Composite: +0.7924

The pitcher edge slightly favored NYM at −0.0457, but it did not overcome the broader TB signal.

Cause classification

Bad input: Not established.
Bad weighting: Plausible. Season/recent offense-quality evidence outweighed the pitcher signal in a way that failed this matchup.
Calibration: Material. TB was published at 0.6655, but lost; Brier was 0.44289.
Starter error: Major and two-sided. Stock allowed 0 instead of 4 projected; Seymour allowed 3 instead of 0.
Offense/bullpen miss: TB scored 2 versus 5 projected, contributing to the miss.
Run-environment miss: Moderate. Total was overprojected by 3.
Irreducible variance: Material, because both starters deviated substantially.

NYY@LAA

Earliest wrong-direction stage

There was a documented intermediate regression:

Historical ML: LAA → TIE
Variance application: tie state
Enhanced prediction: TIE → NYY

The first clear move toward the wrong final result was the enhanced prediction’s recovery toward NYY, after the variance stage had already removed LAA’s direction. This is exactly the pipeline-instability issue: later recovery does not make the intermediate regression harmless.

The base evidence itself strongly favored NYY on season and recent quality:

Season: −1.3728
Recent: −0.696
Composite: −1.0087

Cause classification

Bad input: Not established.
Bad weighting: Possible, because the large season/recent NYY advantage did not survive the realized matchup.
Calibration: Moderate concern. Published NYY probability was 0.550872, not high, but it was still on the losing side.
Starter error: Moderate. Rodríguez allowed 4 versus 2 projected; Ureña allowed 1 versus 2.
Offense/bullpen miss: Major. NYY scored 1 versus 5 projected; LAA scored 10 versus 4 projected.
Run-environment miss: Major. Actual margin was 9 versus predicted margin 1.
Irreducible variance: Present, but the six-run combined offensive miss and nine-run margin miss indicate more than ordinary winner variance.

SF@ATL

Earliest wrong-direction stage

The historical/base direction favored ATL and no later stage corrected it. The earliest wrong-direction evidence was the combination of:

Season edge: +1.2284 home
Recent edge: +0.182
Composite: +0.7181

The pitcher edge favored SF at −0.0832, but was insufficiently weighted to reverse the call.

Cause classification

Bad input: Not established.
Bad weighting: Plausible. The model materially favored ATL’s season/quality evidence while the pitcher edge favored SF.
Calibration: Limited concern, because publication was only 0.530.
Starter error: Primary. Elder allowed 5 versus 1 projected.
Offense/bullpen miss: SF scored 7 versus 4 projected; ATL scored 3 versus 5 projected.
Run-environment miss: Moderate; margin was 4 versus 1 predicted.
Irreducible variance: Present, especially Elder’s four-ER projection error.

Pipeline-stage quality across all stages

Stage	Winner accuracy	Winner hits	Team-run MAE	Total-run MAE	Margin MAE
historical_ml	0.6667	8	3.1283	3.9910	4.7023
variance_applied	0.5000	6	3.1250	3.9167	4.7500
enhanced_prediction	0.6667	8	3.4583	4.2500	4.7500
embedded_savant_run_environment	0.6667	8	3.4167	4.3333	4.6667
fundamental_edge_target_review	0.6667	8	3.4167	4.3333	4.6667
capped_target	0.6667	8	3.4167	4.3333	4.6667
simulation_result	0.6667	8	3.4167	4.3333	4.6667
published_final	0.6667	8	3.4167	4.3333	4.6667

Transitions:

Historical ML → variance:
Winner accuracy −0.1667
Winner hits −2
Margin MAE +0.0477
Regressed: BAL@COL, NYY@LAA
Also changed PHI@ARI from ARI to tie
Variance → enhanced:
Winner accuracy +0.1667
Winner hits +2
BAL@COL and PHI@ARI improved
Team-run MAE worsened by 0.3333
Enhanced → Savant:
No winner-direction changes
Margin MAE improved 0.0833
Team-run MAE improved 0.0416
All subsequent transitions:
No winner changes
No metric changes

The intermediate winner regression remains a genuine process risk even though enhanced prediction later recovered the final winner count.

---

5. Cross-game counterfactuals

Lever tension: widen run distributions

Lever: score.variance_multiplier / score.variance_floor

Helps:

NYM@TB: lost a 0.33-confidence pick; wider tails could reduce overconfidence.
NYY@LAA: lost a 0.10-confidence pick; wider tails could reduce overconfidence.

Hurts:

BAL@COL: strong correct call at 0.50; widening could soften a good edge.
SD@CIN: strong correct call at 0.35; widening could soften a good edge.

Net evidence: 2 helps vs 2 hurts. The pattern supports a bounded shadow experiment, not an unrestricted production increase. It targets Brier score and calibration error, while score-margin compression is also relevant.

Lever tension: pitcher form/prior weighting

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

Helps:

DET@MIN
MIA@WAS
MIL@CHC
NYM@TB
OAK@TEX
PHI@ARI
SD@CIN
SEA@BOS
SF@ATL

Hurts:

BAL@COL, where both starter projections were accurate within one ER.

Net evidence: 9 helps vs 1 hurts, but this is based on starter error observations, not a controlled counterfactual. It supports a diagnostic/backtest, not immediate aggressive reweighting.

Lever tension: recent/season run-mean weighting

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

Helps:

BAL@COL
DET@MIN
MIL@CHC
NYY@LAA
OAK@TEX
PHI@ARI
SD@CIN
SEA@BOS
SF@ATL

Hurts:

CHW@HOU, where team-run error was only 2.
MIA@WAS, where team-run error was only 2.

Net evidence: 9 helps vs 2 hurts for team-run and total-run metrics, but the reported “helps” do not prove a particular direction of weight change. The weak run-model R² argues for feature/target investigation before tuning this lever.

One-game fixes that must not be made

Do not widen every prediction merely because MIL@CHC and SEA@BOS were blowouts.
Do not globally downgrade strong confidence because NYM@TB lost; BAL@COL and SD@CIN were strong correct calls.
Do not remove or sharply reduce all season-quality weighting because NYY@LAA and SF@ATL missed; it supported several correct winners.
Do not treat the 0.1204 SEA@BOS edge as a near coin flip; it is moderate.
Do not force a simulation parameter move solely because simulation produced zero winner flips.
Do not alter calibration weights from this slate alone; the bucket sample sizes are small and inconsistent.
Do not conclude that a later enhanced-stage recovery makes the variance transition safe.

---

6. Data and contract defects

Critical findings

The structural audit summary reports 0 critical. No critical contract failure is evidenced in the supplied JSON.

The missing integrity revision audit is an evidence gap, but the data does not classify it as critical.

Error findings

Structural audit summary: 0 errors.
No per-game error finding details are supplied.

Warning findings

Structural audit summary: 0 warnings.
No per-game warning details are supplied.

Info findings

Structural audit summary: 87 info findings across the slate.
Per-game summaries report 7 or 8 info findings each.
The actual info-finding text is missing from the JSON, so the 87 findings cannot be individually enumerated.
Registered informational findings include:
Complete slate audit coverage
Explained probability-publication movements
Simulation inertness on this slate

Prediction-model defects versus presentation/artifact defects

Prediction-model/process defects

Intermediate winner instability at historical_ml → variance_applied.
Compressed score margins and under-dispersed run distributions.
Very weak run-model holdout signal.
Starter projection errors materially affecting three misses.
Calibration misalignment on the current slate.
Simulation producing no winner-direction interventions over four consecutive slates, requiring execution diagnosis.

Presentation/artifact defects

Missing integrity revision audit artifact.
BAL raw probability field appears inconsistently named or populated between game trace and lineage.
Per-game audit observations claim 7–8 non-pass findings while the embedded findings arrays are empty. This prevents independent audit of those findings.
“PASS” status should not be interpreted as absence of all info-level or non-pass observations; the report itself says complete coverage does not erase findings.

Probability-lineage audit

The complete pick-side lineage contains 12 rows. The recorded layers are:

raw model → calibrated → governor input/output → score coherence and/or recency → published

The data identifies 8 material calibrated-to-published movements, and states that all were explained by named score-coherence and/or recency transformations. No material unexplained delta is marked in the lineage rows.

Examples:

CHW@HOU, DET@MIN, MIA@WAS, MIL@CHC, OAK@TEX, PHI@ARI, SEA@BOS, SF@ATL: recency guardrail and/or score coherence are named.
BAL@COL, NYM@TB, and SD@CIN: score coherence is named; no recency guardrail.
NYY@LAA: score coherence followed by matchup-recency guardrail; publication decreased from calibrated 0.591 to 0.550872.

There is no demonstrated governor bypass: every material published movement has a recorded transformation and unexplained_material_delta: false. The BAL field-label discrepancy is a trace consistency issue, not a material unexplained publication delta.

---

7. Cross-slate persistence

Three prior slates were loaded, producing a four-slate window from 20260828 through 20260831. A single slate is noise, but the escalated findings are explicitly established patterns and must not be treated as single-slate noise.

Raw structural evidence

Margin distribution

Current slate:

Actual margins: 1, 1, 1, 1, 3, 3, 4, 5, 7, 9, 10, 14
Predicted margins: 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 3
Actual margin ≥4: 6/12
Predicted margin ≤2: 11/12
Current mean signed margin error, based on listed rows, is strongly negative for the model’s absolute margin representation: the model systematically undercalled realized separation in most large-margin games.

Pipeline transitions

The raw transition block shows a current winner-direction regression at historical_ml → variance_applied:

BAL@COL and NYY@LAA regressed.
PHI@ARI also changed direction but was later recovered.
Accuracy fell from 8/12 to 6/12 and recovered to 8/12 at enhanced prediction.

This finding has a two-slate consecutive streak and is present now.

Calibration monitoring

Current WACE is 0.1661. Bucket errors alternate between underconfidence and overconfidence, confirming instability rather than a simple universal overconfidence bias.

High-raw cluster

Current high-raw cluster:

3 games
2 hits
Threshold: 0.65
Calibrated probabilities: 0.661, 0.677, 0.7435
Empirical hit rates attached to those rows: 0.5567, 0.5766, 0.5519

The finding streak reports high_raw_probability_cluster at 0 consecutive slates, present current slate false. It is therefore not a current persistent finding.

Simulation control

Current slate:

Observed games: 12
Flip candidates: 0
Published flips: 0
Status: diagnostic only without independent support

Across slates, simulation inertness is established, not noise.

Finding streaks

Finding	Consecutive slates	Present now?	Interpretation
`high_raw_probability_cluster`	0	No	Not persistent/current
`pipeline_winner_instability`	2	Yes	Recurrent intermediate winner-direction regression
`score_margin_compression`	1	Yes	Current issue, not yet a multi-slate streak
`simulation_inert_on_slate`	4	Yes	Established persistent failure
`single_slate_calibration_instability`	1	Yes; incidence 3 in window	Current calibration problem, but not a four-slate persistence pattern
`weak_run_model_signal`	4	Yes	Established persistent weak run-model signal

Escalated findings

`simulation_inert_on_slate`

Consecutive slates: 4
Present now: Yes
Target lever listed: simulation.allow_score
Parameter actionable: No
Metric harmed: team_run_mae, total_run_mae
Direction: confirm simulation module is running and not silently no-oping.
Required interpretation: established code/diagnostic failure, not single-slate noise.

`weak_run_model_signal`

Consecutive slates: 4
Present now: Yes
Listed training levers:
training.xgb.n_estimators
training.xgb.max_depth
training.xgb.learning_rate
Parameter actionable: No
Metrics harmed: team_run_mae, total_run_mae
Direction: investigate features and target through an isolated training backtest.
Required interpretation: established run-model weakness, not a reason to force a parameter move.

Metric trends

margin_bias: Regressing
Series: −1.333, −1.882, −2.786, −3.5
Mean: −2.3752
This is consistent with increasing margin underprediction.
mean_team_run_error_on_losses: Regressing
Series: 5.0, 3.0, 7.0, 8.25
Mean: 5.8125
weighted_calibration_error: Improving
Series: 0.2179, 0.3961, 0.0693, 0.1661
Improvement is directionally favorable over the window, but the current value remains poor and bucket behavior is unstable.
winner_accuracy: Flat
Series: 0.4667, 0.9412, 0.5, 0.6667
Mean: 0.6437
Large variation, no reliable directional improvement.

Top-confidence-pick loss streak

Loss streak: 0
Per-slate results:
20260828: hit
20260829: hit
20260830: miss
20260831: hit

There is no current top-confidence-pick loss streak.

---

8. Adjustments for today

Recommendations are ranked by persistence and evidence strength first, then scope and regression risk.

1. Diagnose and repair simulation execution — highest persistence

Pattern: simulation_inert_on_slate
Persistence: 4 consecutive slates
Exact registered lever: simulation.allow_score
Parameter change: None initially; do not force a parameter move because the finding is explicitly non-parameter-actionable.
Concrete action: run a code-level execution diagnostic confirming that simulation.allow_score is read, the simulation loop executes, score outcomes are generated, and flip candidates are recorded.
Expected metric: team-run MAE and total-run MAE should become measurable at the simulation stage; direction should be improvement or at minimum non-identical behavior when score simulation is enabled.
Helps: all games with score/run errors, particularly MIL@CHC, NYY@LAA, SEA@BOS, BAL@COL, and PHI@ARI.
Hurts/regression risk: an improperly broad simulation could alter strong correct calls such as BAL@COL and SD@CIN.
Gate: shadow execution with stage-by-stage before/after trace; rollback if winner accuracy or Brier worsens without run-error improvement.

2. Isolate the weak run model before tuning training parameters

Pattern: weak_run_model_signal
Persistence: 4 consecutive slates
Exact registered levers: training.xgb.n_estimators, training.xgb.max_depth, training.xgb.learning_rate
Parameter change: No immediate production move. The finding is marked param_actionable: false.
Concrete action: perform the registered isolated training backtest, testing the feature set and target definition before selecting a bounded hyperparameter shadow.
Expected metric: improve away/home/total run-model holdout R² and reduce team-run MAE and total-run MAE.
Helps: the entire slate, especially MIL@CHC, SEA@BOS, NYY@LAA, BAL@COL, and PHI@ARI.
Hurts/regression risk: overfitting could damage close, well-calibrated games such as CHW@HOU and MIA@WAS.
Gate: rolling out-of-sample improvement in MAE and tail coverage, not in-sample fit.

3. Run a bounded shadow increase to score-distribution width

Pattern: score_margin_compression
Persistence: 1 consecutive slate; present now
Exact parameter: score.variance_multiplier
Proposed change: increase by 5% in shadow only. Do not simultaneously change score.variance_floor, so the effect is attributable.
Expected metric: lower margin MAE and total-run MAE; reduce underconfidence/overconfidence distortion caused by compressed score distributions. Winner Brier should be monitored, not assumed to improve.
Helps:
NYM@TB
NYY@LAA
MIL@CHC
SEA@BOS
BAL@COL
SD@CIN
Hurts:
BAL@COL and SD@CIN are explicitly identified as good strong calls that may be softened.
More generally, widening can reduce exact-score sharpness in CHW@HOU and MIA@WAS, although those games are not listed in this lever’s direct tension set.
Evidence strength: moderate for the current slate, weaker than the two four-slate escalations.
Gate: retain only if rolling margin MAE or tail coverage improves without unacceptable Brier or winner-accuracy regression.

4. Do not move calibration caps today; use shadow monitoring

Pattern: single_slate_calibration_instability
Current persistence: 1 consecutive slate, incidence 3 in the four-slate window
Registered parameters: confidence.empirical_max_weight, confidence.base_cap, confidence.moderate_cap
Proposed change: no production parameter move because the issue register explicitly says to monitor the same buckets over the registered rolling window and warns that seven games cannot establish a curve.
Expected metric: future Brier and calibration error should be monitored by bucket.
Helps if later supported: especially high-raw losses such as NYM@TB and NYY@LAA.
Hurts if overcorrected: strong correct calls BAL@COL and SD@CIN.
This is a bounded abstention based on the registered guardrail, not a deferral of diagnosis.

5. Starter model diagnostic, not an immediate global reweight

Single-slate driver: starter projection was the primary driver in 3/4 losses.
Lever tension: player.pitcher.v4 form weights / player.pitcher_prior_games
The supplied causal issue register has no registered target lever for starter projection, so no parameter should be invented or forcibly changed here.
Concrete action: compare form-weight and prior-game variants in a shadow backtest, focusing on Harrison, Stock/Seymour, Rodríguez/Ureña, and Elder-type errors.
Expected metric: lower starter ER MAE and starter outs MAE.
Helps: 9 games in the lever-tension record.
Hurts: BAL@COL, where starter projections were accurate.
Regression control: do not promote unless rolling starter ER MAE improves while winner Brier and team-run MAE do not materially worsen.

---

9. Bottom line

Known

The system correctly picked 8 of 12 winners.
Six of those winner hits had materially wrong scores; winner accuracy therefore overstates prediction quality.
Score margins were systematically compressed: predicted mean margin 1.42 versus actual 4.92, with 11/12 predictions at two runs or fewer and 6/12 actual margins at four runs or more.
Starter projection error was the primary labeled cause in three of four losses.
The historical-to-variance transition caused a real intermediate winner-direction regression of two hits, even though enhanced prediction later recovered them.
The run model has weak persistent holdout signal across four slates.
Simulation has been inert across four consecutive slates and requires code/diagnostic verification.
All 12 games were structurally covered; no critical, error, or warning findings are reported.
The integrity revision artifact is missing, and per-game info findings are not embedded despite being referenced.
Pick-side probability lineage is materially explained; no governor bypass is evidenced.

Uncertain

Whether widened score distributions would improve future Brier/calibration without harming strong calls.
Whether pitcher-form weighting would improve out-of-sample starter accuracy rather than merely explain this slate.
Whether season/recent run weighting should move in a particular direction.
Whether the run-model weakness is caused by features, target construction, hyperparameters, or an upstream execution problem.
Whether simulation is genuinely inert by design on these games or silently failing.

Single safest next action

Perform the four-slate code-and-lineage diagnostic for simulation.allow_score and the weak run-model execution path before changing model weights. It addresses two established persistent findings, has a reversible shadow-gate path, and avoids making an unsupported one-slate calibration or score-distribution adjustment while also checking the most important artifact/process risk.

Deep slate report — 20260831

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 (12)

BAL@COLHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

CHW@HOUHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

DET@MINHIT

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

MIA@WASHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

MIL@CHCMISS

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@TBMISS

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

NYY@LAAMISS

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

OAK@TEXHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

PHI@ARIHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick

Score and confidence evolution

Official postgame stage attribution

Largest player-level adjustments from the ultra-detailed trace

Structural audit findings

SD@CINHIT

Player-level breakdown (projected vs actual)

Fundamental evidence entering the pick