MLB Prediction System

Metric Evolution & Audit

Every number that moved, the stage that moved it, the input it used, and the arithmetic that proves it

Slate 20261007Games 4Checks 627Schema metric_evolution_report_v1Built 2026-10-08 09:14Z

Verdict across the slate

627 invariants checked across 4 games: 3 OF 627 CHECKS FAILED

away staff totals a regulation workload: 30 outs recorded, 27 expected
home staff totals a regulation workload: 30 outs recorded, 27 expected
away staff totals a regulation workload: 26 outs recorded, 24 expected

How to read this, and what a defect looks like

Every figure below is shown with the arithmetic that produced it and a verdict beside it, so a disagreement can be pointed at a specific line rather than at the system as a whole.

Stage chain. Each game lists every stage that touched the score, the score after it, and whether it changed the target. Most stages move nothing on most games; a chain that showed only movement would misrepresent it. A stage reporting a delta while changed_target is false computed an adjustment and did not apply it, which is normal and is stated.
Blend arithmetic. The KPI blend must satisfy blended = baseline + weight x (kpi_only - baseline) for all three targets. A FAIL here means the layer reported a number it did not compute.
Rounding. Published integers must be the rounding of the blended runs under the documented tie rule. A FAIL means the published score and the continuous line describe different games.
Log5 additivity. For every player metric, own skill plus opponent plus the Log5 remainder must equal the total against the league, exactly. A FAIL means the decomposition is not a decomposition.
Direction. A green delta helps the subject, a red one hurts him. Strikeouts invert for a hitter. If an elite arm's opponent shows green strikeouts, the orientation is wrong and the whole player section is suspect.
Measured versus published. The KPI layer moves team runs and the win probability. The per-player metric adjustments here are a measurement of the same catalog at player level, not what the pipeline published; the one club-wide ratio the pipeline actually applied to hitter lines is printed beside them.
Integer versus continuous. The integer player line is a simulation draw reconciled to the box score. The continuous expected_* beside it is the model's own view and is the rankable quantity. Judging the model by the integer is how a working model looks broken.

Slate overview

Game	Published	p(home)	Stages	Stages that moved it	KPI applied	Δ away runs	Δ home runs	Simulation draw chosen on	0-out starter
20261007_CLE@CHW	5 - 4	0.4649	8	2	yes	+0.0005	+0.4454	-	no
20261007_LAD@ATL	6 - 4	0.4928	8	2	yes	+0.0409	+0.6826	-	no
20261007_MIL@SD	6 - 5	0.4950	8	3	yes	-0.1958	+0.3883	-	no
20261007_TB@NYY	3 - 4	0.5440	8	2	yes	+0.4627	+0.1996	-	no

KPI model in force. candidate 20261008T091425Z_3ccbc4c8, 48 features built from 40 registered KPIs, blend weights {"home_runs": 0.38, "away_runs": 0.34, "home_win": 0.0}, frozen evaluation ends 2026-09-27. A target whose weight is zero contributes nothing by construction: the winner probability carries weight zero while its walk-forward verdict is inconclusive, so every win-probability movement below is expected to be exactly zero.

Per game

20261007_CLE@CHW — published 5-4, 157 checks, all pass

Artifact. 20261007_CLE@CHW__20261008T024006Z.json, phase post_start_rerun, generated 2026-10-08T02:40:06, first pitch 2026-10-07T20:00:00.
Published 5 - 4 with p(home) 0.464871. Simulation draw chosen on not recorded; ? of ? draws realized the target exactly.

Stage chain: every stage that touched the score

Stage	Score after	Δ away	Δ home	Cum. away	Cum. home	Changed target	Own adjustment and description
1. `historical_ml`	4.629906778357269 - 2.801656780983749	0	0	0	0	no	Base ML historical model prediction
2. `variance_applied`	5 - 3	+0.3701	+0.1983	+0.3701	+0.1983	yes	Deterministic baseline target (variance factor 1.0)
3. `enhanced_prediction`	5 - 4	0	+1.0000	+0.3701	+1.1983	yes	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
4. `kpi_layer`	5 - 4	0	0	+0.3701	+1.1983	no	away +0.0005; home +0.4454 Savant KPI blend applied to the model target
5. `embedded_savant_run_environment`	5 - 4	0	0	+0.3701	+1.1983	no	away +0.0538; home -0.2578 Target score adjusted by temporal Savant lineup quality versus opposing starter quality
6. `fundamental_edge_target_review`	5 - 4	0	0	+0.3701	+1.1983	no	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
7. `capped_target`	5 - 4	0	0	+0.3701	+1.1983	no	Capped at max 12 realistic runs
8. `simulation_result`	5 - 4	-	-	-	-	no	Final inning-by-inning simulation result

A stage reporting its own adjustment while Changed target reads no computed a movement that a later reconciliation absorbed, or that rounded away at the integer score. That is recorded rather than hidden.

KPI team layer: the blend, with its arithmetic

Target	Baseline	KPI only	Weight	Blended	Computation	Verdict
`away_runs`	5.0000	5.0015	0.34	5.0005	5.0000 + 0.34 × (5.0015 − 5.0000) = 5.0005	PASS
`home_runs`	4.0000	5.1721	0.38	4.4454	4.0000 + 0.38 × (5.1721 − 4.0000) = 4.4454	PASS
`home_win`	0.4514	0.5275	0.00	0.4514	0.4514 + 0.00 × (0.5275 − 0.4514) = 0.4514	PASS

Published by the layer: away 5, home 4, p(home) 0.451379, continuous away 5.0005 / home 4.4454

Inputs and gates. candidate 20261007T091423Z_47b9db81, value age 10 days, lineup coverage away 1.0000 / home 0.8947, movement cap 2.5 runs.
values_present ok values_fresh ok one_source_bundle ok lineup_coverage ok target_eligible ok movement_within_limit ok published_movement_within_limit ok

What drove the KPI forecast for `away_runs`

intercept 4.2358 + sum of contributions = 5.0015; prediction 5.0015; contributions outside the top listed -0.6178 runs

Feature	Raw	Standardised	Runs contributed	Which KPIs it is built from
`home_off_power_spread`	1.1914	1.740	+0.9154	offence group (dispersion across the lineup, not its mean): batter_home_run_rate, batter_bat_speed
`away_off_contact`	-0.3389	-1.678	-0.5764	offence group: batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`away_off_discipline_spread`	0.9673	2.572	+0.5713	offence group (dispersion across the lineup, not its mean): batter_whiff_rate_vs_fastball, batter_whiff_rate_vs_breaking, batter_whiff_rate_vs_offspeed, batter_strikeout_rate ...
`away_pen_share`	0.6084	3.094	-0.5330	composite feature
`home_platoon_whiff`	-0.2395	-1.044	+0.5264	platoon term, selected by the opposing starter's throwing hand
`home_off_ground_spread`	1.1711	2.365	+0.5243	offence group (dispersion across the lineup, not its mean): batter_ground_ball_rate_vs_slider, batter_ground_ball_rate_vs_fastball
`home_off_top3_contact`	1.3435	1.885	-0.4188	mean of the three best contact bats, a tail a club mean cannot express
`away_off_discipline`	0.5348	2.248	+0.3998	offence group: batter_whiff_rate_vs_fastball, batter_whiff_rate_vs_breaking, batter_whiff_rate_vs_offspeed, batter_strikeout_rate ...
`home_platoon_barrel`	0.3669	1.083	-0.3665	platoon term, selected by the opposing starter's throwing hand
`home_x_fastball`	-0.2385	-1.268	+0.3409	batter x pitcher interaction: pitcher_whiff_rate_fastball, batter_whiff_rate_vs_fastball

What drove the KPI forecast for `home_runs`

intercept 4.5238 + sum of contributions = 5.1721; prediction 5.1721; contributions outside the top listed 0.0145 runs

Feature	Raw	Standardised	Runs contributed	Which KPIs it is built from
`away_pen_share`	0.6084	3.094	+0.6775	composite feature
`home_off_ground_spread`	1.1711	2.365	+0.5782	offence group (dispersion across the lineup, not its mean): batter_ground_ball_rate_vs_slider, batter_ground_ball_rate_vs_fastball
`home_pen_share`	0.5399	2.299	+0.5465	composite feature
`home_off_contact_spread`	0.9836	1.624	+0.4690	offence group (dispersion across the lineup, not its mean): batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`home_off_power_spread`	1.1914	1.740	-0.4472	offence group (dispersion across the lineup, not its mean): batter_home_run_rate, batter_bat_speed
`home_off_worst3_discipline`	-1.1139	-2.305	-0.4261	mean of the three weakest plate approaches
`home_platoon_whiff`	-0.2395	-1.044	-0.3999	platoon term, selected by the opposing starter's throwing hand
`home_off_top3_contact`	1.3435	1.885	-0.3397	mean of the three best contact bats, a tail a club mean cannot express
`away_platoon_whiff`	0.5920	1.831	-0.3076	platoon term, selected by the opposing starter's throwing hand
`home_off_qualified_share`	0.8947	1.086	+0.2830	share of the lineup carrying a qualifying sample, not a skill measurement

These are Ridge coefficients on standardised inputs, so each contribution is that feature's own push on the run forecast, in runs. They sum exactly to the prediction minus the intercept, which makes the list an accounting of the forecast rather than a ranking of vaguely important things.

Per-player metric movement

CLE against Matt Festa (RHP)

The pipeline applied no club-wide rescale to CLE hitter lines.
Opposing staff exposure. These hitters face the starter for 13.8787 of 35.458 projected batters (39.1%) and the 3 projected relievers for the remaining 60.9%. Every metric below is the exposure-weighted blend of both matchups; the starter-only figure is shown beside it. Real 2026 game logs put a starter at 21.75 batters faced on average (median 23, p10-p90 17-27) and a whole staff at 37.9: this starter is OUTSIDE that band and this staff is inside its. The per-player columns are the measured alternative: each subject moved by his own KPI pairing. Exposure for a hitter is his projected plate appearances; for the starter it is batters faced.

Player	Pos	Metric	Published → KPI	Δ	League	Own rate	Opponent rate	Log5	Own skill	Opponent	Remainder
Steven Kwan	CF	Hitswhole staff	0.955 → 0.929	-0.025	0.2169	0.2296n=620 Good	0.2048n=369 Good	0.2170	+0.056	-0.053	-0.002 ok
		split	starter x0.9451 at 0.391 share (log5 0.2170 vs his 0.2048) + bullpen x0.9916 at 0.609 share (log5 0.2277 vs pen 0.2150) = x0.9734 \| starter alone would have been -0.052
		Strikeoutswhole staff	0.477 → 0.499	+0.022	0.2212	0.1013n=620 Very good	0.2265n=369 Average	0.1042	-0.529	+0.023	-0.011 ok
		split	starter x1.0277 at 0.391 share (log5 0.1042 vs his 0.2265) + bullpen x1.0573 at 0.609 share (log5 0.1072 vs pen 0.2321) = x1.0457 \| starter alone would have been +0.013
		Walkswhole staff	0.532 → 0.544	+0.012	0.0889	0.1280n=620 Very good	0.0912n=369 Average	0.1311	+0.172	+0.010	+0.004 ok
		split	starter x1.0242 at 0.391 share (log5 0.1311 vs his 0.0912) + bullpen x1.0224 at 0.609 share (log5 0.1309 vs pen 0.0910) = x1.0231 \| starter alone would have been +0.013
		Home runswhole staff	0.053 → 0.045	-0.008	0.0304	0.0125n=620 Poor	0.0215n=369 Very good	0.0088	-0.079	-0.039	+0.023 ok
		split	starter x0.7052 at 0.391 share (log5 0.0088 vs his 0.0215) + bullpen x0.9488 at 0.609 share (log5 0.0118 vs pen 0.0289) = x0.8535 \| starter alone would have been -0.016
Neutralising Steven Kwan to the qualified-population mean moves the club's KPI run forecast by 0.4174 runs, and the weighted stage by 0.1419. Interactions make these non-additive across players, so they do not sum to the published movement.
José Ramírez	3B	Hitswhole staff	0.870 → 0.846	-0.024	0.2169	0.2095n=531 Below average	0.2048n=369 Good	0.1977	-0.033	-0.054	+0.001 ok
		split	starter x0.9437 at 0.391 share (log5 0.1977 vs his 0.2048) + bullpen x0.9914 at 0.609 share (log5 0.2077 vs pen 0.2150) = x0.9727 \| starter alone would have been -0.049
		Strikeoutswhole staff	0.627 → 0.655	+0.028	0.2212	0.1376n=531 Very good	0.2265n=369 Average	0.1413	-0.377	+0.024	-0.007 ok
		split	starter x1.0265 at 0.391 share (log5 0.1413 vs his 0.2265) + bullpen x1.0549 at 0.609 share (log5 0.1452 vs pen 0.2321) = x1.0438 \| starter alone would have been +0.017
		Walkswhole staff	0.545 → 0.558	+0.013	0.0889	0.1086n=531 Good	0.0912n=369 Average	0.1113	+0.089	+0.010	+0.002 ok
		split	starter x1.0248 at 0.391 share (log5 0.1113 vs his 0.0912) + bullpen x1.0229 at 0.609 share (log5 0.1111 vs pen 0.0910) = x1.0236 \| starter alone would have been +0.013
		Home runswhole staff	0.115 → 0.098	-0.017	0.0304	0.0252n=531 Below average	0.0215n=369 Very good	0.0178	-0.023	-0.040	+0.007 ok
		split	starter x0.7079 at 0.391 share (log5 0.0178 vs his 0.0215) + bullpen x0.9495 at 0.609 share (log5 0.0239 vs pen 0.0289) = x0.8549 \| starter alone would have been -0.034
Neutralising José Ramírez to the qualified-population mean moves the club's KPI run forecast by 0.0749 runs, and the weighted stage by 0.0255. Interactions make these non-additive across players, so they do not sum to the published movement.
Chase DeLauter	RF	Hitswhole staff	0.964 → 0.938	-0.025	0.2169	0.2364n=543 Very good	0.2048n=369 Good	0.2236	+0.083	-0.051	-0.003 ok
		split	starter x0.9455 at 0.391 share (log5 0.2236 vs his 0.2048) + bullpen x0.9917 at 0.609 share (log5 0.2345 vs pen 0.2150) = x0.9736 \| starter alone would have been -0.052
		Strikeoutswhole staff	0.709 → 0.739	+0.030	0.2212	0.1534n=543 Good	0.2265n=369 Average	0.1574	-0.287	+0.022	-0.005 ok
		split	starter x1.026 at 0.391 share (log5 0.1574 vs his 0.2265) + bullpen x1.0538 at 0.609 share (log5 0.1617 vs pen 0.2321) = x1.0429 \| starter alone would have been +0.018
		Walkswhole staff	0.418 → 0.428	+0.010	0.0889	0.1051n=543 Good	0.0912n=369 Average	0.1077	+0.069	+0.010	+0.002 ok
		split	starter x1.0249 at 0.391 share (log5 0.1077 vs his 0.0912) + bullpen x1.023 at 0.609 share (log5 0.1075 vs pen 0.0910) = x1.0237 \| starter alone would have been +0.010
		Home runswhole staff	0.125 → 0.107	-0.018	0.0304	0.0298n=543 Average	0.0215n=369 Very good	0.0211	-0.003	-0.037	+0.001 ok
		split	starter x0.7089 at 0.391 share (log5 0.0211 vs his 0.0215) + bullpen x0.9497 at 0.609 share (log5 0.0283 vs pen 0.0289) = x0.8555 \| starter alone would have been -0.036
Neutralising Chase DeLauter to the qualified-population mean moves the club's KPI run forecast by 0.2651 runs, and the weighted stage by 0.0901. Interactions make these non-additive across players, so they do not sum to the published movement.
Jo Adell	DH	Hitswhole staff	0.871 → 0.847	-0.024	0.2169	0.2185n=656 Average	0.2048n=369 Good	0.2063	+0.006	-0.050	-0.000 ok
		split	starter x0.9443 at 0.391 share (log5 0.2063 vs his 0.2048) + bullpen x0.9915 at 0.609 share (log5 0.2166 vs pen 0.2150) = x0.973 \| starter alone would have been -0.049
		Strikeoutswhole staff	1.044 → 1.084	+0.040	0.2212	0.2367n=656 Average	0.2265n=369 Average	0.2422	+0.063	+0.022	+0.001 ok
		split	starter x1.0234 at 0.391 share (log5 0.2422 vs his 0.2265) + bullpen x1.0483 at 0.609 share (log5 0.2481 vs pen 0.2321) = x1.0386 \| starter alone would have been +0.024
		Walkswhole staff	0.281 → 0.288	+0.007	0.0889	0.0575n=656 Poor	0.0912n=369 Average	0.0590	-0.129	+0.009	-0.003 ok
		split	starter x1.0263 at 0.391 share (log5 0.0590 vs his 0.0912) + bullpen x1.0243 at 0.609 share (log5 0.0589 vs pen 0.0910) = x1.0251 \| starter alone would have been +0.007
		Home runswhole staff	0.149 → 0.127	-0.021	0.0304	0.0337n=656 Average	0.0215n=369 Very good	0.0239	+0.014	-0.036	-0.004 ok
		split	starter x0.7097 at 0.391 share (log5 0.0239 vs his 0.0215) + bullpen x0.9499 at 0.609 share (log5 0.0320 vs pen 0.0289) = x0.8559 \| starter alone would have been -0.043
Neutralising Jo Adell to the qualified-population mean moves the club's KPI run forecast by -0.0412 runs, and the weighted stage by -0.0140. Interactions make these non-additive across players, so they do not sum to the published movement.
Nathaniel Lowe	1B	Hitswhole staff	0.832 → 0.809	-0.022	0.2169	0.2187n=417 Average	0.2048n=369 Good	0.2065	+0.007	-0.047	-0.000 ok
		split	starter x0.9443 at 0.391 share (log5 0.2065 vs his 0.2048) + bullpen x0.9915 at 0.609 share (log5 0.2168 vs pen 0.2150) = x0.973 \| starter alone would have been -0.046
		Strikeoutswhole staff	0.992 → 1.030	+0.038	0.2212	0.2293n=417 Average	0.2265n=369 Average	0.2347	+0.031	+0.021	+0.001 ok
		split	starter x1.0236 at 0.391 share (log5 0.2347 vs his 0.2265) + bullpen x1.0487 at 0.609 share (log5 0.2404 vs pen 0.2321) = x1.0389 \| starter alone would have been +0.023
		Walkswhole staff	0.332 → 0.340	+0.008	0.0889	0.0906n=417 Average	0.0912n=369 Average	0.0929	+0.007	+0.009	+0.000 ok
		split	starter x1.0253 at 0.391 share (log5 0.0929 vs his 0.0912) + bullpen x1.0234 at 0.609 share (log5 0.0928 vs pen 0.0910) = x1.0241 \| starter alone would have been +0.008
		Home runswhole staff	0.138 → 0.118	-0.020	0.0304	0.0353n=417 Good	0.0215n=369 Very good	0.0250	+0.019	-0.034	-0.005 ok
		split	starter x0.7101 at 0.391 share (log5 0.0250 vs his 0.0215) + bullpen x0.95 at 0.609 share (log5 0.0335 vs pen 0.0289) = x0.8561 \| starter alone would have been -0.040
Neutralising Nathaniel Lowe to the qualified-population mean moves the club's KPI run forecast by 0.0927 runs, and the weighted stage by 0.0315. Interactions make these non-additive across players, so they do not sum to the published movement.
Angel Martínez	LF	Hitswhole staff	0.869 → 0.846	-0.023	0.2169	0.2293n=408 Good	0.2048n=369 Good	0.2167	+0.048	-0.047	-0.002 ok
		split	starter x0.9451 at 0.391 share (log5 0.2167 vs his 0.2048) + bullpen x0.9916 at 0.609 share (log5 0.2273 vs pen 0.2150) = x0.9734 \| starter alone would have been -0.048
		Strikeoutswhole staff	0.941 → 0.978	+0.037	0.2212	0.2185n=408 Average	0.2265n=369 Average	0.2237	-0.011	+0.021	-0.000 ok
		split	starter x1.024 at 0.391 share (log5 0.2237 vs his 0.2265) + bullpen x1.0495 at 0.609 share (log5 0.2293 vs pen 0.2321) = x1.0395 \| starter alone would have been +0.023
		Walkswhole staff	0.205 → 0.210	+0.005	0.0889	0.0429n=408 Poor	0.0912n=369 Average	0.0440	-0.179	+0.009	-0.004 ok
		split	starter x1.0267 at 0.391 share (log5 0.0440 vs his 0.0912) + bullpen x1.0247 at 0.609 share (log5 0.0439 vs pen 0.0910) = x1.0255 \| starter alone would have been +0.005
		Home runswhole staff	0.164 → 0.140	-0.024	0.0304	0.0387n=408 Good	0.0215n=369 Very good	0.0275	+0.032	-0.034	-0.009 ok
		split	starter x0.7108 at 0.391 share (log5 0.0275 vs his 0.0215) + bullpen x0.9501 at 0.609 share (log5 0.0368 vs pen 0.0289) = x0.8564 \| starter alone would have been -0.047
Neutralising Angel Martínez to the qualified-population mean moves the club's KPI run forecast by 0.0306 runs, and the weighted stage by 0.0104. Interactions make these non-additive across players, so they do not sum to the published movement.
Travis Bazzana	2B	Hitswhole staff	0.788 → 0.767	-0.021	0.2169	0.2129n=506 Average	0.2048n=369 Good	0.2010	-0.015	-0.046	+0.001 ok
		split	starter x0.944 at 0.391 share (log5 0.2010 vs his 0.2048) + bullpen x0.9914 at 0.609 share (log5 0.2111 vs pen 0.2150) = x0.9728 \| starter alone would have been -0.044
		Strikeoutswhole staff	0.873 → 0.908	+0.035	0.2212	0.2087n=506 Average	0.2265n=369 Average	0.2138	-0.048	+0.020	-0.001 ok
		split	starter x1.0243 at 0.391 share (log5 0.2138 vs his 0.2265) + bullpen x1.0501 at 0.609 share (log5 0.2192 vs pen 0.2321) = x1.04 \| starter alone would have been +0.021
		Walkswhole staff	0.388 → 0.397	+0.009	0.0889	0.1033n=506 Good	0.0912n=369 Average	0.1059	+0.055	+0.009	+0.001 ok
		split	starter x1.0249 at 0.391 share (log5 0.1059 vs his 0.0912) + bullpen x1.0231 at 0.609 share (log5 0.1057 vs pen 0.0910) = x1.0238 \| starter alone would have been +0.010
		Home runswhole staff	0.113 → 0.097	-0.016	0.0304	0.0299n=506 Average	0.0215n=369 Very good	0.0212	-0.002	-0.034	+0.001 ok
		split	starter x0.7089 at 0.391 share (log5 0.0212 vs his 0.0215) + bullpen x0.9497 at 0.609 share (log5 0.0284 vs pen 0.0289) = x0.8555 \| starter alone would have been -0.033
Neutralising Travis Bazzana to the qualified-population mean moves the club's KPI run forecast by -0.0246 runs, and the weighted stage by -0.0084. Interactions make these non-additive across players, so they do not sum to the published movement.
Patrick Bailey	C	Hitswhole staff	0.735 → 0.715	-0.020	0.2169	0.2062n=333 Below average	0.2048n=369 Good	0.1946	-0.039	-0.045	+0.002 ok
		split	starter x0.9435 at 0.391 share (log5 0.1946 vs his 0.2048) + bullpen x0.9913 at 0.609 share (log5 0.2045 vs pen 0.2150) = x0.9726 \| starter alone would have been -0.042
		Strikeoutswhole staff	1.008 → 1.045	+0.037	0.2212	0.2666n=333 Below average	0.2265n=369 Average	0.2726	+0.168	+0.019	+0.003 ok
		split	starter x1.0225 at 0.391 share (log5 0.2726 vs his 0.2265) + bullpen x1.0463 at 0.609 share (log5 0.2790 vs pen 0.2321) = x1.037 \| starter alone would have been +0.023
		Walkswhole staff	0.336 → 0.344	+0.008	0.0889	0.0809n=333 Average	0.0912n=369 Average	0.0830	-0.029	+0.008	-0.001 ok
		split	starter x1.0256 at 0.391 share (log5 0.0830 vs his 0.0912) + bullpen x1.0237 at 0.609 share (log5 0.0829 vs pen 0.0910) = x1.0244 \| starter alone would have been +0.009
		Home runswhole staff	0.119 → 0.102	-0.017	0.0304	0.0302n=333 Average	0.0215n=369 Very good	0.0214	-0.001	-0.033	+0.000 ok
		split	starter x0.709 at 0.391 share (log5 0.0214 vs his 0.0215) + bullpen x0.9497 at 0.609 share (log5 0.0287 vs pen 0.0289) = x0.8555 \| starter alone would have been -0.035
Neutralising Patrick Bailey to the qualified-population mean moves the club's KPI run forecast by -0.0759 runs, and the weighted stage by -0.0258. Interactions make these non-additive across players, so they do not sum to the published movement.
Brayan Rocchio	SS	Hitswhole staff	0.745 → 0.725	-0.020	0.2169	0.2238n=576 Average	0.2048n=369 Good	0.2115	+0.024	-0.042	-0.001 ok
		split	starter x0.9447 at 0.391 share (log5 0.2115 vs his 0.2048) + bullpen x0.9915 at 0.609 share (log5 0.2219 vs pen 0.2150) = x0.9732 \| starter alone would have been -0.041
		Strikeoutswhole staff	0.552 → 0.576	+0.024	0.2212	0.1501n=576 Good	0.2265n=369 Average	0.1540	-0.246	+0.018	-0.005 ok
		split	starter x1.0261 at 0.391 share (log5 0.1540 vs his 0.2265) + bullpen x1.054 at 0.609 share (log5 0.1582 vs pen 0.2321) = x1.0431 \| starter alone would have been +0.014
		Walkswhole staff	0.275 → 0.282	+0.007	0.0889	0.0699n=576 Below average	0.0912n=369 Average	0.0717	-0.066	+0.008	-0.002 ok
		split	starter x1.0259 at 0.391 share (log5 0.0717 vs his 0.0912) + bullpen x1.024 at 0.609 share (log5 0.0716 vs pen 0.0910) = x1.0247 \| starter alone would have been +0.007
		Home runswhole staff	0.093 → 0.079	-0.013	0.0304	0.0250n=576 Below average	0.0215n=369 Very good	0.0177	-0.019	-0.031	+0.005 ok
		split	starter x0.7079 at 0.391 share (log5 0.0177 vs his 0.0215) + bullpen x0.9494 at 0.609 share (log5 0.0238 vs pen 0.0289) = x0.8549 \| starter alone would have been -0.027
Neutralising Brayan Rocchio to the qualified-population mean moves the club's KPI run forecast by -0.1313 runs, and the weighted stage by -0.0446. Interactions make these non-additive across players, so they do not sum to the published movement.
STARTER Daniel Espino17.1684 BF	Starter	Hits	3.623 → 3.620	-0.003	0.2169	0.2169n=0 averagebelow floor	0.21678/9 qualified lineup mean	0.2167	0	-0.003	0 ok
		Strikeouts	3.940 → 4.083	+0.143	0.2212	0.2212n=0 averagebelow floor	0.22938/9 qualified lineup mean	0.2293	0	+0.138	0 ok
		Walks	1.731 → 1.913	+0.181	0.0889	0.0889n=0 averagebelow floor	0.09828/9 qualified lineup mean	0.0982	0	+0.160	0 ok
		Home runs	0.637 → 0.720	+0.083	0.0304	0.0304n=0 averagebelow floor	0.03438/9 qualified lineup mean	0.0343	0	+0.068	0 ok
RELIEVER Tim Herrin3.9958 BF	Middle Reliever	Hits	0.868 → 0.867	-0.001	0.2169	0.2029n=237 Good	0.21678/9 qualified lineup mean	0.2028	-0.056	-0.001	0 ok
		Strikeouts	0.972 → 1.007	+0.035	0.2212	0.2262n=237 Average	0.22938/9 qualified lineup mean	0.2343	+0.020	+0.032	+0.001 ok
		Walks	0.491 → 0.541	+0.050	0.0889	0.1065n=237 Below average	0.09828/9 qualified lineup mean	0.1174	+0.070	+0.037	+0.006 ok
		Home runs	0.070 → 0.079	+0.009	0.0304	0.0272n=237 Average	0.03438/9 qualified lineup mean	0.0307	-0.013	+0.016	-0.002 ok
RELIEVER Erik Sabrowski3.6485 BF	Middle Reliever (High Leverage)	Hits	0.525 → 0.524	-0.001	0.2169	0.1819n=194 Very good	0.21678/9 qualified lineup mean	0.1817	-0.128	-0.001	+0.000 ok
		Strikeouts	1.568 → 1.617	+0.049	0.2212	0.3346n=194 Very good	0.22938/9 qualified lineup mean	0.3449	+0.414	+0.029	+0.008 ok
		Walks	0.625 → 0.688	+0.062	0.0889	0.1280n=194 Poor	0.09828/9 qualified lineup mean	0.1408	+0.143	+0.034	+0.013 ok
		Home runs	0.046 → 0.052	+0.006	0.0304	0.0259n=194 Good	0.03438/9 qualified lineup mean	0.0292	-0.017	+0.014	-0.002 ok
RELIEVER Cade Smith3.53 BF	Reliever	Hits	0.674 → 0.673	-0.001	0.2169	0.1909n=281 Very good	0.21678/9 qualified lineup mean	0.1907	-0.092	-0.001	+0.000 ok
		Strikeouts	1.170 → 1.207	+0.036	0.2212	0.3315n=281 Very good	0.22938/9 qualified lineup mean	0.3418	+0.389	+0.028	+0.008 ok
		Walks	0.274 → 0.303	+0.029	0.0889	0.0776n=281 Good	0.09828/9 qualified lineup mean	0.0858	-0.040	+0.033	-0.004 ok
		Home runs	0.090 → 0.102	+0.012	0.0304	0.0255n=281 Good	0.03438/9 qualified lineup mean	0.0289	-0.017	+0.014	-0.002 ok
RELIEVER Hunter Gaddis3.4396 BF	Reliever	Hits	0.736 → 0.735	-0.001	0.2169	0.2138n=252 Average	0.21678/9 qualified lineup mean	0.2137	-0.010	-0.001	0 ok
		Strikeouts	0.763 → 0.791	+0.028	0.2212	0.2220n=252 Average	0.22938/9 qualified lineup mean	0.2300	+0.003	+0.028	+0.000 ok
		Walks	0.302 → 0.334	+0.032	0.0889	0.0879n=252 Average	0.09828/9 qualified lineup mean	0.0971	-0.003	+0.032	-0.000 ok
		Home runs	0.101 → 0.114	+0.013	0.0304	0.0294n=252 Average	0.03438/9 qualified lineup mean	0.0333	-0.003	+0.014	-0.000 ok
RELIEVER Shawn Armstrong3.2931 BF	Middle Reliever (High Leverage)	Hits	0.684 → 0.683	-0.001	0.2169	0.2076n=148 Average	0.21678/9 qualified lineup mean	0.2074	-0.030	-0.001	0 ok
		Strikeouts	0.881 → 0.913	+0.032	0.2212	0.2180n=148 Average	0.22938/9 qualified lineup mean	0.2259	-0.011	+0.026	-0.000 ok
		Walks	0.472 → 0.521	+0.049	0.0889	0.1048n=148 Below average	0.09828/9 qualified lineup mean	0.1155	+0.052	+0.031	+0.005 ok
		Home runs	0.066 → 0.075	+0.009	0.0304	0.0279n=148 Average	0.03438/9 qualified lineup mean	0.0315	-0.008	+0.013	-0.001 ok
RELIEVER Joey Cantillo1.3855 BF	Reliever	Hits	0.293 → 0.293	-0.000	0.2169	0.2118n=687 Average	0.21678/9 qualified lineup mean	0.2116	-0.007	-0.000	0 ok
		Strikeouts	0.341 → 0.353	+0.012	0.2212	0.2460n=687 Good	0.22938/9 qualified lineup mean	0.2547	+0.034	+0.011	+0.001 ok
		Walks	0.156 → 0.172	+0.016	0.0889	0.1125n=687 Below average	0.09828/9 qualified lineup mean	0.1239	+0.033	+0.013	+0.003 ok
		Home runs	0.037 → 0.042	+0.005	0.0304	0.0270n=687 Good	0.03438/9 qualified lineup mean	0.0305	-0.005	+0.005	-0.001 ok
RELIEVER Matt Festa0.7911 BF	Reliever	Hits	0.174 → 0.174	-0.000	0.2169	0.2199n=278 Average	0.21678/9 qualified lineup mean	0.2197	+0.002	-0.000	0 ok
		Strikeouts	0.183 → 0.189	+0.007	0.2212	0.2310n=278 Average	0.22938/9 qualified lineup mean	0.2393	+0.008	+0.006	+0.000 ok
		Walks	0.064 → 0.070	+0.007	0.0889	0.0804n=278 Average	0.09828/9 qualified lineup mean	0.0889	-0.007	+0.007	-0.001 ok
		Home runs	0.021 → 0.024	+0.003	0.0304	0.0270n=278 Good	0.03438/9 qualified lineup mean	0.0305	-0.003	+0.003	-0.000 ok

CHW against David Sandlin (RHP)

The pipeline applied one club-wide ×1.1114 to every CHW hitter line, identical for the leadoff hitter and the ninth bat.
Opposing staff exposure. These hitters face the starter for 17.1684 of 37.248 projected batters (46.1%) and the 3 projected relievers for the remaining 53.9%. Every metric below is the exposure-weighted blend of both matchups; the starter-only figure is shown beside it. Real 2026 game logs put a starter at 21.75 batters faced on average (median 23, p10-p90 17-27) and a whole staff at 37.9: this starter is inside that band and this staff is inside its. The per-player columns are the measured alternative: each subject moved by his own KPI pairing. Exposure for a hitter is his projected plate appearances; for the starter it is batters faced.

Player	Pos	Metric	Published → KPI	Δ	League	Own rate	Opponent rate	Log5	Own skill	Opponent	Remainder
Sam Antonacci	DH	Hitswhole staff	0.974 → 0.936	-0.038	0.2169	0.2300n=553 Good	0.2169n=0 averagebelow floor	0.2300	+0.056	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2300 vs his 0.2169) + bullpen x0.9276 at 0.539 share (log5 0.2134 vs pen 0.2009) = x0.961 \| starter alone would have been 0
		Strikeoutswhole staff	0.751 → 0.838	+0.086	0.2212	0.1525n=553 Good	0.2212n=0 averagebelow floor	0.1525	-0.293	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.1525 vs his 0.2212) + bullpen x1.2135 at 0.539 share (log5 0.1851 vs pen 0.2639) = x1.1151 \| starter alone would have been 0
		Walkswhole staff	0.411 → 0.442	+0.031	0.0889	0.0797n=553 Average	0.0889n=0 averagebelow floor	0.0797	-0.039	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0797 vs his 0.0889) + bullpen x1.14 at 0.539 share (log5 0.0909 vs pen 0.1012) = x1.0755 \| starter alone would have been 0
		Home runswhole staff	0.091 → 0.086	-0.005	0.0304	0.0196n=553 Poor	0.0304n=0 averagebelow floor	0.0196	-0.046	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0196 vs his 0.0304) + bullpen x0.8917 at 0.539 share (log5 0.0175 vs pen 0.0271) = x0.9416 \| starter alone would have been 0
Neutralising Sam Antonacci to the qualified-population mean moves the club's KPI run forecast by 0.1818 runs, and the weighted stage by 0.0691. Interactions make these non-additive across players, so they do not sum to the published movement.
Miguel Vargas	3B	Hitswhole staff	0.868 → 0.834	-0.035	0.2169	0.2136n=695 Average	0.2169n=0 averagebelow floor	0.2136	-0.014	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2136 vs his 0.2169) + bullpen x0.9262 at 0.539 share (log5 0.1978 vs pen 0.2009) = x0.9602 \| starter alone would have been 0
		Strikeoutswhole staff	0.881 → 0.978	+0.097	0.2212	0.1826n=695 Good	0.2212n=0 averagebelow floor	0.1826	-0.168	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.1826 vs his 0.2212) + bullpen x1.2044 at 0.539 share (log5 0.2199 vs pen 0.2639) = x1.1102 \| starter alone would have been 0
		Walkswhole staff	0.636 → 0.681	+0.045	0.0889	0.1346n=695 Very good	0.0889n=0 averagebelow floor	0.1346	+0.199	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.1346 vs his 0.0889) + bullpen x1.1306 at 0.539 share (log5 0.1522 vs pen 0.1012) = x1.0704 \| starter alone would have been 0
		Home runswhole staff	0.187 → 0.176	-0.011	0.0304	0.0428n=695 Very good	0.0304n=0 averagebelow floor	0.0428	+0.054	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0428 vs his 0.0304) + bullpen x0.894 at 0.539 share (log5 0.0382 vs pen 0.0271) = x0.9429 \| starter alone would have been 0
Neutralising Miguel Vargas to the qualified-population mean moves the club's KPI run forecast by 0.5581 runs, and the weighted stage by 0.2121. Interactions make these non-additive across players, so they do not sum to the published movement.
Munetaka Murakami	1B	Hitswhole staff	0.811 → 0.778	-0.033	0.2169	0.1915n=544 Poor	0.2169n=0 averagebelow floor	0.1915	-0.115	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.1915 vs his 0.2169) + bullpen x0.9242 at 0.539 share (log5 0.1770 vs pen 0.2009) = x0.9591 \| starter alone would have been 0
		Strikeoutswhole staff	1.652 → 1.792	+0.140	0.2212	0.3438n=544 Poor	0.2212n=0 averagebelow floor	0.3438	+0.554	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.3438 vs his 0.2212) + bullpen x1.1577 at 0.539 share (log5 0.3980 vs pen 0.2639) = x1.085 \| starter alone would have been 0
		Walkswhole staff	0.753 → 0.805	+0.052	0.0889	0.1532n=544 Very good	0.0889n=0 averagebelow floor	0.1532	+0.291	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.1532 vs his 0.0889) + bullpen x1.1274 at 0.539 share (log5 0.1727 vs pen 0.1012) = x1.0687 \| starter alone would have been 0
		Home runswhole staff	0.245 → 0.231	-0.014	0.0304	0.0532n=544 Very good	0.0304n=0 averagebelow floor	0.0532	+0.103	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0532 vs his 0.0304) + bullpen x0.895 at 0.539 share (log5 0.0476 vs pen 0.0271) = x0.9434 \| starter alone would have been 0
Neutralising Munetaka Murakami to the qualified-population mean moves the club's KPI run forecast by 0.0164 runs, and the weighted stage by 0.0062. Interactions make these non-additive across players, so they do not sum to the published movement.
Kyle Teel	C	Hitswhole staff	0.833 → 0.800	-0.033	0.2169	0.2169n=0 averagebelow floor	0.2169n=0 averagebelow floor	0.2169	0	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2169 vs his 0.2169) + bullpen x0.9264 at 0.539 share (log5 0.2009 vs pen 0.2009) = x0.9603 \| starter alone would have been 0
		Strikeoutswhole staff	1.496 → 1.652	+0.156	0.2212	0.2212n=0 averagebelow floor	0.2212n=0 averagebelow floor	0.2212	0	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2212 vs his 0.2212) + bullpen x1.1929 at 0.539 share (log5 0.2639 vs pen 0.2639) = x1.104 \| starter alone would have been 0
		Walkswhole staff	0.488 → 0.525	+0.036	0.0889	0.0889n=0 averagebelow floor	0.0889n=0 averagebelow floor	0.0889	0	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0889 vs his 0.0889) + bullpen x1.1385 at 0.539 share (log5 0.1012 vs pen 0.1012) = x1.0747 \| starter alone would have been 0
		Home runswhole staff	0.136 → 0.128	-0.008	0.0304	0.0304n=0 averagebelow floor	0.0304n=0 averagebelow floor	0.0304	0	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0304 vs his 0.0304) + bullpen x0.8928 at 0.539 share (log5 0.0271 vs pen 0.0271) = x0.9422 \| starter alone would have been 0
Neutralising Kyle Teel to the qualified-population mean moves the club's KPI run forecast by 0.0777 runs, and the weighted stage by 0.0295. Interactions make these non-additive across players, so they do not sum to the published movement.
Chase Meidroth	2B	Hitswhole staff	0.936 → 0.899	-0.036	0.2169	0.2330n=635 Good	0.2169n=0 averagebelow floor	0.2330	+0.065	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2330 vs his 0.2169) + bullpen x0.9278 at 0.539 share (log5 0.2161 vs pen 0.2009) = x0.9611 \| starter alone would have been 0
		Strikeoutswhole staff	1.042 → 1.151	+0.109	0.2212	0.2176n=635 Average	0.2212n=0 averagebelow floor	0.2176	-0.015	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2176 vs his 0.2212) + bullpen x1.1939 at 0.539 share (log5 0.2598 vs pen 0.2639) = x1.1045 \| starter alone would have been 0
		Walkswhole staff	0.418 → 0.449	+0.031	0.0889	0.0936n=635 Average	0.0889n=0 averagebelow floor	0.0936	+0.019	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0936 vs his 0.0889) + bullpen x1.1376 at 0.539 share (log5 0.1065 vs pen 0.1012) = x1.0742 \| starter alone would have been 0
		Home runswhole staff	0.119 → 0.112	-0.007	0.0304	0.0267n=635 Average	0.0304n=0 averagebelow floor	0.0267	-0.015	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0267 vs his 0.0304) + bullpen x0.8924 at 0.539 share (log5 0.0238 vs pen 0.0271) = x0.942 \| starter alone would have been 0
Neutralising Chase Meidroth to the qualified-population mean moves the club's KPI run forecast by 0.0580 runs, and the weighted stage by 0.0220. Interactions make these non-additive across players, so they do not sum to the published movement.
Tristan Peters	CF	Hitswhole staff	0.863 → 0.829	-0.034	0.2169	0.2266n=484 Good	0.2169n=0 averagebelow floor	0.2266	+0.037	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2266 vs his 0.2169) + bullpen x0.9273 at 0.539 share (log5 0.2102 vs pen 0.2009) = x0.9608 \| starter alone would have been 0
		Strikeoutswhole staff	0.822 → 0.912	+0.090	0.2212	0.1835n=484 Good	0.2212n=0 averagebelow floor	0.1835	-0.142	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.1835 vs his 0.2212) + bullpen x1.2041 at 0.539 share (log5 0.2210 vs pen 0.2639) = x1.11 \| starter alone would have been 0
		Walkswhole staff	0.354 → 0.380	+0.027	0.0889	0.0789n=484 Average	0.0889n=0 averagebelow floor	0.0789	-0.038	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0789 vs his 0.0889) + bullpen x1.1402 at 0.539 share (log5 0.0900 vs pen 0.1012) = x1.0756 \| starter alone would have been 0
		Home runswhole staff	0.111 → 0.104	-0.006	0.0304	0.0268n=484 Average	0.0304n=0 averagebelow floor	0.0268	-0.014	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0268 vs his 0.0304) + bullpen x0.8924 at 0.539 share (log5 0.0239 vs pen 0.0271) = x0.942 \| starter alone would have been 0
Neutralising Tristan Peters to the qualified-population mean moves the club's KPI run forecast by 0.2924 runs, and the weighted stage by 0.1111. Interactions make these non-additive across players, so they do not sum to the published movement.
Braden Montgomery	RF	Hitswhole staff	0.799 → 0.767	-0.032	0.2169	0.2100n=362 Below average	0.2169n=0 averagebelow floor	0.2100	-0.027	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2100 vs his 0.2169) + bullpen x0.9258 at 0.539 share (log5 0.1944 vs pen 0.2009) = x0.96 \| starter alone would have been 0
		Strikeoutswhole staff	0.902 → 0.999	+0.097	0.2212	0.1991n=362 Average	0.2212n=0 averagebelow floor	0.1991	-0.086	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.1991 vs his 0.2212) + bullpen x1.1994 at 0.539 share (log5 0.2388 vs pen 0.2639) = x1.1075 \| starter alone would have been 0
		Walkswhole staff	0.444 → 0.477	+0.033	0.0889	0.1010n=362 Good	0.0889n=0 averagebelow floor	0.1010	+0.047	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.1010 vs his 0.0889) + bullpen x1.1364 at 0.539 share (log5 0.1148 vs pen 0.1012) = x1.0735 \| starter alone would have been 0
		Home runswhole staff	0.091 → 0.086	-0.005	0.0304	0.0224n=362 Below average	0.0304n=0 averagebelow floor	0.0224	-0.031	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0224 vs his 0.0304) + bullpen x0.892 at 0.539 share (log5 0.0200 vs pen 0.0271) = x0.9418 \| starter alone would have been 0
Neutralising Braden Montgomery to the qualified-population mean moves the club's KPI run forecast by 0.1599 runs, and the weighted stage by 0.0608. Interactions make these non-additive across players, so they do not sum to the published movement.
Randal Grichuk	LF	Hitswhole staff	0.868 → 0.834	-0.034	0.2169	0.2384n=279 Very good	0.2169n=0 averagebelow floor	0.2384	+0.077	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2384 vs his 0.2169) + bullpen x0.9283 at 0.539 share (log5 0.2213 vs pen 0.2009) = x0.9613 \| starter alone would have been 0
		Strikeoutswhole staff	0.955 → 1.054	+0.100	0.2212	0.2190n=279 Average	0.2212n=0 averagebelow floor	0.2190	-0.008	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.2190 vs his 0.2212) + bullpen x1.1935 at 0.539 share (log5 0.2614 vs pen 0.2639) = x1.1043 \| starter alone would have been 0
		Walkswhole staff	0.231 → 0.249	+0.018	0.0889	0.0542n=279 Poor	0.0889n=0 averagebelow floor	0.0542	-0.123	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0542 vs his 0.0889) + bullpen x1.1445 at 0.539 share (log5 0.0621 vs pen 0.1012) = x1.0779 \| starter alone would have been 0
		Home runswhole staff	0.168 → 0.159	-0.010	0.0304	0.0424n=279 Very good	0.0304n=0 averagebelow floor	0.0424	+0.043	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0424 vs his 0.0304) + bullpen x0.8939 at 0.539 share (log5 0.0379 vs pen 0.0271) = x0.9428 \| starter alone would have been 0
Neutralising Randal Grichuk to the qualified-population mean moves the club's KPI run forecast by 0.4031 runs, and the weighted stage by 0.1532. Interactions make these non-additive across players, so they do not sum to the published movement.
Colson Montgomery	SS	Hitswhole staff	0.697 → 0.669	-0.028	0.2169	0.1929n=630 Poor	0.2169n=0 averagebelow floor	0.1929	-0.086	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.1929 vs his 0.2169) + bullpen x0.9244 at 0.539 share (log5 0.1783 vs pen 0.2009) = x0.9592 \| starter alone would have been 0
		Strikeoutswhole staff	1.404 → 1.522	+0.119	0.2212	0.3459n=630 Poor	0.2212n=0 averagebelow floor	0.3459	+0.447	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.3459 vs his 0.2212) + bullpen x1.1571 at 0.539 share (log5 0.4003 vs pen 0.2639) = x1.0847 \| starter alone would have been 0
		Walkswhole staff	0.366 → 0.394	+0.028	0.0889	0.0836n=630 Average	0.0889n=0 averagebelow floor	0.0836	-0.019	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0836 vs his 0.0889) + bullpen x1.1394 at 0.539 share (log5 0.0952 vs pen 0.1012) = x1.0752 \| starter alone would have been 0
		Home runswhole staff	0.174 → 0.164	-0.010	0.0304	0.0436n=630 Very good	0.0304n=0 averagebelow floor	0.0436	+0.048	0	0 ok
		split	starter x1.0 at 0.461 share (log5 0.0436 vs his 0.0304) + bullpen x0.8941 at 0.539 share (log5 0.0390 vs pen 0.0271) = x0.9429 \| starter alone would have been 0
Neutralising Colson Montgomery to the qualified-population mean moves the club's KPI run forecast by 0.3800 runs, and the weighted stage by 0.1444. Interactions make these non-additive across players, so they do not sum to the published movement.
STARTER Sean Newcomb13.8787 BF	Starter	Hits	2.562 → 2.608	+0.046	0.2169	0.2048n=369 Good	0.22079/9 qualified lineup mean	0.2084	-0.168	+0.053	-0.002 ok
		Strikeouts	3.281 → 2.775	-0.506	0.2212	0.2265n=369 Average	0.18699/9 qualified lineup mean	0.1915	+0.073	-0.476	-0.009 ok
		Walks	1.294 → 1.289	-0.004	0.0889	0.0912n=369 Average	0.08869/9 qualified lineup mean	0.0909	+0.031	-0.004	-0.000 ok
		Home runs	0.306 → 0.289	-0.017	0.0304	0.0215n=369 Very good	0.02879/9 qualified lineup mean	0.0203	-0.123	-0.024	+0.007 ok
RELIEVER Grant Taylor4.3461 BF	Reliever	Hits	0.858 → 0.873	+0.015	0.2169	0.1974n=333 Good	0.22079/9 qualified lineup mean	0.2010	-0.085	+0.017	-0.001 ok
		Strikeouts	1.291 → 1.107	-0.184	0.2212	0.2970n=333 Very good	0.18699/9 qualified lineup mean	0.2548	+0.330	-0.149	-0.034 ok
		Walks	0.390 → 0.389	-0.001	0.0889	0.0897n=333 Average	0.08869/9 qualified lineup mean	0.0894	+0.003	-0.001	0 ok
		Home runs	0.115 → 0.108	-0.006	0.0304	0.0264n=333 Good	0.02879/9 qualified lineup mean	0.0249	-0.017	-0.007	+0.001 ok
RELIEVER Hagen Smith4.1778 BF	Long Reliever	Hits	0.469 → 0.477	+0.008	0.2169	0.2169n=0 averagebelow floor	0.22079/9 qualified lineup mean	0.2207	0	+0.016	0 ok
		Strikeouts	1.866 → 1.577	-0.290	0.2212	0.2212n=0 averagebelow floor	0.18699/9 qualified lineup mean	0.1869	0	-0.143	0 ok
		Walks	0.937 → 0.934	-0.003	0.0889	0.0889n=0 averagebelow floor	0.08869/9 qualified lineup mean	0.0886	0	-0.001	0 ok
		Home runs	0.000 → 0.000	0	0.0304	0.0304n=0 averagebelow floor	0.02879/9 qualified lineup mean	0.0287	0	-0.007	0 ok
RELIEVER Anthony Kay3.4829 BF	Reliever	Hits	0.781 → 0.795	+0.014	0.2169	0.2242n=686 Below average	0.22079/9 qualified lineup mean	0.2282	+0.026	+0.013	+0.000 ok
		Strikeouts	0.640 → 0.536	-0.103	0.2212	0.1837n=686 Below average	0.18699/9 qualified lineup mean	0.1540	-0.131	-0.120	+0.016 ok
		Walks	0.285 → 0.284	-0.001	0.0889	0.0818n=686 Average	0.08869/9 qualified lineup mean	0.0815	-0.025	-0.001	+0.000 ok
		Home runs	0.106 → 0.100	-0.006	0.0304	0.0305n=686 Average	0.02879/9 qualified lineup mean	0.0288	+0.001	-0.006	0 ok
RELIEVER Jordan Hicks3.4138 BF	Middle Reliever (High Leverage)	Hits	0.913 → 0.929	+0.016	0.2169	0.2054n=216 Good	0.22079/9 qualified lineup mean	0.2091	-0.039	+0.013	-0.001 ok
		Strikeouts	1.235 → 1.055	-0.180	0.2212	0.2760n=216 Good	0.18699/9 qualified lineup mean	0.2357	+0.187	-0.117	-0.020 ok
		Walks	0.599 → 0.597	-0.002	0.0889	0.1258n=216 Poor	0.08869/9 qualified lineup mean	0.1254	+0.126	-0.001	-0.000 ok
		Home runs	0.098 → 0.092	-0.005	0.0304	0.0265n=216 Good	0.02879/9 qualified lineup mean	0.0250	-0.013	-0.006	+0.001 ok
RELIEVER Bryan Hudson3.1398 BF	Middle Reliever (High Leverage)	Hits	1.024 → 1.042	+0.018	0.2169	0.2287n=282 Below average	0.22079/9 qualified lineup mean	0.2327	+0.037	+0.012	+0.001 ok
		Strikeouts	0.787 → 0.664	-0.123	0.2212	0.2119n=282 Average	0.18699/9 qualified lineup mean	0.1787	-0.029	-0.108	+0.004 ok
		Walks	0.229 → 0.228	-0.001	0.0889	0.0728n=282 Good	0.08869/9 qualified lineup mean	0.0726	-0.051	-0.001	+0.000 ok
		Home runs	0.073 → 0.069	-0.004	0.0304	0.0255n=282 Good	0.02879/9 qualified lineup mean	0.0241	-0.015	-0.005	+0.001 ok
RELIEVER Erick Fedde2.501 BF	Reliever	Hits	0.563 → 0.573	+0.010	0.2169	0.2253n=606 Below average	0.22079/9 qualified lineup mean	0.2292	+0.021	+0.010	+0.000 ok
		Strikeouts	0.425 → 0.356	-0.070	0.2212	0.1701n=606 Poor	0.18699/9 qualified lineup mean	0.1423	-0.128	-0.086	+0.016 ok
		Walks	0.216 → 0.215	-0.001	0.0889	0.0865n=606 Average	0.08869/9 qualified lineup mean	0.0862	-0.006	-0.001	0 ok
		Home runs	0.087 → 0.082	-0.005	0.0304	0.0347n=606 Below average	0.02879/9 qualified lineup mean	0.0328	+0.011	-0.004	-0.001 ok
RELIEVER David Sandlin0.5179 BF	Reliever	Hits	0.112 → 0.114	+0.002	0.2169	0.2164n=167 Average	0.22079/9 qualified lineup mean	0.2202	-0.000	+0.002	0 ok
		Strikeouts	0.121 → 0.102	-0.018	0.2212	0.2336n=167 Average	0.18699/9 qualified lineup mean	0.1978	+0.006	-0.018	-0.001 ok
		Walks	0.043 → 0.043	-0.000	0.0889	0.0832n=167 Average	0.08869/9 qualified lineup mean	0.0829	-0.003	-0.000	0 ok
		Home runs	0.017 → 0.016	-0.001	0.0304	0.0335n=167 Below average	0.02879/9 qualified lineup mean	0.0316	+0.002	-0.001	-0.000 ok

Published pitching, integer draw beside the continuous model line

Pitcher	Role	BF	IP	expected IP	H int/cont	SO int/cont	BB int/cont	ER int/cont	Rate source
Daniel Espino	Starter	17.2	4.0	3.938	3 / 3.623	3 / 3.940	3 / 1.731	2 / 2.157	all_players_stats_recent_season_blend
Tim Herrin	Middle Reliever	4.4	1.0	1.000	1 / 0.868	1 / 0.972	1 / 0.491	0 / 0.312	all_players_stats_recent_season_blend
Erik Sabrowski	Middle Reliever (High Leverage)	4.2	1.0	1.000	1 / 0.525	1 / 1.568	0 / 0.625	1 / 0.376	all_players_stats_recent_season_blend
Shawn Armstrong	Middle Reliever (High Leverage)	4.2	3.0	1.000	2 / 0.684	3 / 0.881	0 / 0.472	0 / 0.402	all_players_stats_recent_season_blend
Sean Newcomb	Starter	13.9	3.1	3.341	2 / 2.562	2 / 3.281	2 / 1.294	1 / 1.187	all_players_stats_recent_season_blend
Hagen Smith	Long Reliever	5.4	1.1	1.337	1 / 0.469	0 / 1.866	1 / 0.937	0 / 0.279	all_players_stats_recent_season_blend
Jordan Hicks	Middle Reliever (High Leverage)	4.5	1.0	1.003	1 / 0.913	1 / 1.235	1 / 0.599	2 / 0.488	all_players_stats_recent_season_blend
Bryan Hudson	Middle Reliever (High Leverage)	4.3	3.1	1.003	4 / 1.024	3 / 0.787	1 / 0.229	2 / 0.317	all_players_stats_recent_season_blend

Audit: NO DEFECT FOUND

Invariant	Verdict	Evidence
blend arithmetic away_runs	PASS	5.0000 + 0.34 x (5.0015 - 5.0000) = 5.0005, recorded 5.0005
blend arithmetic home_runs	PASS	4.0000 + 0.38 x (5.1721 - 4.0000) = 4.4454, recorded 4.4454
blend arithmetic home_win	PASS	0.4514 + 0.00 x (0.5275 - 0.4514) = 0.4514, recorded 0.4514
published rounding faithful to the blend	PASS	blended 5.0005-4.4454 rounds to 5-4, layer published 5-4
published probability on the score's side	PASS	5-4 with p(home) 0.451379
Log5 additivity136 checks	PASS	136 of 136 pass
away club fields exactly one starter	PASS	1 starters in a staff of 4
away staff totals a regulation workload	PASS	27 outs recorded, 27 expected
away starter carries a continuous line	PASS	expected_IP 3.938
away starter batters faced inside the real band	PASS	Daniel Espino projected 17.2 batters faced; real starters run p01-p99 4-30, median 23, mean 21.75
away reliever batters faced inside the real band3 checks	PASS	3 of 3 pass
home club fields exactly one starter	PASS	1 starters in a staff of 4
home staff totals a regulation workload	PASS	27 outs recorded, 27 expected
home starter carries a continuous line	PASS	expected_IP 3.341
home starter batters faced inside the real band	PASS	Sean Newcomb projected 13.9 batters faced; real starters run p01-p99 4-30, median 23, mean 21.75
home reliever batters faced inside the real band3 checks	PASS	3 of 3 pass
away hitters satisfy every coherence rule	PASS	9 hitters checked; H<=AB, HR<=H, 2B+3B+HR<=H, SO<=AB, continuous line present
home hitters satisfy every coherence rule	PASS	9 hitters checked; H<=AB, HR<=H, 2B+3B+HR<=H, SO<=AB, continuous line present

20261007_LAD@ATL — published 6-4, 156 checks, all pass

Artifact. 20261007_LAD@ATL__20261008T024931Z.json, phase post_start_rerun, generated 2026-10-08T02:49:31, first pitch 2026-10-07T22:00:00.
Published 6 - 4 with p(home) 0.492823. Simulation draw chosen on not recorded; ? of ? draws realized the target exactly.

Stage chain: every stage that touched the score

Stage	Score after	Δ away	Δ home	Cum. away	Cum. home	Changed target	Own adjustment and description
1. `historical_ml`	4.695043839817618 - 3.9708099350753012	0	0	0	0	no	Base ML historical model prediction
2. `variance_applied`	5 - 4	+0.3050	+0.0292	+0.3050	+0.0292	yes	Deterministic baseline target (variance factor 1.0)
3. `enhanced_prediction`	5 - 4	0	0	+0.3050	+0.0292	no	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by matchup_recency_guardrail)
4. `kpi_layer`	5 - 4	0	0	+0.3050	+0.0292	no	away +0.0409; home +0.6826 Savant KPI blend applied to the model target
5. `embedded_savant_run_environment`	6 - 4	+1.0000	0	+1.3050	+0.0292	yes	away +0.5111; home -0.2760 Target score adjusted by temporal Savant lineup quality versus opposing starter quality
6. `fundamental_edge_target_review`	6 - 4	0	0	+1.3050	+0.0292	no	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
7. `capped_target`	6 - 4	0	0	+1.3050	+0.0292	no	Capped at max 12 realistic runs
8. `simulation_result`	6 - 4	-	-	-	-	no	Final inning-by-inning simulation result

KPI team layer: the blend, with its arithmetic

Target	Baseline	KPI only	Weight	Blended	Computation	Verdict
`away_runs`	4.0000	4.1204	0.34	4.0409	4.0000 + 0.34 × (4.1204 − 4.0000) = 4.0409	PASS
`home_runs`	3.0000	4.7962	0.38	3.6826	3.0000 + 0.38 × (4.7962 − 3.0000) = 3.6826	PASS
`home_win`	0.4916	0.6876	0.00	0.4916	0.4916 + 0.00 × (0.6876 − 0.4916) = 0.4916	PASS

Published by the layer: away 5, home 4, p(home) 0.491557, continuous away 4.0409 / home 3.6826

Inputs and gates. candidate 20261007T091423Z_47b9db81, value age 10 days, lineup coverage away 0.8830 / home 0.8947, movement cap 2.5 runs.
values_present ok values_fresh ok one_source_bundle ok lineup_coverage ok target_eligible ok movement_within_limit ok published_movement_within_limit ok

What drove the KPI forecast for `away_runs`

intercept 4.2358 + sum of contributions = 4.1204; prediction 4.1204; contributions outside the top listed -0.0462 runs

Feature	Raw	Standardised	Runs contributed	Which KPIs it is built from
`away_off_contact`	0.5130	1.979	+0.6800	offence group: batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`home_platoon_barrel`	0.6390	1.994	-0.6751	platoon term, selected by the opposing starter's throwing hand
`home_x_breaking`	0.4955	2.484	+0.5070	batter x pitcher interaction: pitcher_whiff_rate_breaking, batter_whiff_rate_vs_breaking
`home_off_top3_contact`	1.2701	1.656	-0.3680	mean of the three best contact bats, a tail a club mean cannot express
`home_off_qualified_share`	0.8947	1.086	-0.3364	share of the lineup carrying a qualifying sample, not a skill measurement
`home_sp_whiff`	1.0583	1.948	-0.3142	starter group: pitcher_whiff_rate, pitcher_whiff_rate_fastball, pitcher_whiff_rate_breaking
`home_off_contact`	0.3966	1.442	+0.2872	offence group: batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`away_off_power_spread`	0.5497	-0.956	-0.2260	offence group (dispersion across the lineup, not its mean): batter_home_run_rate, batter_bat_speed
`home_sp_command`	0.7673	1.214	+0.1930	starter group: pitcher_strikeout_rate, pitcher_walk_rate
`away_off_discipline`	0.2776	1.031	+0.1834	offence group: batter_whiff_rate_vs_fastball, batter_whiff_rate_vs_breaking, batter_whiff_rate_vs_offspeed, batter_strikeout_rate ...

What drove the KPI forecast for `home_runs`

intercept 4.5238 + sum of contributions = 4.7962; prediction 4.7962; contributions outside the top listed 0.1885 runs

Feature	Raw	Standardised	Runs contributed	Which KPIs it is built from
`home_off_power`	0.5244	2.160	+0.5036	offence group: batter_home_run_rate, batter_bat_speed
`away_off_top3_contact`	1.3012	1.698	-0.3773	mean of the three best contact bats, a tail a club mean cannot express
`away_off_ground`	0.6496	2.051	+0.3746	offence group: batter_ground_ball_rate_vs_slider, batter_ground_ball_rate_vs_fastball
`home_sp_whiff`	1.0583	1.948	-0.3729	starter group: pitcher_whiff_rate, pitcher_whiff_rate_fastball, pitcher_whiff_rate_breaking
`home_x_breaking`	0.4955	2.484	-0.3591	batter x pitcher interaction: pitcher_whiff_rate_breaking, batter_whiff_rate_vs_breaking
`home_off_contact_spread`	0.8940	1.138	+0.3286	offence group (dispersion across the lineup, not its mean): batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`away_off_contact`	0.5130	1.979	+0.3203	offence group: batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`home_platoon_barrel`	0.6390	1.994	-0.3183	platoon term, selected by the opposing starter's throwing hand
`home_off_top3_contact`	1.2701	1.656	-0.2985	mean of the three best contact bats, a tail a club mean cannot express
`home_off_qualified_share`	0.8947	1.086	+0.2830	share of the lineup carrying a qualifying sample, not a skill measurement

Per-player metric movement

LAD against Edgardo Henriquez (RHP)

The pipeline applied no club-wide rescale to LAD hitter lines.
Opposing staff exposure. These hitters face the starter for 20.8794 of 37.04 projected batters (56.4%) and the 2 projected relievers for the remaining 43.6%. Every metric below is the exposure-weighted blend of both matchups; the starter-only figure is shown beside it. Real 2026 game logs put a starter at 21.75 batters faced on average (median 23, p10-p90 17-27) and a whole staff at 37.9: this starter is inside that band and this staff is inside its. The per-player columns are the measured alternative: each subject moved by his own KPI pairing. Exposure for a hitter is his projected plate appearances; for the starter it is batters faced.

Player	Pos	Metric	Published → KPI	Δ	League	Own rate	Opponent rate	Log5	Own skill	Opponent	Remainder
Mookie Betts	SS	Hitswhole staff	1.057 → 1.056	-0.000	0.2169	0.2268n=523 Good	0.2222n=614 Average	0.2323	+0.042	+0.023	+0.001 ok
		split	starter x1.0243 at 0.564 share (log5 0.2323 vs his 0.2222) + bullpen x0.9682 at 0.436 share (log5 0.2196 vs pen 0.2099) = x0.9998 \| starter alone would have been +0.026
		Strikeoutswhole staff	0.524 → 0.531	+0.006	0.2212	0.1168n=523 Very good	0.2229n=614 Average	0.1178	-0.443	+0.007	-0.003 ok
		split	starter x1.0087 at 0.564 share (log5 0.1178 vs his 0.2229) + bullpen x1.0164 at 0.436 share (log5 0.1187 vs pen 0.2244) = x1.0121 \| starter alone would have been +0.005
		Walkswhole staff	0.408 → 0.373	-0.036	0.0889	0.0866n=523 Average	0.0778n=614 Good	0.0757	-0.010	-0.047	+0.001 ok
		split	starter x0.8744 at 0.564 share (log5 0.0757 vs his 0.0778) + bullpen x0.9627 at 0.436 share (log5 0.0833 vs pen 0.0856) = x0.9129 \| starter alone would have been -0.051
		Home runswhole staff	0.158 → 0.159	+0.001	0.0304	0.0394n=523 Good	0.0317n=614 Average	0.0410	+0.038	+0.005	+0.002 ok
		split	starter x1.0412 at 0.564 share (log5 0.0410 vs his 0.0317) + bullpen x0.9639 at 0.436 share (log5 0.0380 vs pen 0.0293) = x1.0075 \| starter alone would have been +0.006
Neutralising Mookie Betts to the qualified-population mean moves the club's KPI run forecast by 0.3096 runs, and the weighted stage by 0.1053. Interactions make these non-additive across players, so they do not sum to the published movement.
Freddie Freeman	1B	Hitswhole staff	1.060 → 1.060	-0.000	0.2169	0.2394n=652 Very good	0.2222n=614 Average	0.2451	+0.098	+0.023	+0.002 ok
		split	starter x1.0239 at 0.564 share (log5 0.2451 vs his 0.2222) + bullpen x0.9687 at 0.436 share (log5 0.2319 vs pen 0.2099) = x0.9998 \| starter alone would have been +0.025
		Strikeoutswhole staff	0.729 → 0.738	+0.008	0.2212	0.1781n=652 Good	0.2229n=614 Average	0.1795	-0.188	+0.007	-0.001 ok
		split	starter x1.0081 at 0.564 share (log5 0.1795 vs his 0.2229) + bullpen x1.0152 at 0.436 share (log5 0.1808 vs pen 0.2244) = x1.0112 \| starter alone would have been +0.006
		Walkswhole staff	0.448 → 0.409	-0.038	0.0889	0.0993n=652 Average	0.0778n=614 Good	0.0870	+0.045	-0.049	-0.005 ok
		split	starter x0.8759 at 0.564 share (log5 0.0870 vs his 0.0778) + bullpen x0.9632 at 0.436 share (log5 0.0957 vs pen 0.0856) = x0.914 \| starter alone would have been -0.056
		Home runswhole staff	0.139 → 0.140	+0.001	0.0304	0.0262n=652 Average	0.0317n=614 Average	0.0273	-0.018	+0.005	-0.001 ok
		split	starter x1.0418 at 0.564 share (log5 0.0273 vs his 0.0317) + bullpen x0.9634 at 0.436 share (log5 0.0253 vs pen 0.0293) = x1.0076 \| starter alone would have been +0.006
Neutralising Freddie Freeman to the qualified-population mean moves the club's KPI run forecast by 0.1949 runs, and the weighted stage by 0.0662. Interactions make these non-additive across players, so they do not sum to the published movement.
Will Smith	C	Hitswhole staff	0.935 → 0.934	-0.000	0.2169	0.2185n=276 Average	0.2222n=614 Average	0.2238	+0.007	+0.022	+0.000 ok
		split	starter x1.0245 at 0.564 share (log5 0.2238 vs his 0.2222) + bullpen x0.9679 at 0.436 share (log5 0.2114 vs pen 0.2099) = x0.9998 \| starter alone would have been +0.023
		Strikeoutswhole staff	0.699 → 0.707	+0.008	0.2212	0.1718n=276 Good	0.2229n=614 Average	0.1732	-0.204	+0.007	-0.001 ok
		split	starter x1.0082 at 0.564 share (log5 0.1732 vs his 0.2229) + bullpen x1.0153 at 0.436 share (log5 0.1744 vs pen 0.2244) = x1.0113 \| starter alone would have been +0.006
		Walkswhole staff	0.437 → 0.400	-0.037	0.0889	0.1129n=276 Good	0.0778n=614 Good	0.0990	+0.099	-0.046	-0.011 ok
		split	starter x0.8776 at 0.564 share (log5 0.0990 vs his 0.0778) + bullpen x0.9638 at 0.436 share (log5 0.1088 vs pen 0.0856) = x0.9152 \| starter alone would have been -0.053
		Home runswhole staff	0.159 → 0.160	+0.001	0.0304	0.0334n=276 Average	0.0317n=614 Average	0.0348	+0.012	+0.005	+0.001 ok
		split	starter x1.0415 at 0.564 share (log5 0.0348 vs his 0.0317) + bullpen x0.9637 at 0.436 share (log5 0.0322 vs pen 0.0293) = x1.0076 \| starter alone would have been +0.007
Neutralising Will Smith to the qualified-population mean moves the club's KPI run forecast by 0.1339 runs, and the weighted stage by 0.0455. Interactions make these non-additive across players, so they do not sum to the published movement.
Shohei Ohtani	DH	Hitswhole staff	0.896 → 0.896	-0.000	0.2169	0.2255n=613 Good	0.2222n=614 Average	0.2310	+0.036	+0.023	+0.001 ok
		split	starter x1.0243 at 0.564 share (log5 0.2310 vs his 0.2222) + bullpen x0.9682 at 0.436 share (log5 0.2183 vs pen 0.2099) = x0.9998 \| starter alone would have been +0.022
		Strikeoutswhole staff	1.012 → 1.023	+0.010	0.2212	0.2445n=613 Average	0.2229n=614 Average	0.2463	+0.099	+0.007	+0.001 ok
		split	starter x1.0074 at 0.564 share (log5 0.2463 vs his 0.2229) + bullpen x1.014 at 0.436 share (log5 0.2479 vs pen 0.2244) = x1.0103 \| starter alone would have been +0.007
		Walkswhole staff	0.487 → 0.446	-0.041	0.0889	0.1251n=613 Very good	0.0778n=614 Good	0.1100	+0.154	-0.047	-0.017 ok
		split	starter x0.8791 at 0.564 share (log5 0.1100 vs his 0.0778) + bullpen x0.9642 at 0.436 share (log5 0.1206 vs pen 0.0856) = x0.9162 \| starter alone would have been -0.059
		Home runswhole staff	0.181 → 0.182	+0.001	0.0304	0.0434n=613 Very good	0.0317n=614 Average	0.0451	+0.055	+0.005	+0.002 ok
		split	starter x1.041 at 0.564 share (log5 0.0451 vs his 0.0317) + bullpen x0.9641 at 0.436 share (log5 0.0418 vs pen 0.0293) = x1.0074 \| starter alone would have been +0.007
Neutralising Shohei Ohtani to the qualified-population mean moves the club's KPI run forecast by -0.0401 runs, and the weighted stage by -0.0136. Interactions make these non-additive across players, so they do not sum to the published movement.
Teoscar Hernández	LF	Hitswhole staff	0.952 → 0.951	-0.000	0.2169	0.2292n=487 Good	0.2222n=614 Average	0.2347	+0.049	+0.021	+0.001 ok
		split	starter x1.0242 at 0.564 share (log5 0.2347 vs his 0.2222) + bullpen x0.9683 at 0.436 share (log5 0.2219 vs pen 0.2099) = x0.9998 \| starter alone would have been +0.023
		Strikeoutswhole staff	1.057 → 1.067	+0.011	0.2212	0.2684n=487 Below average	0.2229n=614 Average	0.2703	+0.186	+0.007	+0.001 ok
		split	starter x1.0072 at 0.564 share (log5 0.2703 vs his 0.2229) + bullpen x1.0135 at 0.436 share (log5 0.2720 vs pen 0.2244) = x1.0099 \| starter alone would have been +0.008
		Walkswhole staff	0.280 → 0.255	-0.024	0.0889	0.0868n=487 Average	0.0778n=614 Good	0.0759	-0.009	-0.044	+0.001 ok
		split	starter x0.8744 at 0.564 share (log5 0.0759 vs his 0.0778) + bullpen x0.9627 at 0.436 share (log5 0.0835 vs pen 0.0856) = x0.9129 \| starter alone would have been -0.035
		Home runswhole staff	0.134 → 0.135	+0.001	0.0304	0.0320n=487 Average	0.0317n=614 Average	0.0333	+0.006	+0.005	+0.000 ok
		split	starter x1.0416 at 0.564 share (log5 0.0333 vs his 0.0317) + bullpen x0.9637 at 0.436 share (log5 0.0308 vs pen 0.0293) = x1.0076 \| starter alone would have been +0.006
Neutralising Teoscar Hernández to the qualified-population mean moves the club's KPI run forecast by -0.0332 runs, and the weighted stage by -0.0113. Interactions make these non-additive across players, so they do not sum to the published movement.
Max Muncy	3B	Hitswhole staff	0.913 → 0.913	-0.000	0.2169	0.2173n=556 Average	0.2222n=614 Average	0.2226	+0.002	+0.021	0 ok
		split	starter x1.0246 at 0.564 share (log5 0.2226 vs his 0.2222) + bullpen x0.9678 at 0.436 share (log5 0.2103 vs pen 0.2099) = x0.9998 \| starter alone would have been +0.022
		Strikeoutswhole staff	0.966 → 0.976	+0.010	0.2212	0.2493n=556 Below average	0.2229n=614 Average	0.2512	+0.111	+0.007	+0.001 ok
		split	starter x1.0074 at 0.564 share (log5 0.2512 vs his 0.2229) + bullpen x1.0139 at 0.436 share (log5 0.2528 vs pen 0.2244) = x1.0102 \| starter alone would have been +0.007
		Walkswhole staff	0.381 → 0.349	-0.032	0.0889	0.1105n=556 Good	0.0778n=614 Good	0.0969	+0.085	-0.044	-0.009 ok
		split	starter x0.8773 at 0.564 share (log5 0.0969 vs his 0.0778) + bullpen x0.9637 at 0.436 share (log5 0.1065 vs pen 0.0856) = x0.915 \| starter alone would have been -0.047
		Home runswhole staff	0.184 → 0.185	+0.001	0.0304	0.0451n=556 Very good	0.0317n=614 Average	0.0470	+0.058	+0.005	+0.002 ok
		split	starter x1.041 at 0.564 share (log5 0.0470 vs his 0.0317) + bullpen x0.9641 at 0.436 share (log5 0.0435 vs pen 0.0293) = x1.0074 \| starter alone would have been +0.007
Neutralising Max Muncy to the qualified-population mean moves the club's KPI run forecast by 0.0012 runs, and the weighted stage by 0.0004. Interactions make these non-additive across players, so they do not sum to the published movement.
Enrique Hernández	2B	Hitswhole staff	0.805 → 0.805	-0.000	0.2169	0.2169n=0 averagebelow floor	0.2222n=614 Average	0.2222	0	+0.018	0 ok
		split	starter x1.0246 at 0.564 share (log5 0.2222 vs his 0.2222) + bullpen x0.9678 at 0.436 share (log5 0.2099 vs pen 0.2099) = x0.9998 \| starter alone would have been +0.020
		Strikeoutswhole staff	0.980 → 0.990	+0.010	0.2212	0.2212n=0 averagebelow floor	0.2229n=614 Average	0.2229	0	+0.006	0 ok
		split	starter x1.0077 at 0.564 share (log5 0.2229 vs his 0.2229) + bullpen x1.0144 at 0.436 share (log5 0.2244 vs pen 0.2244) = x1.0106 \| starter alone would have been +0.007
		Walkswhole staff	0.223 → 0.204	-0.019	0.0889	0.0889n=0 averagebelow floor	0.0778n=614 Good	0.0778	0	-0.039	0 ok
		split	starter x0.8747 at 0.564 share (log5 0.0778 vs his 0.0778) + bullpen x0.9628 at 0.436 share (log5 0.0856 vs pen 0.0856) = x0.9131 \| starter alone would have been -0.028
		Home runswhole staff	0.137 → 0.138	+0.001	0.0304	0.0304n=0 averagebelow floor	0.0317n=614 Average	0.0317	0	+0.004	0 ok
		split	starter x1.0416 at 0.564 share (log5 0.0317 vs his 0.0317) + bullpen x0.9636 at 0.436 share (log5 0.0293 vs pen 0.0293) = x1.0076 \| starter alone would have been +0.006
Neutralising Enrique Hernández to the qualified-population mean moves the club's KPI run forecast by 0.0000 runs, and the weighted stage by 0.0000. Interactions make these non-additive across players, so they do not sum to the published movement.
Kyle Tucker	RF	Hitswhole staff	0.791 → 0.791	-0.000	0.2169	0.2076n=622 Below average	0.2222n=614 Average	0.2127	-0.037	+0.021	-0.001 ok
		split	starter x1.0249 at 0.564 share (log5 0.2127 vs his 0.2222) + bullpen x0.9674 at 0.436 share (log5 0.2008 vs pen 0.2099) = x0.9998 \| starter alone would have been +0.020
		Strikeoutswhole staff	0.659 → 0.666	+0.007	0.2212	0.1711n=622 Good	0.2229n=614 Average	0.1725	-0.197	+0.007	-0.001 ok
		split	starter x1.0082 at 0.564 share (log5 0.1725 vs his 0.2229) + bullpen x1.0154 at 0.436 share (log5 0.1737 vs pen 0.2244) = x1.0113 \| starter alone would have been +0.005
		Walkswhole staff	0.455 → 0.416	-0.038	0.0889	0.1195n=622 Very good	0.0778n=614 Good	0.1050	+0.120	-0.044	-0.013 ok
		split	starter x0.8784 at 0.564 share (log5 0.1050 vs his 0.0778) + bullpen x0.964 at 0.436 share (log5 0.1152 vs pen 0.0856) = x0.9157 \| starter alone would have been -0.055
		Home runswhole staff	0.119 → 0.120	+0.001	0.0304	0.0282n=622 Average	0.0317n=614 Average	0.0294	-0.008	+0.005	-0.000 ok
		split	starter x1.0417 at 0.564 share (log5 0.0294 vs his 0.0317) + bullpen x0.9635 at 0.436 share (log5 0.0272 vs pen 0.0293) = x1.0076 \| starter alone would have been +0.005
Neutralising Kyle Tucker to the qualified-population mean moves the club's KPI run forecast by -0.0192 runs, and the weighted stage by -0.0065. Interactions make these non-additive across players, so they do not sum to the published movement.
Andy Pages	CF	Hitswhole staff	0.840 → 0.839	-0.000	0.2169	0.2323n=569 Good	0.2222n=614 Average	0.2379	+0.056	+0.019	+0.001 ok
		split	starter x1.0241 at 0.564 share (log5 0.2379 vs his 0.2222) + bullpen x0.9684 at 0.436 share (log5 0.2250 vs pen 0.2099) = x0.9998 \| starter alone would have been +0.020
		Strikeoutswhole staff	0.731 → 0.739	+0.008	0.2212	0.1971n=569 Average	0.2229n=614 Average	0.1987	-0.087	+0.006	-0.001 ok
		split	starter x1.0079 at 0.564 share (log5 0.1987 vs his 0.2229) + bullpen x1.0149 at 0.436 share (log5 0.2001 vs pen 0.2244) = x1.011 \| starter alone would have been +0.006
		Walkswhole staff	0.337 → 0.308	-0.029	0.0889	0.0866n=569 Average	0.0778n=614 Good	0.0757	-0.008	-0.040	+0.001 ok
		split	starter x0.8744 at 0.564 share (log5 0.0757 vs his 0.0778) + bullpen x0.9627 at 0.436 share (log5 0.0834 vs pen 0.0856) = x0.9129 \| starter alone would have been -0.042
		Home runswhole staff	0.136 → 0.137	+0.001	0.0304	0.0372n=569 Good	0.0317n=614 Average	0.0388	+0.025	+0.005	+0.001 ok
		split	starter x1.0413 at 0.564 share (log5 0.0388 vs his 0.0317) + bullpen x0.9638 at 0.436 share (log5 0.0359 vs pen 0.0293) = x1.0075 \| starter alone would have been +0.006
Neutralising Andy Pages to the qualified-population mean moves the club's KPI run forecast by 0.1699 runs, and the weighted stage by 0.0578. Interactions make these non-additive across players, so they do not sum to the published movement.
STARTER Tyler Glasnow20.9371 BF	Starter	Hits	3.503 → 3.637	+0.134	0.2169	0.1869n=288 Very good	0.22498/9 qualified lineup mean	0.1941	-0.628	+0.167	-0.017 ok
		Strikeouts	6.814 → 6.540	-0.275	0.2212	0.3143n=288 Very good	0.21118/9 qualified lineup mean	0.3017	+1.950	-0.211	-0.054 ok
		Walks	1.829 → 1.698	-0.131	0.0889	0.0876n=288 Average	0.08268/9 qualified lineup mean	0.0813	-0.029	-0.133	+0.002 ok
		Home runs	0.393 → 0.454	+0.061	0.0304	0.0280n=288 Average	0.03518/9 qualified lineup mean	0.0324	-0.050	+0.099	-0.007 ok
RELIEVER Kris Bubic2.7485 BF	Reliever	Hits	0.548 → 0.569	+0.021	0.2169	0.1996n=226 Good	0.22498/9 qualified lineup mean	0.2071	-0.048	+0.022	-0.001 ok
		Strikeouts	0.662 → 0.632	-0.029	0.2212	0.2408n=226 Average	0.21118/9 qualified lineup mean	0.2301	+0.054	-0.028	-0.002 ok
		Walks	0.287 → 0.267	-0.020	0.0889	0.1043n=226 Below average	0.08268/9 qualified lineup mean	0.0970	+0.042	-0.018	-0.003 ok
		Home runs	0.072 → 0.083	+0.011	0.0304	0.0261n=226 Good	0.03518/9 qualified lineup mean	0.0302	-0.012	+0.013	-0.002 ok
RELIEVER Brock Stewart2.2263 BF	Middle Reliever (High Leverage)	Hits	0.522 → 0.541	+0.019	0.2169	0.2169n=0 averagebelow floor	0.22498/9 qualified lineup mean	0.2249	0	+0.018	0 ok
		Strikeouts	0.968 → 0.924	-0.044	0.2212	0.2212n=0 averagebelow floor	0.21118/9 qualified lineup mean	0.2111	0	-0.022	0 ok
		Walks	0.330 → 0.306	-0.024	0.0889	0.0889n=0 averagebelow floor	0.08268/9 qualified lineup mean	0.0826	0	-0.014	0 ok
		Home runs	0.058 → 0.067	+0.009	0.0304	0.0304n=0 averagebelow floor	0.03518/9 qualified lineup mean	0.0351	0	+0.011	0 ok
RELIEVER Roki Sasaki2.2029 BF	Middle Reliever	Hits	0.868 → 0.900	+0.032	0.2169	0.2166n=522 Average	0.22498/9 qualified lineup mean	0.2246	-0.001	+0.018	0 ok
		Strikeouts	1.578 → 1.507	-0.071	0.2212	0.2337n=522 Average	0.21118/9 qualified lineup mean	0.2232	+0.028	-0.022	-0.001 ok
		Walks	0.502 → 0.466	-0.036	0.0889	0.0913n=522 Average	0.08268/9 qualified lineup mean	0.0847	+0.005	-0.014	-0.000 ok
		Home runs	0.088 → 0.102	+0.014	0.0304	0.0367n=522 Poor	0.03518/9 qualified lineup mean	0.0424	+0.014	+0.010	+0.002 ok
RELIEVER Alex Vesia2.1606 BF	Middle Reliever	Hits	0.387 → 0.402	+0.015	0.2169	0.1766n=237 Very good	0.22498/9 qualified lineup mean	0.1835	-0.087	+0.017	-0.003 ok
		Strikeouts	1.340 → 1.284	-0.056	0.2212	0.2892n=237 Very good	0.21118/9 qualified lineup mean	0.2772	+0.147	-0.022	-0.004 ok
		Walks	0.450 → 0.419	-0.031	0.0889	0.1166n=237 Poor	0.08268/9 qualified lineup mean	0.1085	+0.060	-0.014	-0.004 ok
		Home runs	0.052 → 0.060	+0.008	0.0304	0.0257n=237 Good	0.03518/9 qualified lineup mean	0.0297	-0.010	+0.010	-0.002 ok
RELIEVER Evan Phillips1.9783 BF	Reliever	Hits	0.429 → 0.445	+0.016	0.2169	0.2169n=0 averagebelow floor	0.22498/9 qualified lineup mean	0.2249	0	+0.016	0 ok
		Strikeouts	0.438 → 0.418	-0.020	0.2212	0.2212n=0 averagebelow floor	0.21118/9 qualified lineup mean	0.2111	0	-0.020	0 ok
		Walks	0.176 → 0.163	-0.013	0.0889	0.0889n=0 averagebelow floor	0.08268/9 qualified lineup mean	0.0826	0	-0.013	0 ok
		Home runs	0.060 → 0.069	+0.009	0.0304	0.0304n=0 averagebelow floor	0.03518/9 qualified lineup mean	0.0351	0	+0.009	0 ok
RELIEVER Tanner Scott1.7025 BF	Reliever	Hits	0.349 → 0.362	+0.013	0.2169	0.2051n=251 Good	0.22498/9 qualified lineup mean	0.2127	-0.020	+0.014	-0.001 ok
		Strikeouts	0.513 → 0.492	-0.021	0.2212	0.3011n=251 Very good	0.21118/9 qualified lineup mean	0.2888	+0.136	-0.017	-0.004 ok
		Walks	0.121 → 0.112	-0.009	0.0889	0.0710n=251 Good	0.08268/9 qualified lineup mean	0.0658	-0.030	-0.011	+0.002 ok
		Home runs	0.043 → 0.050	+0.007	0.0304	0.0252n=251 Good	0.03518/9 qualified lineup mean	0.0291	-0.009	+0.008	-0.001 ok
RELIEVER Edgardo Henriquez1.5698 BF	Reliever	Hits	0.295 → 0.306	+0.011	0.2169	0.1880n=242 Very good	0.22498/9 qualified lineup mean	0.1952	-0.045	+0.013	-0.001 ok
		Strikeouts	0.432 → 0.414	-0.018	0.2212	0.2754n=242 Good	0.21118/9 qualified lineup mean	0.2637	+0.085	-0.016	-0.003 ok
		Walks	0.153 → 0.142	-0.011	0.0889	0.0976n=242 Average	0.08268/9 qualified lineup mean	0.0907	+0.014	-0.010	-0.001 ok
		Home runs	0.033 → 0.038	+0.005	0.0304	0.0212n=242 Very good	0.03518/9 qualified lineup mean	0.0245	-0.015	+0.007	-0.002 ok

ATL against Dylan Lee (LHP)

The pipeline applied one club-wide ×1.2275 to every ATL hitter line, identical for the leadoff hitter and the ninth bat.
Opposing staff exposure. These hitters face the starter for 20.9371 of 35.526 projected batters (58.9%) and the 3 projected relievers for the remaining 41.1%. Every metric below is the exposure-weighted blend of both matchups; the starter-only figure is shown beside it. Real 2026 game logs put a starter at 21.75 batters faced on average (median 23, p10-p90 17-27) and a whole staff at 37.9: this starter is inside that band and this staff is inside its. The per-player columns are the measured alternative: each subject moved by his own KPI pairing. Exposure for a hitter is his projected plate appearances; for the starter it is batters faced.

Player	Pos	Metric	Published → KPI	Δ	League	Own rate	Opponent rate	Log5	Own skill	Opponent	Remainder
Michael Harris II	CF	Hitswhole staff	1.014 → 0.910	-0.104	0.2169	0.2533n=608 Very good	0.1869n=288 Very good	0.2197	+0.156	-0.129	-0.015 ok
		split	starter x0.8674 at 0.589 share (log5 0.2197 vs his 0.1869) + bullpen x0.9394 at 0.411 share (log5 0.2379 vs pen 0.2031) = x0.897 \| starter alone would have been -0.135
		Strikeoutswhole staff	1.099 → 1.453	+0.353	0.2212	0.1915n=608 Average	0.3143n=288 Very good	0.2766	-0.127	+0.399	-0.035 ok
		split	starter x1.4442 at 0.589 share (log5 0.2766 vs his 0.3143) + bullpen x1.1454 at 0.411 share (log5 0.2194 vs pen 0.2520) = x1.3215 \| starter alone would have been +0.488
		Walkswhole staff	0.232 → 0.237	+0.005	0.0889	0.0462n=608 Poor	0.0876n=288 Average	0.0455	-0.183	-0.006	+0.003 ok
		split	starter x0.9839 at 0.589 share (log5 0.0455 vs his 0.0876) + bullpen x1.0732 at 0.411 share (log5 0.0496 vs pen 0.0951) = x1.0206 \| starter alone would have been -0.004
		Home runswhole staff	0.197 → 0.182	-0.015	0.0304	0.0379n=608 Good	0.0280n=288 Average	0.0349	+0.032	-0.010	-0.002 ok
		split	starter x0.9221 at 0.589 share (log5 0.0349 vs his 0.0280) + bullpen x0.9301 at 0.411 share (log5 0.0352 vs pen 0.0282) = x0.9254 \| starter alone would have been -0.015
Neutralising Michael Harris II to the qualified-population mean moves the club's KPI run forecast by 0.3276 runs, and the weighted stage by 0.1245. Interactions make these non-additive across players, so they do not sum to the published movement.
Ronald Acuña Jr.	RF	Hitswhole staff	0.952 → 0.851	-0.102	0.2169	0.2212n=457 Average	0.1869n=288 Very good	0.1908	+0.019	-0.130	-0.002 ok
		split	starter x0.8624 at 0.589 share (log5 0.1908 vs his 0.1869) + bullpen x0.9369 at 0.411 share (log5 0.2073 vs pen 0.2031) = x0.893 \| starter alone would have been -0.131
		Strikeoutswhole staff	1.124 → 1.468	+0.344	0.2212	0.2190n=457 Average	0.3143n=288 Very good	0.3116	-0.009	+0.402	-0.002 ok
		split	starter x1.4227 at 0.589 share (log5 0.3116 vs his 0.3143) + bullpen x1.1398 at 0.411 share (log5 0.2497 vs pen 0.2520) = x1.3065 \| starter alone would have been +0.475
		Walkswhole staff	0.458 → 0.467	+0.009	0.0889	0.1104n=457 Good	0.0876n=288 Average	0.1087	+0.093	-0.006	-0.001 ok
		split	starter x0.985 at 0.589 share (log5 0.1087 vs his 0.0876) + bullpen x1.0679 at 0.411 share (log5 0.1179 vs pen 0.0951) = x1.019 \| starter alone would have been -0.007
		Home runswhole staff	0.175 → 0.162	-0.013	0.0304	0.0389n=457 Good	0.0280n=288 Average	0.0358	+0.037	-0.010	-0.003 ok
		split	starter x0.9222 at 0.589 share (log5 0.0358 vs his 0.0280) + bullpen x0.9302 at 0.411 share (log5 0.0361 vs pen 0.0282) = x0.9255 \| starter alone would have been -0.014
Neutralising Ronald Acuña Jr. to the qualified-population mean moves the club's KPI run forecast by 0.2291 runs, and the weighted stage by 0.0871. Interactions make these non-additive across players, so they do not sum to the published movement.
Matt Olson	1B	Hitswhole staff	0.865 → 0.773	-0.092	0.2169	0.2232n=682 Average	0.1869n=288 Very good	0.1926	+0.027	-0.128	-0.003 ok
		split	starter x0.8627 at 0.589 share (log5 0.1926 vs his 0.1869) + bullpen x0.9371 at 0.411 share (log5 0.2092 vs pen 0.2031) = x0.8933 \| starter alone would have been -0.119
		Strikeoutswhole staff	1.184 → 1.533	+0.349	0.2212	0.2406n=682 Average	0.3143n=288 Very good	0.3383	+0.083	+0.397	+0.020 ok
		split	starter x1.4063 at 0.589 share (log5 0.3383 vs his 0.3143) + bullpen x1.1354 at 0.411 share (log5 0.2731 vs pen 0.2520) = x1.295 \| starter alone would have been +0.481
		Walkswhole staff	0.446 → 0.454	+0.009	0.0889	0.0919n=682 Average	0.0876n=288 Average	0.0905	+0.013	-0.006	-0.000 ok
		split	starter x0.9847 at 0.589 share (log5 0.0905 vs his 0.0876) + bullpen x1.0695 at 0.411 share (log5 0.0982 vs pen 0.0951) = x1.0195 \| starter alone would have been -0.007
		Home runswhole staff	0.232 → 0.215	-0.017	0.0304	0.0518n=682 Very good	0.0280n=288 Average	0.0478	+0.091	-0.010	-0.007 ok
		split	starter x0.9231 at 0.589 share (log5 0.0478 vs his 0.0280) + bullpen x0.931 at 0.411 share (log5 0.0482 vs pen 0.0282) = x0.9263 \| starter alone would have been -0.018
Neutralising Matt Olson to the qualified-population mean moves the club's KPI run forecast by 0.4574 runs, and the weighted stage by 0.1738. Interactions make these non-additive across players, so they do not sum to the published movement.
Ozzie Albies	2B	Hitswhole staff	0.773 → 0.690	-0.083	0.2169	0.2154n=667 Average	0.1869n=288 Very good	0.1856	-0.006	-0.119	+0.001 ok
		split	starter x0.8616 at 0.589 share (log5 0.1856 vs his 0.1869) + bullpen x0.9365 at 0.411 share (log5 0.2017 vs pen 0.2031) = x0.8924 \| starter alone would have been -0.107
		Strikeoutswhole staff	0.781 → 1.047	+0.266	0.2212	0.1576n=667 Good	0.3143n=288 Very good	0.2319	-0.253	+0.371	-0.075 ok
		split	starter x1.4717 at 0.589 share (log5 0.2319 vs his 0.3143) + bullpen x1.1524 at 0.411 share (log5 0.1816 vs pen 0.2520) = x1.3406 \| starter alone would have been +0.368
		Walkswhole staff	0.345 → 0.352	+0.007	0.0889	0.0707n=667 Below average	0.0876n=288 Average	0.0696	-0.072	-0.005	+0.001 ok
		split	starter x0.9843 at 0.589 share (log5 0.0696 vs his 0.0876) + bullpen x1.0712 at 0.411 share (log5 0.0758 vs pen 0.0951) = x1.02 \| starter alone would have been -0.005
		Home runswhole staff	0.140 → 0.130	-0.011	0.0304	0.0322n=667 Average	0.0280n=288 Average	0.0297	+0.007	-0.009	-0.001 ok
		split	starter x0.9217 at 0.589 share (log5 0.0297 vs his 0.0280) + bullpen x0.9297 at 0.411 share (log5 0.0300 vs pen 0.0282) = x0.925 \| starter alone would have been -0.011
Neutralising Ozzie Albies to the qualified-population mean moves the club's KPI run forecast by 0.4581 runs, and the weighted stage by 0.1741. Interactions make these non-additive across players, so they do not sum to the published movement.
Drake Baldwin	DH	Hitswhole staff	0.933 → 0.835	-0.098	0.2169	0.2374n=598 Very good	0.1869n=288 Very good	0.2053	+0.088	-0.128	-0.009 ok
		split	starter x0.8649 at 0.589 share (log5 0.2053 vs his 0.1869) + bullpen x0.9382 at 0.411 share (log5 0.2227 vs pen 0.2031) = x0.895 \| starter alone would have been -0.126
		Strikeoutswhole staff	1.070 → 1.403	+0.333	0.2212	0.2099n=598 Average	0.3143n=288 Very good	0.3002	-0.048	+0.399	-0.013 ok
		split	starter x1.4297 at 0.589 share (log5 0.3002 vs his 0.3143) + bullpen x1.1416 at 0.411 share (log5 0.2397 vs pen 0.2520) = x1.3114 \| starter alone would have been +0.460
		Walkswhole staff	0.389 → 0.397	+0.008	0.0889	0.0915n=598 Average	0.0876n=288 Average	0.0901	+0.011	-0.006	-0.000 ok
		split	starter x0.9847 at 0.589 share (log5 0.0901 vs his 0.0876) + bullpen x1.0695 at 0.411 share (log5 0.0978 vs pen 0.0951) = x1.0195 \| starter alone would have been -0.006
		Home runswhole staff	0.177 → 0.164	-0.013	0.0304	0.0371n=598 Good	0.0280n=288 Average	0.0343	+0.029	-0.010	-0.002 ok
		split	starter x0.9221 at 0.589 share (log5 0.0343 vs his 0.0280) + bullpen x0.9301 at 0.411 share (log5 0.0345 vs pen 0.0282) = x0.9254 \| starter alone would have been -0.014
Neutralising Drake Baldwin to the qualified-population mean moves the club's KPI run forecast by 0.2741 runs, and the weighted stage by 0.1042. Interactions make these non-additive across players, so they do not sum to the published movement.
Mike Yastrzemski	LF	Hitswhole staff	0.705 → 0.628	-0.076	0.2169	0.2093n=398 Below average	0.1869n=288 Very good	0.1802	-0.029	-0.117	+0.003 ok
		split	starter x0.8606 at 0.589 share (log5 0.1802 vs his 0.1869) + bullpen x0.936 at 0.411 share (log5 0.1959 vs pen 0.2031) = x0.8916 \| starter alone would have been -0.098
		Strikeoutswhole staff	1.138 → 1.481	+0.342	0.2212	0.2301n=398 Average	0.3143n=288 Very good	0.3254	+0.035	+0.363	+0.009 ok
		split	starter x1.4142 at 0.589 share (log5 0.3254 vs his 0.3143) + bullpen x1.1375 at 0.411 share (log5 0.2617 vs pen 0.2520) = x1.3006 \| starter alone would have been +0.472
		Walkswhole staff	0.350 → 0.356	+0.007	0.0889	0.0998n=398 Average	0.0876n=288 Average	0.0983	+0.042	-0.005	-0.001 ok
		split	starter x0.9848 at 0.589 share (log5 0.0983 vs his 0.0876) + bullpen x1.0688 at 0.411 share (log5 0.1066 vs pen 0.0951) = x1.0193 \| starter alone would have been -0.005
		Home runswhole staff	0.129 → 0.119	-0.010	0.0304	0.0302n=398 Average	0.0280n=288 Average	0.0279	-0.001	-0.009	0 ok
		split	starter x0.9215 at 0.589 share (log5 0.0279 vs his 0.0280) + bullpen x0.9296 at 0.411 share (log5 0.0281 vs pen 0.0282) = x0.9248 \| starter alone would have been -0.010
Neutralising Mike Yastrzemski to the qualified-population mean moves the club's KPI run forecast by 0.1894 runs, and the weighted stage by 0.0720. Interactions make these non-additive across players, so they do not sum to the published movement.
Mauricio Dubón	SS	Hitswhole staff	0.950 → 0.850	-0.100	0.2169	0.2382n=603 Very good	0.1869n=288 Very good	0.2060	+0.082	-0.115	-0.008 ok
		split	starter x0.865 at 0.589 share (log5 0.2060 vs his 0.1869) + bullpen x0.9382 at 0.411 share (log5 0.2235 vs pen 0.2031) = x0.8951 \| starter alone would have been -0.128
		Strikeoutswhole staff	0.628 → 0.853	+0.225	0.2212	0.1270n=603 Very good	0.3143n=288 Very good	0.1902	-0.361	+0.356	-0.115 ok
		split	starter x1.4973 at 0.589 share (log5 0.1902 vs his 0.3143) + bullpen x1.1589 at 0.411 share (log5 0.1472 vs pen 0.2520) = x1.3583 \| starter alone would have been +0.312
		Walkswhole staff	0.247 → 0.252	+0.005	0.0889	0.0631n=603 Below average	0.0876n=288 Average	0.0621	-0.099	-0.005	+0.001 ok
		split	starter x0.9842 at 0.589 share (log5 0.0621 vs his 0.0876) + bullpen x1.0718 at 0.411 share (log5 0.0677 vs pen 0.0951) = x1.0202 \| starter alone would have been -0.004
		Home runswhole staff	0.104 → 0.096	-0.008	0.0304	0.0231n=603 Below average	0.0280n=288 Average	0.0213	-0.028	-0.009	+0.002 ok
		split	starter x0.921 at 0.589 share (log5 0.0213 vs his 0.0280) + bullpen x0.9291 at 0.411 share (log5 0.0215 vs pen 0.0282) = x0.9243 \| starter alone would have been -0.008
Neutralising Mauricio Dubón to the qualified-population mean moves the club's KPI run forecast by 0.0844 runs, and the weighted stage by 0.0321. Interactions make these non-additive across players, so they do not sum to the published movement.
Austin Riley	3B	Hitswhole staff	0.685 → 0.611	-0.075	0.2169	0.2034n=614 Below average	0.1869n=288 Very good	0.1749	-0.049	-0.109	+0.005 ok
		split	starter x0.8597 at 0.589 share (log5 0.1749 vs his 0.1869) + bullpen x0.9356 at 0.411 share (log5 0.1903 vs pen 0.2031) = x0.8909 \| starter alone would have been -0.096
		Strikeoutswhole staff	1.455 → 1.833	+0.379	0.2212	0.3090n=614 Poor	0.3143n=288 Very good	0.4192	+0.320	+0.340	+0.062 ok
		split	starter x1.3566 at 0.589 share (log5 0.4192 vs his 0.3143) + bullpen x1.1217 at 0.411 share (log5 0.3466 vs pen 0.2520) = x1.2601 \| starter alone would have been +0.519
		Walkswhole staff	0.317 → 0.323	+0.006	0.0889	0.0799n=614 Average	0.0876n=288 Average	0.0787	-0.033	-0.005	+0.001 ok
		split	starter x0.9845 at 0.589 share (log5 0.0787 vs his 0.0876) + bullpen x1.0704 at 0.411 share (log5 0.0856 vs pen 0.0951) = x1.0198 \| starter alone would have been -0.005
		Home runswhole staff	0.126 → 0.116	-0.009	0.0304	0.0308n=614 Average	0.0280n=288 Average	0.0284	+0.001	-0.009	-0.000 ok
		split	starter x0.9216 at 0.589 share (log5 0.0284 vs his 0.0280) + bullpen x0.9296 at 0.411 share (log5 0.0286 vs pen 0.0282) = x0.9249 \| starter alone would have been -0.010
Neutralising Austin Riley to the qualified-population mean moves the club's KPI run forecast by 0.3284 runs, and the weighted stage by 0.1248. Interactions make these non-additive across players, so they do not sum to the published movement.
Sean Murphy	C	Hitswhole staff	0.616 → 0.550	-0.066	0.2169	0.2169n=0 averagebelow floor	0.1869n=288 Very good	0.1869	0	-0.102	0 ok
		split	starter x0.8618 at 0.589 share (log5 0.1869 vs his 0.1869) + bullpen x0.9366 at 0.411 share (log5 0.2031 vs pen 0.2031) = x0.8925 \| starter alone would have been -0.085
		Strikeoutswhole staff	1.135 → 1.482	+0.347	0.2212	0.2212n=0 averagebelow floor	0.3143n=288 Very good	0.3143	0	+0.318	0 ok
		split	starter x1.421 at 0.589 share (log5 0.3143 vs his 0.3143) + bullpen x1.1393 at 0.411 share (log5 0.2520 vs pen 0.2520) = x1.3053 \| starter alone would have been +0.478
		Walkswhole staff	0.322 → 0.328	+0.006	0.0889	0.0889n=0 averagebelow floor	0.0876n=288 Average	0.0876	0	-0.005	0 ok
		split	starter x0.9846 at 0.589 share (log5 0.0876 vs his 0.0876) + bullpen x1.0697 at 0.411 share (log5 0.0951 vs pen 0.0951) = x1.0196 \| starter alone would have been -0.005
		Home runswhole staff	0.119 → 0.110	-0.009	0.0304	0.0304n=0 averagebelow floor	0.0280n=288 Average	0.0280	0	-0.008	0 ok
		split	starter x0.9215 at 0.589 share (log5 0.0280 vs his 0.0280) + bullpen x0.9296 at 0.411 share (log5 0.0282 vs pen 0.0282) = x0.9248 \| starter alone would have been -0.009
Neutralising Sean Murphy to the qualified-population mean moves the club's KPI run forecast by -0.0006 runs, and the weighted stage by -0.0002. Interactions make these non-additive across players, so they do not sum to the published movement.
STARTER Tyler Mahle20.8794 BF	Starter	Hits	4.443 → 4.587	+0.143	0.2169	0.2222n=614 Average	0.22398/9 qualified lineup mean	0.2294	+0.111	+0.147	+0.003 ok
		Strikeouts	4.119 → 3.745	-0.374	0.2212	0.2229n=614 Average	0.20118/9 qualified lineup mean	0.2027	+0.035	-0.420	-0.002 ok
		Walks	1.445 → 1.663	+0.219	0.0889	0.0778n=614 Good	0.10228/9 qualified lineup mean	0.0896	-0.233	+0.277	-0.031 ok
		Home runs	0.602 → 0.695	+0.093	0.0304	0.0317n=614 Average	0.03518/9 qualified lineup mean	0.0365	+0.026	+0.098	+0.004 ok
RELIEVER AJ Smith-Shawver3.3824 BF	Reliever	Hits	0.740 → 0.763	+0.024	0.2169	0.2186n=144 Average	0.22398/9 qualified lineup mean	0.2257	+0.006	+0.024	+0.000 ok
		Strikeouts	0.675 → 0.612	-0.063	0.2212	0.1996n=144 Below average	0.20118/9 qualified lineup mean	0.1809	-0.073	-0.068	+0.005 ok
		Walks	0.337 → 0.386	+0.050	0.0889	0.0995n=144 Below average	0.10228/9 qualified lineup mean	0.1142	+0.036	+0.045	+0.005 ok
		Home runs	0.118 → 0.136	+0.018	0.0304	0.0348n=144 Below average	0.03518/9 qualified lineup mean	0.0402	+0.015	+0.016	+0.002 ok
RELIEVER Grant Holmes3.1853 BF	Long Reliever	Hits	1.515 → 1.564	+0.049	0.2169	0.2162n=607 Average	0.22398/9 qualified lineup mean	0.2232	-0.002	+0.022	-0.000 ok
		Strikeouts	1.085 → 0.982	-0.103	0.2212	0.1859n=607 Below average	0.20118/9 qualified lineup mean	0.1683	-0.113	-0.064	+0.008 ok
		Walks	0.966 → 1.109	+0.144	0.0889	0.0916n=607 Average	0.10228/9 qualified lineup mean	0.1052	+0.008	+0.042	+0.001 ok
		Home runs	0.379 → 0.437	+0.058	0.0304	0.0338n=607 Below average	0.03518/9 qualified lineup mean	0.0390	+0.011	+0.015	+0.002 ok
RELIEVER Robert Suarez2.5097 BF	Reliever	Hits	0.519 → 0.537	+0.017	0.2169	0.2070n=130 Good	0.22398/9 qualified lineup mean	0.2138	-0.025	+0.018	-0.001 ok
		Strikeouts	0.522 → 0.474	-0.048	0.2212	0.2081n=130 Average	0.20118/9 qualified lineup mean	0.1889	-0.033	-0.051	+0.002 ok
		Walks	0.175 → 0.201	+0.027	0.0889	0.0696n=130 Good	0.10228/9 qualified lineup mean	0.0802	-0.049	+0.033	-0.007 ok
		Home runs	0.063 → 0.073	+0.010	0.0304	0.0253n=130 Good	0.03518/9 qualified lineup mean	0.0292	-0.013	+0.012	-0.002 ok
RELIEVER Victor Mederos2.1995 BF	Reliever	Hits	0.452 → 0.467	+0.015	0.2169	0.2054n=148 Good	0.22398/9 qualified lineup mean	0.2121	-0.025	+0.015	-0.001 ok
		Strikeouts	0.499 → 0.454	-0.045	0.2212	0.2267n=148 Average	0.20118/9 qualified lineup mean	0.2062	+0.012	-0.044	-0.001 ok
		Walks	0.202 → 0.232	+0.030	0.0889	0.0917n=148 Average	0.10228/9 qualified lineup mean	0.1053	+0.006	+0.029	+0.001 ok
		Home runs	0.054 → 0.062	+0.008	0.0304	0.0245n=148 Very good	0.03518/9 qualified lineup mean	0.0283	-0.013	+0.010	-0.002 ok
RELIEVER Didier Fuentes2.1397 BF	Reliever	Hits	0.453 → 0.468	+0.015	0.2169	0.2119n=290 Average	0.22398/9 qualified lineup mean	0.2189	-0.011	+0.015	-0.000 ok
		Strikeouts	0.600 → 0.549	-0.051	0.2212	0.2802n=290 Very good	0.20118/9 qualified lineup mean	0.2565	+0.126	-0.043	-0.008 ok
		Walks	0.163 → 0.187	+0.025	0.0889	0.0760n=290 Good	0.10228/9 qualified lineup mean	0.0875	-0.028	+0.028	-0.004 ok
		Home runs	0.057 → 0.066	+0.009	0.0304	0.0266n=290 Good	0.03518/9 qualified lineup mean	0.0307	-0.008	+0.010	-0.001 ok
RELIEVER Dylan Dodd1.5304 BF	Middle Reliever	Hits	0.707 → 0.731	+0.024	0.2169	0.1933n=186 Very good	0.22398/9 qualified lineup mean	0.1997	-0.036	+0.011	-0.001 ok
		Strikeouts	1.112 → 1.016	-0.096	0.2212	0.2620n=186 Good	0.20118/9 qualified lineup mean	0.2393	+0.062	-0.031	-0.004 ok
		Walks	0.213 → 0.245	+0.032	0.0889	0.0815n=186 Average	0.10228/9 qualified lineup mean	0.0938	-0.011	+0.020	-0.002 ok
		Home runs	0.148 → 0.171	+0.023	0.0304	0.0294n=186 Average	0.03518/9 qualified lineup mean	0.0339	-0.002	+0.007	-0.000 ok
RELIEVER Dylan Lee1.2137 BF	Reliever	Hits	0.244 → 0.252	+0.008	0.2169	0.2007n=253 Good	0.22398/9 qualified lineup mean	0.2073	-0.020	+0.009	-0.001 ok
		Strikeouts	0.338 → 0.309	-0.029	0.2212	0.2783n=253 Very good	0.20118/9 qualified lineup mean	0.2547	+0.069	-0.024	-0.004 ok
		Walks	0.092 → 0.106	+0.014	0.0889	0.0755n=253 Good	0.10228/9 qualified lineup mean	0.0870	-0.016	+0.016	-0.002 ok
		Home runs	0.029 → 0.033	+0.004	0.0304	0.0237n=253 Very good	0.03518/9 qualified lineup mean	0.0274	-0.008	+0.006	-0.001 ok

Published pitching, integer draw beside the continuous model line

Pitcher	Role	BF	IP	expected IP	H int/cont	SO int/cont	BB int/cont	ER int/cont	Rate source
Tyler Glasnow	Starter	20.9	5.1	5.202	4 / 3.503	5 / 6.814	0 / 1.829	2 / 2.010	all_players_stats_recent_season_blend
Alex Vesia	Middle Reliever	3.9	1.0	1.014	0 / 0.387	1 / 1.340	0 / 0.450	0 / 0.218	all_players_stats_recent_season_blend
Roki Sasaki	Middle Reliever	5.4	1.1	1.351	1 / 0.868	2 / 1.578	0 / 0.502	1 / 0.305	all_players_stats_recent_season_blend
Brock Stewart	Middle Reliever (High Leverage)	3.9	1.1	1.014	2 / 0.522	4 / 0.968	1 / 0.330	1 / 0.246	all_players_stats_recent_season_blend
Tyler Mahle	Starter	20.9	5.0	4.997	6 / 4.443	4 / 4.119	2 / 1.445	3 / 1.853	all_players_stats_recent_season_blend
Grant Holmes	Long Reliever	8.4	2.0	1.985	1 / 1.515	2 / 1.085	1 / 0.966	2 / 0.571	all_players_stats_recent_season_blend
Dylan Dodd	Middle Reliever	3.9	2.0	0.993	2 / 0.707	2 / 1.112	0 / 0.213	1 / 0.354	all_players_stats_recent_season_blend

Audit: NO DEFECT FOUND

Invariant	Verdict	Evidence
blend arithmetic away_runs	PASS	4.0000 + 0.34 x (4.1204 - 4.0000) = 4.0409, recorded 4.0409
blend arithmetic home_runs	PASS	3.0000 + 0.38 x (4.7962 - 3.0000) = 3.6826, recorded 3.6826
blend arithmetic home_win	PASS	0.4916 + 0.00 x (0.6876 - 0.4916) = 0.4916, recorded 0.4916
published rounding faithful to the blend	PASS	blended 4.0409-3.6826 rounds to 5-4, layer published 5-4
published probability on the score's side	PASS	5-4 with p(home) 0.491557
Log5 additivity136 checks	PASS	136 of 136 pass
away club fields exactly one starter	PASS	1 starters in a staff of 4
away staff totals a regulation workload	PASS	27 outs recorded, 27 expected
away starter carries a continuous line	PASS	expected_IP 5.2016
away starter batters faced inside the real band	PASS	Tyler Glasnow projected 20.9 batters faced; real starters run p01-p99 4-30, median 23, mean 21.75
away reliever batters faced inside the real band3 checks	PASS	3 of 3 pass
home club fields exactly one starter	PASS	1 starters in a staff of 3
home staff totals a regulation workload	PASS	27 outs recorded, 27 expected
home starter carries a continuous line	PASS	expected_IP 4.9971
home starter batters faced inside the real band	PASS	Tyler Mahle projected 20.9 batters faced; real starters run p01-p99 4-30, median 23, mean 21.75
home reliever batters faced inside the real band2 checks	PASS	2 of 2 pass
away hitters satisfy every coherence rule	PASS	9 hitters checked; H<=AB, HR<=H, 2B+3B+HR<=H, SO<=AB, continuous line present
home hitters satisfy every coherence rule	PASS	9 hitters checked; H<=AB, HR<=H, 2B+3B+HR<=H, SO<=AB, continuous line present

20261007_MIL@SD — published 6-5, 157 checks, 2 failed

Artifact. 20261007_MIL@SD__20261008T024942Z.json, phase post_start_rerun, generated 2026-10-08T02:49:42, first pitch 2026-10-08T02:00:00.
Published 6 - 5 with p(home) 0.495000. Simulation draw chosen on not recorded; ? of ? draws realized the target exactly.

Stage chain: every stage that touched the score

Stage	Score after	Δ away	Δ home	Cum. away	Cum. home	Changed target	Own adjustment and description
1. `historical_ml`	4.316793501587018 - 3.9583418010956	0	0	0	0	no	Base ML historical model prediction
2. `variance_applied`	4 - 4	-0.3168	+0.0417	-0.3168	+0.0417	yes	Deterministic baseline target (variance factor 1.0)
3. `enhanced_prediction`	4 - 5	0	+1.0000	-0.3168	+1.0417	yes	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run)
4. `kpi_layer`	4 - 5	0	0	-0.3168	+1.0417	no	away -0.1958; home +0.3883 Savant KPI blend applied to the model target
5. `embedded_savant_run_environment`	4 - 5	0	0	-0.3168	+1.0417	no	away +0.4139; home +0.2587 Target score adjusted by temporal Savant lineup quality versus opposing starter quality
6. `fundamental_edge_target_review`	6 - 5	+2.0000	0	+1.6832	+1.0417	yes	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
7. `capped_target`	6 - 5	0	0	+1.6832	+1.0417	no	Capped at max 12 realistic runs
8. `simulation_result`	6 - 5	-	-	-	-	no	Final inning-by-inning simulation result

KPI team layer: the blend, with its arithmetic

Target	Baseline	KPI only	Weight	Blended	Computation	Verdict
`away_runs`	4.0000	3.4240	0.34	3.8042	4.0000 + 0.34 × (3.4240 − 4.0000) = 3.8042	PASS
`home_runs`	5.0000	6.0217	0.38	5.3883	5.0000 + 0.38 × (6.0217 − 5.0000) = 5.3883	PASS
`home_win`	0.5050	0.7071	0.00	0.5050	0.5050 + 0.00 × (0.7071 − 0.5050) = 0.5050	PASS

Published by the layer: away 4, home 5, p(home) 0.505000, continuous away 3.8042 / home 5.3883

Inputs and gates. candidate 20261007T091423Z_47b9db81, value age 10 days, lineup coverage away 1.0000 / home 0.7719, movement cap 2.5 runs.
values_present ok values_fresh ok one_source_bundle ok lineup_coverage ok target_eligible ok movement_within_limit ok published_movement_within_limit ok

What drove the KPI forecast for `away_runs`

intercept 4.2358 + sum of contributions = 3.4240; prediction 3.4240; contributions outside the top listed -0.4461 runs

Feature	Raw	Standardised	Runs contributed	Which KPIs it is built from
`away_sp_ground`	1.0175	1.612	-0.5250	starter group: pitcher_ground_ball_rate_induced
`away_off_contact`	0.3325	1.204	+0.4138	offence group: batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`home_off_power_spread`	0.5785	-0.773	-0.4064	offence group (dispersion across the lineup, not its mean): batter_home_run_rate, batter_bat_speed
`away_pen_share`	0.5149	1.733	-0.2986	composite feature
`home_off_discipline_spread`	0.2140	-2.368	+0.2839	offence group (dispersion across the lineup, not its mean): batter_whiff_rate_vs_fastball, batter_whiff_rate_vs_breaking, batter_whiff_rate_vs_offspeed, batter_strikeout_rate ...
`away_off_contact_spread`	0.8979	0.993	-0.2761	offence group (dispersion across the lineup, not its mean): batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`home_off_top3_contact`	1.0774	1.055	-0.2345	mean of the three best contact bats, a tail a club mean cannot express
`away_off_discipline_spread`	0.7301	1.054	+0.2340	offence group (dispersion across the lineup, not its mean): batter_whiff_rate_vs_fastball, batter_whiff_rate_vs_breaking, batter_whiff_rate_vs_offspeed, batter_strikeout_rate ...
`home_off_contact`	0.3237	1.140	+0.2272	offence group: batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`away_off_ground`	-0.7385	-2.544	+0.2160	offence group: batter_ground_ball_rate_vs_slider, batter_ground_ball_rate_vs_fastball

What drove the KPI forecast for `home_runs`

intercept 4.5238 + sum of contributions = 6.0217; prediction 6.0217; contributions outside the top listed 0.6565 runs

Feature	Raw	Standardised	Runs contributed	Which KPIs it is built from
`away_off_ground`	-0.7385	-2.544	-0.4647	offence group: batter_ground_ball_rate_vs_slider, batter_ground_ball_rate_vs_fastball
`away_off_contact_spread`	0.8979	0.993	-0.4421	offence group (dispersion across the lineup, not its mean): batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`away_off_top3_contact`	1.3386	1.818	-0.4038	mean of the three best contact bats, a tail a club mean cannot express
`away_pen_share`	0.5149	1.733	+0.3796	composite feature
`home_pen_share`	0.4794	1.511	+0.3592	composite feature
`away_sp_velocity`	0.5712	0.742	+0.3413	starter group: pitcher_release_speed
`home_pen_minus_sp`	0.5408	2.027	+0.3017	composite feature
`away_sp_ground`	1.0175	1.612	+0.2835	starter group: pitcher_ground_ball_rate_induced
`home_off_power`	0.2837	1.084	+0.2527	offence group: batter_home_run_rate, batter_bat_speed
`away_pen_minus_sp`	0.5283	1.683	+0.2339	composite feature

Per-player metric movement

MIL against JoJo Romero (LHP)

The pipeline applied one club-wide ×0.951 to every MIL hitter line, identical for the leadoff hitter and the ninth bat.
Opposing staff exposure. These hitters face the starter for 15.467 of 37.766 projected batters (40.9%) and the 3 projected relievers for the remaining 59.1%. Every metric below is the exposure-weighted blend of both matchups; the starter-only figure is shown beside it. Real 2026 game logs put a starter at 21.75 batters faced on average (median 23, p10-p90 17-27) and a whole staff at 37.9: this starter is OUTSIDE that band and this staff is inside its. The per-player columns are the measured alternative: each subject moved by his own KPI pairing. Exposure for a hitter is his projected plate appearances; for the starter it is batters faced.

Player	Pos	Metric	Published → KPI	Δ	League	Own rate	Opponent rate	Log5	Own skill	Opponent	Remainder
Jackson Chourio	LF	Hitswhole staff	1.087 → 1.075	-0.012	0.2169	0.2457n=567 Very good	0.2262n=651 Below average	0.2559	+0.128	+0.041	+0.004 ok
		split	starter x1.0414 at 0.409 share (log5 0.2559 vs his 0.2262) + bullpen x0.9532 at 0.591 share (log5 0.2342 vs pen 0.2063) = x0.9893 \| starter alone would have been +0.045
		Strikeoutswhole staff	1.017 → 1.046	+0.029	0.2212	0.2351n=567 Average	0.2034n=651 Average	0.2165	+0.062	-0.079	-0.004 ok
		split	starter x0.9209 at 0.409 share (log5 0.2165 vs his 0.2034) + bullpen x1.1037 at 0.591 share (log5 0.2595 vs pen 0.2446) = x1.0288 \| starter alone would have been -0.080
		Walkswhole staff	0.362 → 0.353	-0.009	0.0889	0.0839n=567 Average	0.0878n=651 Average	0.0829	-0.022	-0.005	+0.000 ok
		split	starter x0.9876 at 0.409 share (log5 0.0829 vs his 0.0878) + bullpen x0.968 at 0.591 share (log5 0.0813 vs pen 0.0861) = x0.976 \| starter alone would have been -0.004
		Home runswhole staff	0.171 → 0.161	-0.010	0.0304	0.0397n=567 Good	0.0324n=651 Below average	0.0423	+0.042	+0.009	+0.003 ok
		split	starter x1.0658 at 0.409 share (log5 0.0423 vs his 0.0324) + bullpen x0.8591 at 0.591 share (log5 0.0341 vs pen 0.0261) = x0.9437 \| starter alone would have been +0.011
Neutralising Jackson Chourio to the qualified-population mean moves the club's KPI run forecast by 0.2686 runs, and the weighted stage by 0.0913. Interactions make these non-additive across players, so they do not sum to the published movement.
Christian Yelich	DH	Hitswhole staff	0.855 → 0.846	-0.009	0.2169	0.2108n=487 Average	0.2262n=651 Below average	0.2200	-0.025	+0.039	-0.001 ok
		split	starter x1.0434 at 0.409 share (log5 0.2200 vs his 0.2262) + bullpen x0.9511 at 0.591 share (log5 0.2005 vs pen 0.2063) = x0.9889 \| starter alone would have been +0.037
		Strikeoutswhole staff	1.029 → 1.057	+0.028	0.2212	0.2684n=487 Below average	0.2034n=651 Average	0.2480	+0.199	-0.075	-0.011 ok
		split	starter x0.9241 at 0.409 share (log5 0.2480 vs his 0.2034) + bullpen x1.0987 at 0.591 share (log5 0.2949 vs pen 0.2446) = x1.0272 \| starter alone would have been -0.078
		Walkswhole staff	0.421 → 0.411	-0.010	0.0889	0.0983n=487 Average	0.0878n=651 Average	0.0971	+0.040	-0.005	-0.000 ok
		split	starter x0.9878 at 0.409 share (log5 0.0971 vs his 0.0878) + bullpen x0.9685 at 0.591 share (log5 0.0952 vs pen 0.0861) = x0.9764 \| starter alone would have been -0.005
		Home runswhole staff	0.114 → 0.107	-0.006	0.0304	0.0267n=487 Average	0.0324n=651 Below average	0.0284	-0.016	+0.009	-0.001 ok
		split	starter x1.0667 at 0.409 share (log5 0.0284 vs his 0.0324) + bullpen x0.8575 at 0.591 share (log5 0.0229 vs pen 0.0261) = x0.9432 \| starter alone would have been +0.008
Neutralising Christian Yelich to the qualified-population mean moves the club's KPI run forecast by -0.1141 runs, and the weighted stage by -0.0388. Interactions make these non-additive across players, so they do not sum to the published movement.
William Contreras	C	Hitswhole staff	0.940 → 0.930	-0.010	0.2169	0.2303n=630 Good	0.2262n=651 Below average	0.2400	+0.057	+0.039	+0.002 ok
		split	starter x1.0422 at 0.409 share (log5 0.2400 vs his 0.2262) + bullpen x0.9523 at 0.591 share (log5 0.2193 vs pen 0.2063) = x0.9891 \| starter alone would have been +0.040
		Strikeoutswhole staff	0.621 → 0.642	+0.021	0.2212	0.1529n=630 Good	0.2034n=651 Average	0.1396	-0.289	-0.075	+0.019 ok
		split	starter x0.9132 at 0.409 share (log5 0.1396 vs his 0.2034) + bullpen x1.1161 at 0.591 share (log5 0.1706 vs pen 0.2446) = x1.033 \| starter alone would have been -0.054
		Walkswhole staff	0.410 → 0.400	-0.010	0.0889	0.0942n=630 Average	0.0878n=651 Average	0.0931	+0.022	-0.005	-0.000 ok
		split	starter x0.9877 at 0.409 share (log5 0.0931 vs his 0.0878) + bullpen x0.9684 at 0.591 share (log5 0.0913 vs pen 0.0861) = x0.9763 \| starter alone would have been -0.005
		Home runswhole staff	0.122 → 0.115	-0.007	0.0304	0.0291n=630 Average	0.0324n=651 Below average	0.0310	-0.005	+0.009	-0.000 ok
		split	starter x1.0666 at 0.409 share (log5 0.0310 vs his 0.0324) + bullpen x0.8578 at 0.591 share (log5 0.0250 vs pen 0.0261) = x0.9433 \| starter alone would have been +0.008
Neutralising William Contreras to the qualified-population mean moves the club's KPI run forecast by 0.1979 runs, and the weighted stage by 0.0673. Interactions make these non-additive across players, so they do not sum to the published movement.
Brice Turang	2B	Hitswhole staff	0.898 → 0.888	-0.010	0.2169	0.2264n=666 Good	0.2262n=651 Below average	0.2360	+0.040	+0.039	+0.001 ok
		split	starter x1.0425 at 0.409 share (log5 0.2360 vs his 0.2262) + bullpen x0.952 at 0.591 share (log5 0.2156 vs pen 0.2063) = x0.9891 \| starter alone would have been +0.038
		Strikeoutswhole staff	0.955 → 0.982	+0.027	0.2212	0.2473n=666 Below average	0.2034n=651 Average	0.2281	+0.109	-0.074	-0.006 ok
		split	starter x0.9221 at 0.409 share (log5 0.2281 vs his 0.2034) + bullpen x1.1018 at 0.591 share (log5 0.2725 vs pen 0.2446) = x1.0282 \| starter alone would have been -0.074
		Walkswhole staff	0.501 → 0.489	-0.011	0.0889	0.1243n=666 Very good	0.0878n=651 Average	0.1228	+0.148	-0.005	-0.002 ok
		split	starter x0.9881 at 0.409 share (log5 0.1228 vs his 0.0878) + bullpen x0.9694 at 0.591 share (log5 0.1205 vs pen 0.0861) = x0.9771 \| starter alone would have been -0.006
		Home runswhole staff	0.116 → 0.109	-0.006	0.0304	0.0312n=666 Average	0.0324n=651 Below average	0.0333	+0.003	+0.009	+0.000 ok
		split	starter x1.0664 at 0.409 share (log5 0.0333 vs his 0.0324) + bullpen x0.8581 at 0.591 share (log5 0.0268 vs pen 0.0261) = x0.9434 \| starter alone would have been +0.008
Neutralising Brice Turang to the qualified-population mean moves the club's KPI run forecast by 0.3509 runs, and the weighted stage by 0.1193. Interactions make these non-additive across players, so they do not sum to the published movement.
Jake Bauers	1B	Hitswhole staff	0.803 → 0.794	-0.009	0.2169	0.2179n=567 Average	0.2262n=651 Below average	0.2273	+0.004	+0.039	+0.000 ok
		split	starter x1.0429 at 0.409 share (log5 0.2273 vs his 0.2262) + bullpen x0.9515 at 0.591 share (log5 0.2073 vs pen 0.2063) = x0.9889 \| starter alone would have been +0.035
		Strikeoutswhole staff	1.016 → 1.043	+0.028	0.2212	0.2676n=567 Below average	0.2034n=651 Average	0.2473	+0.193	-0.074	-0.011 ok
		split	starter x0.924 at 0.409 share (log5 0.2473 vs his 0.2034) + bullpen x1.0988 at 0.591 share (log5 0.2941 vs pen 0.2446) = x1.0272 \| starter alone would have been -0.077
		Walkswhole staff	0.554 → 0.542	-0.012	0.0889	0.1510n=567 Very good	0.0878n=651 Average	0.1492	+0.257	-0.005	-0.003 ok
		split	starter x0.9885 at 0.409 share (log5 0.1492 vs his 0.0878) + bullpen x0.9703 at 0.591 share (log5 0.1465 vs pen 0.0861) = x0.9778 \| starter alone would have been -0.006
		Home runswhole staff	0.170 → 0.160	-0.010	0.0304	0.0397n=567 Good	0.0324n=651 Below average	0.0423	+0.039	+0.008	+0.003 ok
		split	starter x1.0658 at 0.409 share (log5 0.0423 vs his 0.0324) + bullpen x0.8591 at 0.591 share (log5 0.0341 vs pen 0.0261) = x0.9437 \| starter alone would have been +0.011
Neutralising Jake Bauers to the qualified-population mean moves the club's KPI run forecast by 0.0609 runs, and the weighted stage by 0.0207. Interactions make these non-additive across players, so they do not sum to the published movement.
Cooper Pratt	SS	Hitswhole staff	0.833 → 0.823	-0.009	0.2169	0.2255n=258 Good	0.2262n=651 Below average	0.2351	+0.033	+0.035	+0.001 ok
		split	starter x1.0425 at 0.409 share (log5 0.2351 vs his 0.2262) + bullpen x0.952 at 0.591 share (log5 0.2147 vs pen 0.2063) = x0.9891 \| starter alone would have been +0.035
		Strikeoutswhole staff	0.700 → 0.722	+0.022	0.2212	0.1851n=258 Good	0.2034n=651 Average	0.1696	-0.136	-0.067	+0.009 ok
		split	starter x0.9162 at 0.409 share (log5 0.1696 vs his 0.2034) + bullpen x1.1112 at 0.591 share (log5 0.2057 vs pen 0.2446) = x1.0313 \| starter alone would have been -0.059
		Walkswhole staff	0.307 → 0.300	-0.007	0.0889	0.0917n=258 Average	0.0878n=651 Average	0.0906	+0.011	-0.004	-0.000 ok
		split	starter x0.9877 at 0.409 share (log5 0.0906 vs his 0.0878) + bullpen x0.9683 at 0.591 share (log5 0.0888 vs pen 0.0861) = x0.9762 \| starter alone would have been -0.004
		Home runswhole staff	0.074 → 0.070	-0.004	0.0304	0.0211n=258 Below average	0.0324n=651 Below average	0.0225	-0.035	+0.008	-0.002 ok
		split	starter x1.0672 at 0.409 share (log5 0.0225 vs his 0.0324) + bullpen x0.8568 at 0.591 share (log5 0.0181 vs pen 0.0261) = x0.943 \| starter alone would have been +0.005
Neutralising Cooper Pratt to the qualified-population mean moves the club's KPI run forecast by 0.1071 runs, and the weighted stage by 0.0364. Interactions make these non-additive across players, so they do not sum to the published movement.
Sal Frelick	RF	Hitswhole staff	0.810 → 0.801	-0.009	0.2169	0.2106n=374 Below average	0.2262n=651 Below average	0.2197	-0.023	+0.034	-0.001 ok
		split	starter x1.0434 at 0.409 share (log5 0.2197 vs his 0.2262) + bullpen x0.9511 at 0.591 share (log5 0.2003 vs pen 0.2063) = x0.9889 \| starter alone would have been +0.035
		Strikeoutswhole staff	0.418 → 0.432	+0.015	0.2212	0.1106n=374 Very good	0.2034n=651 Average	0.1006	-0.408	-0.066	+0.029 ok
		split	starter x0.9092 at 0.409 share (log5 0.1006 vs his 0.2034) + bullpen x1.1226 at 0.591 share (log5 0.1242 vs pen 0.2446) = x1.0352 \| starter alone would have been -0.038
		Walkswhole staff	0.282 → 0.276	-0.007	0.0889	0.0843n=374 Average	0.0878n=651 Average	0.0833	-0.017	-0.004	+0.000 ok
		split	starter x0.9876 at 0.409 share (log5 0.0833 vs his 0.0878) + bullpen x0.968 at 0.591 share (log5 0.0817 vs pen 0.0861) = x0.976 \| starter alone would have been -0.004
		Home runswhole staff	0.068 → 0.064	-0.004	0.0304	0.0173n=374 Poor	0.0324n=651 Below average	0.0184	-0.048	+0.007	-0.003 ok
		split	starter x1.0674 at 0.409 share (log5 0.0184 vs his 0.0324) + bullpen x0.8563 at 0.591 share (log5 0.0148 vs pen 0.0261) = x0.9427 \| starter alone would have been +0.005
Neutralising Sal Frelick to the qualified-population mean moves the club's KPI run forecast by 0.3225 runs, and the weighted stage by 0.1097. Interactions make these non-additive across players, so they do not sum to the published movement.
Garrett Mitchell	CF	Hitswhole staff	0.792 → 0.783	-0.009	0.2169	0.2215n=519 Average	0.2262n=651 Below average	0.2310	+0.018	+0.036	+0.001 ok
		split	starter x1.0427 at 0.409 share (log5 0.2310 vs his 0.2262) + bullpen x0.9517 at 0.591 share (log5 0.2108 vs pen 0.2063) = x0.989 \| starter alone would have been +0.034
		Strikeoutswhole staff	1.110 → 1.138	+0.028	0.2212	0.3061n=519 Poor	0.2034n=651 Average	0.2840	+0.324	-0.068	-0.017 ok
		split	starter x0.9277 at 0.409 share (log5 0.2840 vs his 0.2034) + bullpen x1.0932 at 0.591 share (log5 0.3346 vs pen 0.2446) = x1.0254 \| starter alone would have been -0.080
		Walkswhole staff	0.411 → 0.402	-0.010	0.0889	0.1091n=519 Good	0.0878n=651 Average	0.1077	+0.077	-0.004	-0.001 ok
		split	starter x0.9879 at 0.409 share (log5 0.1077 vs his 0.0878) + bullpen x0.9689 at 0.591 share (log5 0.1057 vs pen 0.0861) = x0.9767 \| starter alone would have been -0.005
		Home runswhole staff	0.086 → 0.081	-0.005	0.0304	0.0243n=519 Below average	0.0324n=651 Below average	0.0259	-0.023	+0.008	-0.002 ok
		split	starter x1.0669 at 0.409 share (log5 0.0259 vs his 0.0324) + bullpen x0.8572 at 0.591 share (log5 0.0208 vs pen 0.0261) = x0.9431 \| starter alone would have been +0.006
Neutralising Garrett Mitchell to the qualified-population mean moves the club's KPI run forecast by -0.1415 runs, and the weighted stage by -0.0481. Interactions make these non-additive across players, so they do not sum to the published movement.
Joey Ortiz	3B	Hitswhole staff	0.785 → 0.777	-0.009	0.2169	0.2271n=403 Good	0.2262n=651 Below average	0.2367	+0.037	+0.034	+0.001 ok
		split	starter x1.0424 at 0.409 share (log5 0.2367 vs his 0.2262) + bullpen x0.9521 at 0.591 share (log5 0.2162 vs pen 0.2063) = x0.9891 \| starter alone would have been +0.033
		Strikeoutswhole staff	0.747 → 0.770	+0.023	0.2212	0.2010n=403 Average	0.2034n=651 Average	0.1844	-0.073	-0.064	+0.004 ok
		split	starter x0.9177 at 0.409 share (log5 0.1844 vs his 0.2034) + bullpen x1.1088 at 0.591 share (log5 0.2228 vs pen 0.2446) = x1.0305 \| starter alone would have been -0.061
		Walkswhole staff	0.278 → 0.271	-0.007	0.0889	0.0873n=403 Average	0.0878n=651 Average	0.0862	-0.006	-0.004	+0.000 ok
		split	starter x0.9876 at 0.409 share (log5 0.0862 vs his 0.0878) + bullpen x0.9681 at 0.591 share (log5 0.0845 vs pen 0.0861) = x0.9761 \| starter alone would have been -0.003
		Home runswhole staff	0.086 → 0.081	-0.005	0.0304	0.0255n=403 Below average	0.0324n=651 Below average	0.0272	-0.018	+0.007	-0.001 ok
		split	starter x1.0668 at 0.409 share (log5 0.0272 vs his 0.0324) + bullpen x0.8573 at 0.591 share (log5 0.0219 vs pen 0.0261) = x0.9431 \| starter alone would have been +0.006
Neutralising Joey Ortiz to the qualified-population mean moves the club's KPI run forecast by 0.0813 runs, and the weighted stage by 0.0276. Interactions make these non-additive across players, so they do not sum to the published movement.
STARTER Robert Gasser15.7744 BF	Starter	Hits	2.886 → 2.934	+0.048	0.2169	0.2100n=462 Average	0.22047/9 qualified lineup mean	0.2135	-0.108	+0.056	-0.001 ok
		Strikeouts	3.300 → 3.185	-0.116	0.2212	0.2227n=462 Average	0.21347/9 qualified lineup mean	0.2149	+0.023	-0.123	-0.001 ok
		Walks	1.658 → 1.738	+0.081	0.0889	0.0775n=462 Good	0.09327/9 qualified lineup mean	0.0813	-0.180	+0.067	-0.008 ok
		Home runs	0.483 → 0.512	+0.029	0.0304	0.0314n=462 Average	0.03227/9 qualified lineup mean	0.0333	+0.017	+0.029	+0.001 ok
RELIEVER Aaron Ashby6.3304 BF	Setup Man	Hits	0.861 → 0.875	+0.014	0.2169	0.2024n=372 Good	0.22047/9 qualified lineup mean	0.2057	-0.092	+0.023	-0.001 ok
		Strikeouts	1.556 → 1.506	-0.050	0.2212	0.2869n=372 Very good	0.21347/9 qualified lineup mean	0.2776	+0.416	-0.049	-0.009 ok
		Walks	0.417 → 0.437	+0.020	0.0889	0.1039n=372 Below average	0.09327/9 qualified lineup mean	0.1088	+0.095	+0.027	+0.004 ok
		Home runs	0.081 → 0.086	+0.005	0.0304	0.0251n=372 Good	0.03227/9 qualified lineup mean	0.0266	-0.033	+0.011	-0.002 ok
RELIEVER Trevor Megill5.3848 BF	Closer	Hits	0.540 → 0.549	+0.009	0.2169	0.1936n=242 Very good	0.22047/9 qualified lineup mean	0.1969	-0.125	+0.019	-0.002 ok
		Strikeouts	1.073 → 1.038	-0.035	0.2212	0.2692n=242 Good	0.21347/9 qualified lineup mean	0.2603	+0.259	-0.042	-0.006 ok
		Walks	0.177 → 0.185	+0.009	0.0889	0.0726n=242 Good	0.09327/9 qualified lineup mean	0.0762	-0.088	+0.023	-0.004 ok
		Home runs	0.012 → 0.013	+0.001	0.0304	0.0212n=242 Very good	0.03227/9 qualified lineup mean	0.0224	-0.050	+0.010	-0.003 ok
RELIEVER Antonio Senzatela4.114 BF	Middle Reliever (High Leverage)	Hits	1.007 → 1.023	+0.016	0.2169	0.2274n=303 Below average	0.22047/9 qualified lineup mean	0.2311	+0.043	+0.015	+0.001 ok
		Strikeouts	0.794 → 0.765	-0.029	0.2212	0.1977n=303 Below average	0.21347/9 qualified lineup mean	0.1905	-0.097	-0.032	+0.003 ok
		Walks	0.397 → 0.416	+0.019	0.0889	0.0847n=303 Average	0.09327/9 qualified lineup mean	0.0888	-0.017	+0.018	-0.001 ok
		Home runs	0.057 → 0.060	+0.003	0.0304	0.0261n=303 Good	0.03227/9 qualified lineup mean	0.0277	-0.018	+0.007	-0.001 ok
RELIEVER Abner Uribe2.1864 BF	Reliever	Hits	0.441 → 0.449	+0.007	0.2169	0.2018n=230 Good	0.22047/9 qualified lineup mean	0.2052	-0.033	+0.008	-0.000 ok
		Strikeouts	0.555 → 0.536	-0.019	0.2212	0.2538n=230 Good	0.21347/9 qualified lineup mean	0.2453	+0.071	-0.017	-0.002 ok
		Walks	0.181 → 0.189	+0.009	0.0889	0.0826n=230 Average	0.09327/9 qualified lineup mean	0.0866	-0.014	+0.009	-0.001 ok
		Home runs	0.054 → 0.057	+0.003	0.0304	0.0245n=230 Very good	0.03227/9 qualified lineup mean	0.0260	-0.013	+0.004	-0.001 ok
RELIEVER Chad Patrick2.0067 BF	Reliever	Hits	0.400 → 0.406	+0.007	0.2169	0.1991n=438 Good	0.22047/9 qualified lineup mean	0.2025	-0.036	+0.007	-0.000 ok
		Strikeouts	0.429 → 0.414	-0.015	0.2212	0.2137n=438 Average	0.21347/9 qualified lineup mean	0.2061	-0.015	-0.016	+0.000 ok
		Walks	0.209 → 0.219	+0.010	0.0889	0.1041n=438 Below average	0.09327/9 qualified lineup mean	0.1090	+0.030	+0.009	+0.001 ok
		Home runs	0.056 → 0.059	+0.003	0.0304	0.0278n=438 Average	0.03227/9 qualified lineup mean	0.0294	-0.005	+0.004	-0.000 ok
RELIEVER DL Hall0.7894 BF	Reliever	Hits	0.137 → 0.140	+0.002	0.2169	0.1738n=234 Very good	0.22047/9 qualified lineup mean	0.1768	-0.034	+0.003	-0.000 ok
		Strikeouts	0.200 → 0.194	-0.007	0.2212	0.2538n=234 Good	0.21347/9 qualified lineup mean	0.2452	+0.026	-0.006	-0.001 ok
		Walks	0.097 → 0.101	+0.004	0.0889	0.1226n=234 Poor	0.09327/9 qualified lineup mean	0.1283	+0.027	+0.003	+0.001 ok
		Home runs	0.021 → 0.023	+0.001	0.0304	0.0273n=234 Average	0.03227/9 qualified lineup mean	0.0289	-0.002	+0.001	-0.000 ok
RELIEVER JoJo Romero0.5819 BF	Reliever	Hits	0.126 → 0.128	+0.002	0.2169	0.2159n=233 Average	0.22047/9 qualified lineup mean	0.2194	-0.001	+0.002	0 ok
		Strikeouts	0.131 → 0.127	-0.005	0.2212	0.2259n=233 Average	0.21347/9 qualified lineup mean	0.2180	+0.003	-0.004	-0.000 ok
		Walks	0.046 → 0.048	+0.002	0.0889	0.0794n=233 Average	0.09327/9 qualified lineup mean	0.0833	-0.005	+0.003	-0.000 ok
		Home runs	0.018 → 0.019	+0.001	0.0304	0.0303n=233 Average	0.03227/9 qualified lineup mean	0.0321	-0.000	+0.001	0 ok

SD against Joe Musgrove (RHP)

The pipeline applied one club-wide ×1.0777 to every SD hitter line, identical for the leadoff hitter and the ninth bat.
Opposing staff exposure. These hitters face the starter for 15.7744 of 37.168 projected batters (42.4%) and the 3 projected relievers for the remaining 57.6%. Every metric below is the exposure-weighted blend of both matchups; the starter-only figure is shown beside it. Real 2026 game logs put a starter at 21.75 batters faced on average (median 23, p10-p90 17-27) and a whole staff at 37.9: this starter is OUTSIDE that band and this staff is inside its. The per-player columns are the measured alternative: each subject moved by his own KPI pairing. Exposure for a hitter is his projected plate appearances; for the starter it is batters faced.

Player	Pos	Metric	Published → KPI	Δ	League	Own rate	Opponent rate	Log5	Own skill	Opponent	Remainder
Fernando Tatis Jr.	RF	Hitswhole staff	1.027 → 0.979	-0.048	0.2169	0.2416n=704 Very good	0.2100n=462 Average	0.2342	+0.110	-0.031	-0.002 ok
		split	starter x0.9693 at 0.424 share (log5 0.2342 vs his 0.2100) + bullpen x0.9421 at 0.576 share (log5 0.2276 vs pen 0.2039) = x0.9536 \| starter alone would have been -0.032
		Strikeoutswhole staff	0.828 → 0.901	+0.072	0.2212	0.1870n=704 Good	0.2227n=462 Average	0.1883	-0.152	+0.007	-0.001 ok
		split	starter x1.007 at 0.424 share (log5 0.1883 vs his 0.2227) + bullpen x1.147 at 0.576 share (log5 0.2145 vs pen 0.2522) = x1.0876 \| starter alone would have been +0.006
		Walkswhole staff	0.452 → 0.431	-0.021	0.0889	0.0906n=704 Average	0.0775n=462 Good	0.0790	+0.008	-0.051	-0.001 ok
		split	starter x0.8717 at 0.424 share (log5 0.0790 vs his 0.0775) + bullpen x1.0143 at 0.576 share (log5 0.0919 vs pen 0.0902) = x0.9538 \| starter alone would have been -0.058
		Home runswhole staff	0.156 → 0.141	-0.014	0.0304	0.0341n=704 Average	0.0314n=462 Average	0.0353	+0.017	+0.005	+0.001 ok
		split	starter x1.0346 at 0.424 share (log5 0.0353 vs his 0.0314) + bullpen x0.8142 at 0.576 share (log5 0.0278 vs pen 0.0247) = x0.9077 \| starter alone would have been +0.005
Neutralising Fernando Tatis Jr. to the qualified-population mean moves the club's KPI run forecast by 0.2913 runs, and the weighted stage by 0.1107. Interactions make these non-additive across players, so they do not sum to the published movement.
Jackson Merrill	CF	Hitswhole staff	0.932 → 0.888	-0.044	0.2169	0.2277n=664 Good	0.2100n=462 Average	0.2206	+0.046	-0.029	-0.001 ok
		split	starter x0.9688 at 0.424 share (log5 0.2206 vs his 0.2100) + bullpen x0.9411 at 0.576 share (log5 0.2143 vs pen 0.2039) = x0.9529 \| starter alone would have been -0.029
		Strikeoutswhole staff	1.115 → 1.206	+0.090	0.2212	0.2396n=664 Average	0.2227n=462 Average	0.2412	+0.078	+0.006	+0.000 ok
		split	starter x1.0065 at 0.424 share (log5 0.2412 vs his 0.2227) + bullpen x1.1362 at 0.576 share (log5 0.2722 vs pen 0.2522) = x1.0812 \| starter alone would have been +0.007
		Walkswhole staff	0.280 → 0.267	-0.013	0.0889	0.0659n=664 Below average	0.0775n=462 Good	0.0572	-0.098	-0.049	+0.012 ok
		split	starter x0.8686 at 0.424 share (log5 0.0572 vs his 0.0775) + bullpen x1.0147 at 0.576 share (log5 0.0668 vs pen 0.0902) = x0.9527 \| starter alone would have been -0.037
		Home runswhole staff	0.152 → 0.138	-0.014	0.0304	0.0367n=664 Good	0.0314n=462 Average	0.0379	+0.027	+0.004	+0.001 ok
		split	starter x1.0346 at 0.424 share (log5 0.0379 vs his 0.0314) + bullpen x0.8146 at 0.576 share (log5 0.0299 vs pen 0.0247) = x0.908 \| starter alone would have been +0.005
Neutralising Jackson Merrill to the qualified-population mean moves the club's KPI run forecast by 0.4964 runs, and the weighted stage by 0.1886. Interactions make these non-additive across players, so they do not sum to the published movement.
Manny Machado	3B	Hitswhole staff	0.867 → 0.825	-0.042	0.2169	0.2000n=677 Poor	0.2100n=462 Average	0.1936	-0.070	-0.029	+0.002 ok
		split	starter x0.9677 at 0.424 share (log5 0.1936 vs his 0.2100) + bullpen x0.9391 at 0.576 share (log5 0.1878 vs pen 0.2039) = x0.9512 \| starter alone would have been -0.028
		Strikeoutswhole staff	0.854 → 0.926	+0.072	0.2212	0.2133n=677 Average	0.2227n=462 Average	0.2147	-0.033	+0.006	-0.000 ok
		split	starter x1.0068 at 0.424 share (log5 0.2147 vs his 0.2227) + bullpen x1.1416 at 0.576 share (log5 0.2435 vs pen 0.2522) = x1.0844 \| starter alone would have been +0.006
		Walkswhole staff	0.420 → 0.401	-0.019	0.0889	0.1088n=677 Good	0.0775n=462 Good	0.0951	+0.083	-0.048	-0.009 ok
		split	starter x0.8739 at 0.424 share (log5 0.0951 vs his 0.0775) + bullpen x1.014 at 0.576 share (log5 0.1103 vs pen 0.0902) = x0.9545 \| starter alone would have been -0.053
		Home runswhole staff	0.172 → 0.156	-0.016	0.0304	0.0404n=677 Good	0.0314n=462 Average	0.0418	+0.042	+0.004	+0.001 ok
		split	starter x1.0344 at 0.424 share (log5 0.0418 vs his 0.0314) + bullpen x0.8151 at 0.576 share (log5 0.0329 vs pen 0.0247) = x0.9082 \| starter alone would have been +0.006
Neutralising Manny Machado to the qualified-population mean moves the club's KPI run forecast by 0.1882 runs, and the weighted stage by 0.0715. Interactions make these non-additive across players, so they do not sum to the published movement.
Ty France	1B	Hitswhole staff	0.977 → 0.932	-0.045	0.2169	0.2398n=500 Very good	0.2100n=462 Average	0.2324	+0.095	-0.029	-0.002 ok
		split	starter x0.9692 at 0.424 share (log5 0.2324 vs his 0.2100) + bullpen x0.942 at 0.576 share (log5 0.2259 vs pen 0.2039) = x0.9535 \| starter alone would have been -0.030
		Strikeoutswhole staff	0.906 → 0.982	+0.076	0.2212	0.2165n=500 Average	0.2227n=462 Average	0.2179	-0.020	+0.006	-0.000 ok
		split	starter x1.0067 at 0.424 share (log5 0.2179 vs his 0.2227) + bullpen x1.1409 at 0.576 share (log5 0.2470 vs pen 0.2522) = x1.0839 \| starter alone would have been +0.006
		Walkswhole staff	0.297 → 0.283	-0.014	0.0889	0.0769n=500 Below average	0.0775n=462 Good	0.0669	-0.050	-0.048	+0.006 ok
		split	starter x0.87 at 0.424 share (log5 0.0669 vs his 0.0775) + bullpen x1.0145 at 0.576 share (log5 0.0780 vs pen 0.0902) = x0.9532 \| starter alone would have been -0.039
		Home runswhole staff	0.153 → 0.139	-0.014	0.0304	0.0380n=500 Good	0.0314n=462 Average	0.0393	+0.032	+0.004	+0.001 ok
		split	starter x1.0345 at 0.424 share (log5 0.0393 vs his 0.0314) + bullpen x0.8148 at 0.576 share (log5 0.0309 vs pen 0.0247) = x0.908 \| starter alone would have been +0.005
Neutralising Ty France to the qualified-population mean moves the club's KPI run forecast by 0.2022 runs, and the weighted stage by 0.0768. Interactions make these non-additive across players, so they do not sum to the published movement.
Austin Hays	DH	Hitswhole staff	0.807 → 0.768	-0.039	0.2169	0.2169n=0 averagebelow floor	0.2100n=462 Average	0.2100	0	-0.025	0 ok
		split	starter x0.9683 at 0.424 share (log5 0.2100 vs his 0.2100) + bullpen x0.9403 at 0.576 share (log5 0.2039 vs pen 0.2039) = x0.9522 \| starter alone would have been -0.025
		Strikeoutswhole staff	0.949 → 1.028	+0.079	0.2212	0.2212n=0 averagebelow floor	0.2227n=462 Average	0.2227	0	+0.005	0 ok
		split	starter x1.0067 at 0.424 share (log5 0.2227 vs his 0.2227) + bullpen x1.1399 at 0.576 share (log5 0.2522 vs pen 0.2522) = x1.0834 \| starter alone would have been +0.006
		Walkswhole staff	0.234 → 0.223	-0.011	0.0889	0.0889n=0 averagebelow floor	0.0775n=462 Good	0.0775	0	-0.042	0 ok
		split	starter x0.8715 at 0.424 share (log5 0.0775 vs his 0.0775) + bullpen x1.0143 at 0.576 share (log5 0.0902 vs pen 0.0902) = x0.9537 \| starter alone would have been -0.030
		Home runswhole staff	0.110 → 0.100	-0.010	0.0304	0.0304n=0 averagebelow floor	0.0314n=462 Average	0.0314	0	+0.004	0 ok
		split	starter x1.0348 at 0.424 share (log5 0.0314 vs his 0.0314) + bullpen x0.8136 at 0.576 share (log5 0.0247 vs pen 0.0247) = x0.9075 \| starter alone would have been +0.004
Neutralising Austin Hays to the qualified-population mean moves the club's KPI run forecast by -0.0600 runs, and the weighted stage by -0.0228. Interactions make these non-additive across players, so they do not sum to the published movement.
Xander Bogaerts	SS	Hitswhole staff	0.817 → 0.777	-0.040	0.2169	0.2059n=606 Below average	0.2100n=462 Average	0.1993	-0.042	-0.027	+0.001 ok
		split	starter x0.9679 at 0.424 share (log5 0.1993 vs his 0.2100) + bullpen x0.9395 at 0.576 share (log5 0.1935 vs pen 0.2039) = x0.9516 \| starter alone would have been -0.026
		Strikeoutswhole staff	0.693 → 0.754	+0.061	0.2212	0.1829n=606 Good	0.2227n=462 Average	0.1842	-0.149	+0.006	-0.001 ok
		split	starter x1.007 at 0.424 share (log5 0.1842 vs his 0.2227) + bullpen x1.1478 at 0.576 share (log5 0.2100 vs pen 0.2522) = x1.088 \| starter alone would have been +0.005
		Walkswhole staff	0.412 → 0.393	-0.018	0.0889	0.1208n=606 Very good	0.0775n=462 Good	0.1057	+0.124	-0.044	-0.014 ok
		split	starter x0.8754 at 0.424 share (log5 0.1057 vs his 0.0775) + bullpen x1.0138 at 0.576 share (log5 0.1225 vs pen 0.0902) = x0.9551 \| starter alone would have been -0.051
		Home runswhole staff	0.096 → 0.087	-0.009	0.0304	0.0242n=606 Below average	0.0314n=462 Average	0.0250	-0.024	+0.004	-0.001 ok
		split	starter x1.035 at 0.424 share (log5 0.0250 vs his 0.0314) + bullpen x0.8126 at 0.576 share (log5 0.0196 vs pen 0.0247) = x0.907 \| starter alone would have been +0.003
Neutralising Xander Bogaerts to the qualified-population mean moves the club's KPI run forecast by 0.2708 runs, and the weighted stage by 0.1029. Interactions make these non-additive across players, so they do not sum to the published movement.
Luis Campusano	C	Hitswhole staff	0.775 → 0.738	-0.037	0.2169	0.2215n=248 Average	0.2100n=462 Average	0.2146	+0.018	-0.027	-0.000 ok
		split	starter x0.9685 at 0.424 share (log5 0.2146 vs his 0.2100) + bullpen x0.9406 at 0.576 share (log5 0.2084 vs pen 0.2039) = x0.9524 \| starter alone would have been -0.024
		Strikeoutswhole staff	0.803 → 0.870	+0.068	0.2212	0.2149n=248 Average	0.2227n=462 Average	0.2164	-0.025	+0.006	-0.000 ok
		split	starter x1.0067 at 0.424 share (log5 0.2164 vs his 0.2227) + bullpen x1.1412 at 0.576 share (log5 0.2453 vs pen 0.2522) = x1.0841 \| starter alone would have been +0.005
		Walkswhole staff	0.348 → 0.332	-0.016	0.0889	0.1106n=248 Good	0.0775n=462 Good	0.0967	+0.085	-0.045	-0.010 ok
		split	starter x0.8741 at 0.424 share (log5 0.0967 vs his 0.0775) + bullpen x1.0139 at 0.576 share (log5 0.1121 vs pen 0.0902) = x0.9546 \| starter alone would have been -0.044
		Home runswhole staff	0.117 → 0.106	-0.011	0.0304	0.0313n=248 Average	0.0314n=462 Average	0.0324	+0.004	+0.004	+0.000 ok
		split	starter x1.0347 at 0.424 share (log5 0.0324 vs his 0.0314) + bullpen x0.8137 at 0.576 share (log5 0.0255 vs pen 0.0247) = x0.9075 \| starter alone would have been +0.004
Neutralising Luis Campusano to the qualified-population mean moves the club's KPI run forecast by 0.3191 runs, and the weighted stage by 0.1213. Interactions make these non-additive across players, so they do not sum to the published movement.
Jake Cronenworth	2B	Hitswhole staff	0.712 → 0.677	-0.034	0.2169	0.2086n=425 Below average	0.2100n=462 Average	0.2019	-0.030	-0.025	+0.001 ok
		split	starter x0.968 at 0.424 share (log5 0.2019 vs his 0.2100) + bullpen x0.9397 at 0.576 share (log5 0.1960 vs pen 0.2039) = x0.9517 \| starter alone would have been -0.023
		Strikeoutswhole staff	0.883 → 0.956	+0.073	0.2212	0.2275n=425 Average	0.2227n=462 Average	0.2290	+0.023	+0.005	+0.000 ok
		split	starter x1.0066 at 0.424 share (log5 0.2290 vs his 0.2227) + bullpen x1.1386 at 0.576 share (log5 0.2591 vs pen 0.2522) = x1.0826 \| starter alone would have been +0.006
		Walkswhole staff	0.307 → 0.293	-0.014	0.0889	0.0893n=425 Average	0.0775n=462 Good	0.0778	+0.001	-0.042	-0.000 ok
		split	starter x0.8715 at 0.424 share (log5 0.0778 vs his 0.0775) + bullpen x1.0143 at 0.576 share (log5 0.0906 vs pen 0.0902) = x0.9537 \| starter alone would have been -0.039
		Home runswhole staff	0.080 → 0.073	-0.007	0.0304	0.0218n=425 Below average	0.0314n=462 Average	0.0226	-0.031	+0.004	-0.001 ok
		split	starter x1.0351 at 0.424 share (log5 0.0226 vs his 0.0314) + bullpen x0.8122 at 0.576 share (log5 0.0177 vs pen 0.0247) = x0.9068 \| starter alone would have been +0.003
Neutralising Jake Cronenworth to the qualified-population mean moves the club's KPI run forecast by 0.4558 runs, and the weighted stage by 0.1732. Interactions make these non-additive across players, so they do not sum to the published movement.
Samad Taylor	LF	Hitswhole staff	0.720 → 0.685	-0.034	0.2169	0.2169n=0 averagebelow floor	0.2100n=462 Average	0.2100	0	-0.024	0 ok
		split	starter x0.9683 at 0.424 share (log5 0.2100 vs his 0.2100) + bullpen x0.9403 at 0.576 share (log5 0.2039 vs pen 0.2039) = x0.9522 \| starter alone would have been -0.023
		Strikeoutswhole staff	0.870 → 0.942	+0.072	0.2212	0.2212n=0 averagebelow floor	0.2227n=462 Average	0.2227	0	+0.005	0 ok
		split	starter x1.0067 at 0.424 share (log5 0.2227 vs his 0.2227) + bullpen x1.1399 at 0.576 share (log5 0.2522 vs pen 0.2522) = x1.0834 \| starter alone would have been +0.006
		Walkswhole staff	0.378 → 0.360	-0.018	0.0889	0.0889n=0 averagebelow floor	0.0775n=462 Good	0.0775	0	-0.041	0 ok
		split	starter x0.8715 at 0.424 share (log5 0.0775 vs his 0.0775) + bullpen x1.0143 at 0.576 share (log5 0.0902 vs pen 0.0902) = x0.9537 \| starter alone would have been -0.049
		Home runswhole staff	0.077 → 0.070	-0.007	0.0304	0.0304n=0 averagebelow floor	0.0314n=462 Average	0.0314	0	+0.004	0 ok
		split	starter x1.0348 at 0.424 share (log5 0.0314 vs his 0.0314) + bullpen x0.8136 at 0.576 share (log5 0.0247 vs pen 0.0247) = x0.9075 \| starter alone would have been +0.003
Neutralising Samad Taylor to the qualified-population mean moves the club's KPI run forecast by 0.0034 runs, and the weighted stage by 0.0013. Interactions make these non-additive across players, so they do not sum to the published movement.
STARTER Walker Buehler15.467 BF	Starter	Hits	3.388 → 3.502	+0.115	0.2169	0.2262n=651 Below average	0.22439/9 qualified lineup mean	0.2338	+0.144	+0.115	+0.004 ok
		Strikeouts	3.233 → 3.224	-0.008	0.2212	0.2034n=651 Average	0.22079/9 qualified lineup mean	0.2029	-0.275	-0.009	+0.001 ok
		Walks	1.264 → 1.465	+0.201	0.0889	0.0878n=651 Average	0.10309/9 qualified lineup mean	0.1018	-0.017	+0.218	-0.002 ok
		Home runs	0.817 → 0.771	-0.046	0.0304	0.0324n=651 Below average	0.02879/9 qualified lineup mean	0.0306	+0.031	-0.026	-0.002 ok
RELIEVER Adrian Morejon9.292 BF	Closer	Hits	1.069 → 1.106	+0.037	0.2169	0.1993n=352 Good	0.22439/9 qualified lineup mean	0.2063	-0.163	+0.069	-0.004 ok
		Strikeouts	1.434 → 1.431	-0.003	0.2212	0.2886n=352 Very good	0.22079/9 qualified lineup mean	0.2880	+0.626	-0.005	-0.001 ok
		Walks	0.419 → 0.486	+0.067	0.0889	0.0824n=352 Average	0.10309/9 qualified lineup mean	0.0956	-0.060	+0.131	-0.009 ok
		Home runs	0.108 → 0.102	-0.006	0.0304	0.0220n=352 Very good	0.02879/9 qualified lineup mean	0.0208	-0.078	-0.016	+0.004 ok
RELIEVER Jason Adam5.6639 BF	Setup Man	Hits	0.745 → 0.770	+0.025	0.2169	0.2150n=142 Average	0.22439/9 qualified lineup mean	0.2224	-0.010	+0.042	-0.000 ok
		Strikeouts	0.475 → 0.473	-0.001	0.2212	0.1789n=142 Below average	0.22079/9 qualified lineup mean	0.1784	-0.240	-0.003	+0.001 ok
		Walks	0.465 → 0.539	+0.075	0.0889	0.0802n=142 Average	0.10309/9 qualified lineup mean	0.0930	-0.050	+0.080	-0.007 ok
		Home runs	0.113 → 0.107	-0.006	0.0304	0.0316n=142 Average	0.02879/9 qualified lineup mean	0.0298	+0.007	-0.010	-0.000 ok
RELIEVER Bradgley Rodriguez2.6224 BF	Long Reliever	Hits	0.941 → 0.974	+0.033	0.2169	0.2063n=301 Good	0.22439/9 qualified lineup mean	0.2135	-0.028	+0.019	-0.001 ok
		Strikeouts	1.524 → 1.520	-0.004	0.2212	0.2590n=301 Good	0.22079/9 qualified lineup mean	0.2584	+0.099	-0.002	-0.000 ok
		Walks	0.450 → 0.522	+0.073	0.0889	0.0742n=301 Good	0.10309/9 qualified lineup mean	0.0862	-0.039	+0.037	-0.006 ok
		Home runs	0.052 → 0.049	-0.003	0.0304	0.0262n=301 Good	0.02879/9 qualified lineup mean	0.0247	-0.011	-0.004	+0.001 ok
RELIEVER David Morgan1.9868 BF	Reliever	Hits	0.439 → 0.454	+0.015	0.2169	0.2210n=162 Average	0.22439/9 qualified lineup mean	0.2285	+0.008	+0.015	+0.000 ok
		Strikeouts	0.466 → 0.464	-0.001	0.2212	0.2343n=162 Average	0.22079/9 qualified lineup mean	0.2337	+0.026	-0.001	0 ok
		Walks	0.224 → 0.258	+0.035	0.0889	0.1127n=162 Below average	0.10309/9 qualified lineup mean	0.1300	+0.047	+0.028	+0.006 ok
		Home runs	0.051 → 0.048	-0.003	0.0304	0.0256n=162 Good	0.02879/9 qualified lineup mean	0.0241	-0.009	-0.003	+0.001 ok
RELIEVER Yuki Matsui1.4985 BF	Reliever	Hits	0.273 → 0.282	+0.010	0.2169	0.1819n=260 Very good	0.22439/9 qualified lineup mean	0.1884	-0.052	+0.011	-0.001 ok
		Strikeouts	0.378 → 0.377	-0.001	0.2212	0.2520n=260 Good	0.22079/9 qualified lineup mean	0.2514	+0.046	-0.001	-0.000 ok
		Walks	0.179 → 0.207	+0.027	0.0889	0.1198n=260 Poor	0.10309/9 qualified lineup mean	0.1380	+0.046	+0.021	+0.006 ok
		Home runs	0.044 → 0.041	-0.003	0.0304	0.0291n=260 Average	0.02879/9 qualified lineup mean	0.0275	-0.002	-0.003	+0.000 ok
RELIEVER Wandy Peralta0.689 BF	Reliever	Hits	0.160 → 0.165	+0.005	0.2169	0.2321n=321 Below average	0.22439/9 qualified lineup mean	0.2399	+0.011	+0.005	+0.000 ok
		Strikeouts	0.116 → 0.116	-0.000	0.2212	0.1688n=321 Poor	0.22079/9 qualified lineup mean	0.1684	-0.036	-0.000	+0.000 ok
		Walks	0.053 → 0.062	+0.009	0.0889	0.0773n=321 Good	0.10309/9 qualified lineup mean	0.0898	-0.008	+0.010	-0.001 ok
		Home runs	0.018 → 0.017	-0.001	0.0304	0.0268n=321 Good	0.02879/9 qualified lineup mean	0.0253	-0.003	-0.001	+0.000 ok
RELIEVER Joe Musgrove0.5486 BF	Reliever	Hits	0.119 → 0.123	+0.004	0.2169	0.2169n=0 averagebelow floor	0.22439/9 qualified lineup mean	0.2243	0	+0.004	0 ok
		Strikeouts	0.121 → 0.121	-0.000	0.2212	0.2212n=0 averagebelow floor	0.22079/9 qualified lineup mean	0.2207	0	-0.000	0 ok
		Walks	0.049 → 0.057	+0.008	0.0889	0.0889n=0 averagebelow floor	0.10309/9 qualified lineup mean	0.1030	0	+0.008	0 ok
		Home runs	0.017 → 0.016	-0.001	0.0304	0.0304n=0 averagebelow floor	0.02879/9 qualified lineup mean	0.0287	0	-0.001	0 ok

Published pitching, integer draw beside the continuous model line

Pitcher	Role	BF	IP	expected IP	H int/cont	SO int/cont	BB int/cont	ER int/cont	Rate source
Robert Gasser	Starter	15.8	3.2	3.744	3 / 2.886	4 / 3.300	1 / 1.658	2 / 1.146	all_players_stats_recent_season_blend
Antonio Senzatela	Middle Reliever (High Leverage)	4.4	1.0	1.013	0 / 1.007	1 / 0.794	1 / 0.397	1 / 0.365	all_players_stats_recent_season_blend
Aaron Ashby	Setup Man	5.3	1.1	1.351	1 / 0.861	2 / 1.556	0 / 0.417	1 / 0.293	all_players_stats_recent_season_blend
Trevor Megill	Closer	3.8	4.0	1.013	3 / 0.540	1 / 1.073	1 / 0.177	0 / 0.164	all_players_stats_recent_season_blend
Walker Buehler	Starter	15.5	3.2	3.605	4 / 3.388	4 / 3.233	3 / 1.264	3 / 1.770	all_players_stats_recent_season_blend
Bradgley Rodriguez	Long Reliever	5.3	1.1	1.318	0 / 0.941	1 / 1.524	0 / 0.450	0 / 0.253	all_players_stats_recent_season_blend
Jason Adam	Setup Man	4.2	1.0	0.989	0 / 0.745	1 / 0.475	0 / 0.465	0 / 0.274	all_players_stats_recent_season_blend
Adrian Morejon	Closer	5.4	4.0	1.318	4 / 1.069	3 / 1.434	0 / 0.419	2 / 0.412	all_players_stats_recent_season_blend

Audit: 2 CHECK(S) FAILED

Invariant	Verdict	Evidence
blend arithmetic away_runs	PASS	4.0000 + 0.34 x (3.4240 - 4.0000) = 3.8042, recorded 3.8042
blend arithmetic home_runs	PASS	5.0000 + 0.38 x (6.0217 - 5.0000) = 5.3883, recorded 5.3883
blend arithmetic home_win	PASS	0.5050 + 0.00 x (0.7071 - 0.5050) = 0.5050, recorded 0.5050
published rounding faithful to the blend	PASS	blended 3.8042-5.3883 rounds to 4-5, layer published 4-5
published probability on the score's side	PASS	4-5 with p(home) 0.505000
Log5 additivity136 checks	PASS	136 of 136 pass
away club fields exactly one starter	PASS	1 starters in a staff of 4
away staff totals a regulation workload	FAIL	30 outs recorded, 27 expected
away starter carries a continuous line	PASS	expected_IP 3.7437
away starter batters faced inside the real band	PASS	Robert Gasser projected 15.8 batters faced; real starters run p01-p99 4-30, median 23, mean 21.75
away reliever batters faced inside the real band3 checks	PASS	3 of 3 pass
home club fields exactly one starter	PASS	1 starters in a staff of 4
home staff totals a regulation workload	FAIL	30 outs recorded, 27 expected
home starter carries a continuous line	PASS	expected_IP 3.605
home starter batters faced inside the real band	PASS	Walker Buehler projected 15.5 batters faced; real starters run p01-p99 4-30, median 23, mean 21.75
home reliever batters faced inside the real band3 checks	PASS	3 of 3 pass
away hitters satisfy every coherence rule	PASS	9 hitters checked; H<=AB, HR<=H, 2B+3B+HR<=H, SO<=AB, continuous line present
home hitters satisfy every coherence rule	PASS	9 hitters checked; H<=AB, HR<=H, 2B+3B+HR<=H, SO<=AB, continuous line present

20261007_TB@NYY — published 3-4, 157 checks, 1 failed

Artifact. 20261007_TB@NYY__20261008T004948Z.json, phase post_start_rerun, generated 2026-10-08T00:49:48, first pitch 2026-10-08T00:00:00.
Published 3 - 4 with p(home) 0.544016. Simulation draw chosen on not recorded; ? of ? draws realized the target exactly.

Stage chain: every stage that touched the score

Stage	Score after	Δ away	Δ home	Cum. away	Cum. home	Changed target	Own adjustment and description
1. `historical_ml`	3.959661691282892 - 5.5685220273548195	0	0	0	0	no	Base ML historical model prediction
2. `variance_applied`	4 - 6	+0.0403	+0.4315	+0.0403	+0.4315	yes	Deterministic baseline target (variance factor 1.0)
3. `enhanced_prediction`	3 - 4	-1.0000	-2.0000	-0.9597	-1.5685	yes	Enhanced ML with recent form, H2H, low-score analysis (confidence shrunk by projected_margin_is_one_run, matchup_recency_guardrail)
4. `kpi_layer`	3 - 4	0	0	-0.9597	-1.5685	no	away +0.4627; home +0.1996 Savant KPI blend applied to the model target
5. `embedded_savant_run_environment`	3 - 4	0	0	-0.9597	-1.5685	no	away -0.2946; home +0.4252 Target score adjusted by temporal Savant lineup quality versus opposing starter quality
6. `fundamental_edge_target_review`	3 - 4	0	0	-0.9597	-1.5685	no	Refreshed score target and win probability after Savant/pitcher/team-strength adjustments
7. `capped_target`	3 - 4	0	0	-0.9597	-1.5685	no	Capped at max 12 realistic runs
8. `simulation_result`	3 - 4	-	-	-	-	no	Final inning-by-inning simulation result

KPI team layer: the blend, with its arithmetic

Target	Baseline	KPI only	Weight	Blended	Computation	Verdict
`away_runs`	3.0000	4.3610	0.34	3.4627	3.0000 + 0.34 × (4.3610 − 3.0000) = 3.4627	PASS
`home_runs`	4.0000	4.5253	0.38	4.1996	4.0000 + 0.38 × (4.5253 − 4.0000) = 4.1996	PASS
`home_win`	0.5665	0.5817	0.00	0.5665	0.5665 + 0.00 × (0.5817 − 0.5665) = 0.5665	PASS

Published by the layer: away 3, home 4, p(home) 0.566481, continuous away 3.4627 / home 4.1996

Inputs and gates. candidate 20261007T091423Z_47b9db81, value age 10 days, lineup coverage away 0.9825 / home 0.8947, movement cap 2.5 runs.
values_present ok values_fresh ok one_source_bundle ok lineup_coverage ok target_eligible ok movement_within_limit ok published_movement_within_limit ok

What drove the KPI forecast for `away_runs`

intercept 4.2358 + sum of contributions = 4.3610; prediction 4.3610; contributions outside the top listed 0.9711 runs

Feature	Raw	Standardised	Runs contributed	Which KPIs it is built from
`away_off_worst3_discipline`	-0.0345	1.951	+0.5862	mean of the three weakest plate approaches
`home_platoon_barrel`	0.5586	1.725	-0.5840	platoon term, selected by the opposing starter's throwing hand
`away_sp_suppress`	1.5422	2.540	-0.4676	starter group: pitcher_barrel_rate_allowed, pitcher_hard_hit_rate_allowed, pitcher_contact_xwoba_allowed, pitcher_home_run_rate_allowed
`away_off_power_spread`	1.2760	1.967	+0.4651	offence group (dispersion across the lineup, not its mean): batter_home_run_rate, batter_bat_speed
`away_off_discipline`	0.6124	2.615	+0.4651	offence group: batter_whiff_rate_vs_fastball, batter_whiff_rate_vs_breaking, batter_whiff_rate_vs_offspeed, batter_strikeout_rate ...
`home_off_top3_contact`	1.3995	2.059	-0.4576	mean of the three best contact bats, a tail a club mean cannot express
`away_sp_ground`	0.8588	1.342	-0.4370	starter group: pitcher_ground_ball_rate_induced
`away_off_contact_spread`	0.9747	1.397	-0.3885	offence group (dispersion across the lineup, not its mean): batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`home_off_qualified_share`	0.8947	1.086	-0.3364	share of the lineup carrying a qualifying sample, not a skill measurement
`home_x_ground`	-0.2803	-1.378	+0.3088	batter x pitcher interaction: pitcher_ground_ball_rate_induced, batter_ground_ball_rate_vs_fastball

What drove the KPI forecast for `home_runs`

intercept 4.5238 + sum of contributions = 4.5253; prediction 4.5253; contributions outside the top listed 0.5791 runs

Feature	Raw	Standardised	Runs contributed	Which KPIs it is built from
`away_off_contact_spread`	0.9747	1.397	-0.6220	offence group (dispersion across the lineup, not its mean): batter_barrel_rate_vs_fastball, batter_barrel_rate_vs_breaking, batter_barrel_rate_vs_offspeed, batter_hard_hit_rate_vs_fastball ...
`away_platoon_whiff`	0.9015	2.892	-0.4857	platoon term, selected by the opposing starter's throwing hand
`away_sp_velocity`	-0.5170	-0.979	-0.4501	starter group: pitcher_release_speed
`home_off_ground`	0.4468	1.342	+0.3886	offence group: batter_ground_ball_rate_vs_slider, batter_ground_ball_rate_vs_fastball
`home_off_top3_contact`	1.3995	2.059	-0.3712	mean of the three best contact bats, a tail a club mean cannot express
`away_off_power_spread`	1.2760	1.967	+0.3502	offence group (dispersion across the lineup, not its mean): batter_home_run_rate, batter_bat_speed
`home_sp_whiff`	-0.7672	-1.609	+0.3079	starter group: pitcher_whiff_rate, pitcher_whiff_rate_fastball, pitcher_whiff_rate_breaking
`away_sp_suppress`	1.5422	2.540	+0.2970	starter group: pitcher_barrel_rate_allowed, pitcher_hard_hit_rate_allowed, pitcher_contact_xwoba_allowed, pitcher_home_run_rate_allowed
`home_off_qualified_share`	0.8947	1.086	+0.2830	share of the lineup carrying a qualifying sample, not a skill measurement
`home_platoon_barrel`	0.5586	1.725	-0.2753	platoon term, selected by the opposing starter's throwing hand

Per-player metric movement

TB against Tyler Wells (RHP)

The pipeline applied one club-wide ×1.1542 to every TB hitter line, identical for the leadoff hitter and the ninth bat.
Opposing staff exposure. These hitters face the starter for 20.9702 of 36.528 projected batters (57.4%) and the 3 projected relievers for the remaining 42.6%. Every metric below is the exposure-weighted blend of both matchups; the starter-only figure is shown beside it. Real 2026 game logs put a starter at 21.75 batters faced on average (median 23, p10-p90 17-27) and a whole staff at 37.9: this starter is inside that band and this staff is inside its. The per-player columns are the measured alternative: each subject moved by his own KPI pairing. Exposure for a hitter is his projected plate appearances; for the starter it is batters faced.

Player	Pos	Metric	Published → KPI	Δ	League	Own rate	Opponent rate	Log5	Own skill	Opponent	Remainder
Yandy Díaz	DH	Hitswhole staff	1.054 → 0.996	-0.058	0.2169	0.2464n=678 Very good	0.1991n=418 Good	0.2269	+0.132	-0.080	-0.008 ok
		split	starter x0.921 at 0.574 share (log5 0.2269 vs his 0.1991) + bullpen x0.9766 at 0.426 share (log5 0.2406 vs pen 0.2116) = x0.9447 \| starter alone would have been -0.083
		Strikeoutswhole staff	0.735 → 0.773	+0.038	0.2212	0.1455n=678 Very good	0.2303n=418 Average	0.1521	-0.339	+0.041	-0.011 ok
		split	starter x1.0453 at 0.574 share (log5 0.1521 vs his 0.2303) + bullpen x1.0606 at 0.426 share (log5 0.1544 vs pen 0.2334) = x1.0518 \| starter alone would have been +0.033
		Walkswhole staff	0.426 → 0.394	-0.033	0.0889	0.0835n=678 Average	0.0799n=418 Average	0.0751	-0.024	-0.040	+0.002 ok
		split	starter x0.8985 at 0.574 share (log5 0.0751 vs his 0.0799) + bullpen x0.9572 at 0.426 share (log5 0.0800 vs pen 0.0851) = x0.9235 \| starter alone would have been -0.043
		Home runswhole staff	0.163 → 0.126	-0.037	0.0304	0.0319n=678 Average	0.0203n=418 Very good	0.0213	+0.007	-0.045	-0.002 ok
		split	starter x0.6692 at 0.574 share (log5 0.0213 vs his 0.0203) + bullpen x0.9058 at 0.426 share (log5 0.0289 vs pen 0.0275) = x0.77 \| starter alone would have been -0.054
Neutralising Yandy Díaz to the qualified-population mean moves the club's KPI run forecast by 0.7051 runs, and the weighted stage by 0.2397. Interactions make these non-additive across players, so they do not sum to the published movement.
Jonathan Aranda	1B	Hitswhole staff	0.922 → 0.870	-0.052	0.2169	0.2302n=661 Good	0.1991n=418 Good	0.2116	+0.057	-0.076	-0.003 ok
		split	starter x0.9194 at 0.574 share (log5 0.2116 vs his 0.1991) + bullpen x0.9761 at 0.426 share (log5 0.2247 vs pen 0.2116) = x0.9435 \| starter alone would have been -0.074
		Strikeoutswhole staff	1.272 → 1.330	+0.058	0.2212	0.2420n=661 Average	0.2303n=418 Average	0.2517	+0.089	+0.039	+0.003 ok
		split	starter x1.04 at 0.574 share (log5 0.2517 vs his 0.2303) + bullpen x1.0534 at 0.426 share (log5 0.2550 vs pen 0.2334) = x1.0457 \| starter alone would have been +0.051
		Walkswhole staff	0.408 → 0.377	-0.031	0.0889	0.1007n=661 Good	0.0799n=418 Average	0.0907	+0.051	-0.038	-0.005 ok
		split	starter x0.9002 at 0.574 share (log5 0.0907 vs his 0.0799) + bullpen x0.958 at 0.426 share (log5 0.0965 vs pen 0.0851) = x0.9248 \| starter alone would have been -0.041
		Home runswhole staff	0.168 → 0.130	-0.039	0.0304	0.0325n=661 Average	0.0203n=418 Very good	0.0217	+0.009	-0.043	-0.003 ok
		split	starter x0.6693 at 0.574 share (log5 0.0217 vs his 0.0203) + bullpen x0.9059 at 0.426 share (log5 0.0294 vs pen 0.0275) = x0.7701 \| starter alone would have been -0.056
Neutralising Jonathan Aranda to the qualified-population mean moves the club's KPI run forecast by -0.0272 runs, and the weighted stage by -0.0093. Interactions make these non-additive across players, so they do not sum to the published movement.
Junior Caminero	3B	Hitswhole staff	0.951 → 0.897	-0.053	0.2169	0.2330n=695 Good	0.1991n=418 Good	0.2143	+0.067	-0.073	-0.004 ok
		split	starter x0.9197 at 0.574 share (log5 0.2143 vs his 0.1991) + bullpen x0.9762 at 0.426 share (log5 0.2274 vs pen 0.2116) = x0.9438 \| starter alone would have been -0.076
		Strikeoutswhole staff	0.803 → 0.843	+0.040	0.2212	0.1853n=695 Good	0.2303n=418 Average	0.1932	-0.148	+0.038	-0.005 ok
		split	starter x1.0431 at 0.574 share (log5 0.1932 vs his 0.2303) + bullpen x1.0576 at 0.426 share (log5 0.1959 vs pen 0.2334) = x1.0493 \| starter alone would have been +0.035
		Walkswhole staff	0.435 → 0.403	-0.032	0.0889	0.1100n=695 Good	0.0799n=418 Average	0.0992	+0.087	-0.037	-0.008 ok
		split	starter x0.9011 at 0.574 share (log5 0.0992 vs his 0.0799) + bullpen x0.9584 at 0.426 share (log5 0.1055 vs pen 0.0851) = x0.9255 \| starter alone would have been -0.043
		Home runswhole staff	0.231 → 0.179	-0.052	0.0304	0.0522n=695 Very good	0.0203n=418 Very good	0.0351	+0.090	-0.042	-0.029 ok
		split	starter x0.6739 at 0.574 share (log5 0.0351 vs his 0.0203) + bullpen x0.9076 at 0.426 share (log5 0.0473 vs pen 0.0275) = x0.7734 \| starter alone would have been -0.075
Neutralising Junior Caminero to the qualified-population mean moves the club's KPI run forecast by 0.2500 runs, and the weighted stage by 0.0850. Interactions make these non-additive across players, so they do not sum to the published movement.
Victor Mesa Jr.	LF	Hitswhole staff	0.822 → 0.775	-0.047	0.2169	0.2223n=259 Average	0.1991n=418 Good	0.2043	+0.022	-0.070	-0.001 ok
		split	starter x0.9186 at 0.574 share (log5 0.2043 vs his 0.1991) + bullpen x0.9759 at 0.426 share (log5 0.2170 vs pen 0.2116) = x0.943 \| starter alone would have been -0.067
		Strikeoutswhole staff	0.972 → 1.019	+0.047	0.2212	0.2073n=259 Average	0.2303n=418 Average	0.2159	-0.055	+0.036	-0.002 ok
		split	starter x1.0419 at 0.574 share (log5 0.2159 vs his 0.2303) + bullpen x1.056 at 0.426 share (log5 0.2189 vs pen 0.2334) = x1.0479 \| starter alone would have been +0.041
		Walkswhole staff	0.267 → 0.246	-0.020	0.0889	0.0888n=259 Average	0.0799n=418 Average	0.0799	-0.000	-0.035	0 ok
		split	starter x0.899 at 0.574 share (log5 0.0799 vs his 0.0799) + bullpen x0.9574 at 0.426 share (log5 0.0851 vs pen 0.0851) = x0.9239 \| starter alone would have been -0.027
		Home runswhole staff	0.189 → 0.146	-0.043	0.0304	0.0421n=259 Good	0.0203n=418 Very good	0.0283	+0.046	-0.040	-0.015 ok
		split	starter x0.6715 at 0.574 share (log5 0.0283 vs his 0.0203) + bullpen x0.9067 at 0.426 share (log5 0.0382 vs pen 0.0275) = x0.7717 \| starter alone would have been -0.062
Neutralising Victor Mesa Jr. to the qualified-population mean moves the club's KPI run forecast by -0.0919 runs, and the weighted stage by -0.0313. Interactions make these non-additive across players, so they do not sum to the published movement.
Ryan Vilade	RF	Hitswhole staff	0.763 → 0.720	-0.044	0.2169	0.2163n=274 Average	0.1991n=418 Good	0.1986	-0.002	-0.067	+0.000 ok
		split	starter x0.9181 at 0.574 share (log5 0.1986 vs his 0.1991) + bullpen x0.9757 at 0.426 share (log5 0.2110 vs pen 0.2116) = x0.9426 \| starter alone would have been -0.063
		Strikeoutswhole staff	0.971 → 1.015	+0.045	0.2212	0.2348n=274 Average	0.2303n=418 Average	0.2443	+0.051	+0.034	+0.001 ok
		split	starter x1.0404 at 0.574 share (log5 0.2443 vs his 0.2303) + bullpen x1.0539 at 0.426 share (log5 0.2475 vs pen 0.2334) = x1.0461 \| starter alone would have been +0.039
		Walkswhole staff	0.298 → 0.276	-0.023	0.0889	0.0931n=274 Average	0.0799n=418 Average	0.0837	+0.016	-0.034	-0.001 ok
		split	starter x0.8994 at 0.574 share (log5 0.0837 vs his 0.0799) + bullpen x0.9576 at 0.426 share (log5 0.0891 vs pen 0.0851) = x0.9242 \| starter alone would have been -0.030
		Home runswhole staff	0.167 → 0.128	-0.038	0.0304	0.0354n=274 Good	0.0203n=418 Very good	0.0237	+0.019	-0.038	-0.006 ok
		split	starter x0.67 at 0.574 share (log5 0.0237 vs his 0.0203) + bullpen x0.9061 at 0.426 share (log5 0.0320 vs pen 0.0275) = x0.7706 \| starter alone would have been -0.055
Neutralising Ryan Vilade to the qualified-population mean moves the club's KPI run forecast by 0.1709 runs, and the weighted stage by 0.0581. Interactions make these non-additive across players, so they do not sum to the published movement.
Liam Hicks	C	Hitswhole staff	0.805 → 0.760	-0.045	0.2169	0.2335n=551 Good	0.1991n=418 Good	0.2148	+0.064	-0.069	-0.004 ok
		split	starter x0.9197 at 0.574 share (log5 0.2148 vs his 0.1991) + bullpen x0.9762 at 0.426 share (log5 0.2280 vs pen 0.2116) = x0.9438 \| starter alone would have been -0.065
		Strikeoutswhole staff	0.501 → 0.528	+0.027	0.2212	0.1080n=551 Very good	0.2303n=418 Average	0.1131	-0.437	+0.035	-0.015 ok
		split	starter x1.0474 at 0.574 share (log5 0.1131 vs his 0.2303) + bullpen x1.0634 at 0.426 share (log5 0.1149 vs pen 0.2334) = x1.0542 \| starter alone would have been +0.024
		Walkswhole staff	0.365 → 0.337	-0.028	0.0889	0.0994n=551 Average	0.0799n=418 Average	0.0894	+0.040	-0.035	-0.004 ok
		split	starter x0.9 at 0.574 share (log5 0.0894 vs his 0.0799) + bullpen x0.9579 at 0.426 share (log5 0.0952 vs pen 0.0851) = x0.9247 \| starter alone would have been -0.036
		Home runswhole staff	0.121 → 0.093	-0.028	0.0304	0.0319n=551 Average	0.0203n=418 Very good	0.0214	+0.006	-0.039	-0.002 ok
		split	starter x0.6692 at 0.574 share (log5 0.0214 vs his 0.0203) + bullpen x0.9058 at 0.426 share (log5 0.0289 vs pen 0.0275) = x0.77 \| starter alone would have been -0.040
Neutralising Liam Hicks to the qualified-population mean moves the club's KPI run forecast by 0.2182 runs, and the weighted stage by 0.0742. Interactions make these non-additive across players, so they do not sum to the published movement.
Jonny DeLuca	CF	Hitswhole staff	0.899 → 0.849	-0.051	0.2169	0.2346n=363 Good	0.1991n=418 Good	0.2158	+0.068	-0.068	-0.004 ok
		split	starter x0.9198 at 0.574 share (log5 0.2158 vs his 0.1991) + bullpen x0.9763 at 0.426 share (log5 0.2291 vs pen 0.2116) = x0.9439 \| starter alone would have been -0.072
		Strikeoutswhole staff	0.993 → 1.040	+0.047	0.2212	0.2132n=363 Average	0.2303n=418 Average	0.2221	-0.030	+0.035	-0.001 ok
		split	starter x1.0416 at 0.574 share (log5 0.2221 vs his 0.2303) + bullpen x1.0555 at 0.426 share (log5 0.2250 vs pen 0.2334) = x1.0475 \| starter alone would have been +0.041
		Walkswhole staff	0.253 → 0.233	-0.020	0.0889	0.0572n=363 Poor	0.0799n=418 Average	0.0513	-0.121	-0.034	+0.011 ok
		split	starter x0.8959 at 0.574 share (log5 0.0513 vs his 0.0799) + bullpen x0.956 at 0.426 share (log5 0.0547 vs pen 0.0851) = x0.9215 \| starter alone would have been -0.026
		Home runswhole staff	0.142 → 0.109	-0.033	0.0304	0.0303n=363 Average	0.0203n=418 Very good	0.0203	-0.000	-0.038	+0.000 ok
		split	starter x0.6688 at 0.574 share (log5 0.0203 vs his 0.0203) + bullpen x0.9057 at 0.426 share (log5 0.0275 vs pen 0.0275) = x0.7697 \| starter alone would have been -0.047
Neutralising Jonny DeLuca to the qualified-population mean moves the club's KPI run forecast by -0.0170 runs, and the weighted stage by -0.0058. Interactions make these non-additive across players, so they do not sum to the published movement.
Richie Palacios	2B	Hitswhole staff	0.755 → 0.711	-0.043	0.2169	0.2110n=396 Average	0.1991n=418 Good	0.1936	-0.021	-0.063	+0.001 ok
		split	starter x0.9176 at 0.574 share (log5 0.1936 vs his 0.1991) + bullpen x0.9756 at 0.426 share (log5 0.2059 vs pen 0.2116) = x0.9423 \| starter alone would have been -0.062
		Strikeoutswhole staff	0.738 → 0.775	+0.037	0.2212	0.1749n=396 Good	0.2303n=418 Average	0.1825	-0.165	+0.032	-0.005 ok
		split	starter x1.0437 at 0.574 share (log5 0.1825 vs his 0.2303) + bullpen x1.0584 at 0.426 share (log5 0.1851 vs pen 0.2334) = x1.05 \| starter alone would have been +0.032
		Walkswhole staff	0.371 → 0.343	-0.028	0.0889	0.1099n=396 Good	0.0799n=418 Average	0.0990	+0.074	-0.032	-0.007 ok
		split	starter x0.9011 at 0.574 share (log5 0.0990 vs his 0.0799) + bullpen x0.9584 at 0.426 share (log5 0.1053 vs pen 0.0851) = x0.9255 \| starter alone would have been -0.037
		Home runswhole staff	0.114 → 0.088	-0.026	0.0304	0.0273n=396 Average	0.0203n=418 Very good	0.0182	-0.011	-0.036	+0.004 ok
		split	starter x0.6681 at 0.574 share (log5 0.0182 vs his 0.0203) + bullpen x0.9054 at 0.426 share (log5 0.0247 vs pen 0.0275) = x0.7692 \| starter alone would have been -0.038
Neutralising Richie Palacios to the qualified-population mean moves the club's KPI run forecast by -0.0171 runs, and the weighted stage by -0.0058. Interactions make these non-additive across players, so they do not sum to the published movement.
Taylor Walls	SS	Hitswhole staff	0.658 → 0.620	-0.038	0.2169	0.2028n=430 Below average	0.1991n=418 Good	0.1859	-0.049	-0.062	+0.003 ok
		split	starter x0.9168 at 0.574 share (log5 0.1859 vs his 0.1991) + bullpen x0.9753 at 0.426 share (log5 0.1978 vs pen 0.2116) = x0.9417 \| starter alone would have been -0.055
		Strikeoutswhole staff	0.928 → 0.970	+0.042	0.2212	0.2418n=430 Average	0.2303n=418 Average	0.2515	+0.072	+0.032	+0.002 ok
		split	starter x1.04 at 0.574 share (log5 0.2515 vs his 0.2303) + bullpen x1.0534 at 0.426 share (log5 0.2548 vs pen 0.2334) = x1.0457 \| starter alone would have been +0.037
		Walkswhole staff	0.356 → 0.329	-0.027	0.0889	0.1012n=430 Good	0.0799n=418 Average	0.0911	+0.043	-0.031	-0.004 ok
		split	starter x0.9002 at 0.574 share (log5 0.0911 vs his 0.0799) + bullpen x0.958 at 0.426 share (log5 0.0970 vs pen 0.0851) = x0.9248 \| starter alone would have been -0.035
		Home runswhole staff	0.059 → 0.046	-0.014	0.0304	0.0144n=430 Poor	0.0203n=418 Very good	0.0096	-0.056	-0.035	+0.018 ok
		split	starter x0.6652 at 0.574 share (log5 0.0096 vs his 0.0203) + bullpen x0.9043 at 0.426 share (log5 0.0131 vs pen 0.0275) = x0.767 \| starter alone would have been -0.020
Neutralising Taylor Walls to the qualified-population mean moves the club's KPI run forecast by -0.0179 runs, and the weighted stage by -0.0061. Interactions make these non-additive across players, so they do not sum to the published movement.
STARTER Nick Martinez18.0453 BF	Starter	Hits	4.238 → 4.243	+0.005	0.2169	0.2316n=698 Below average	0.21718/9 qualified lineup mean	0.2319	+0.266	+0.005	+0.000 ok
		Strikeouts	3.171 → 3.257	+0.087	0.2212	0.1539n=698 Poor	0.22688/9 qualified lineup mean	0.1581	-1.215	+0.100	-0.024 ok
		Walks	0.896 → 0.971	+0.075	0.0889	0.0456n=698 Very good	0.09608/9 qualified lineup mean	0.0494	-0.782	+0.127	-0.059 ok
		Home runs	0.675 → 0.783	+0.108	0.0304	0.0302n=698 Average	0.03528/9 qualified lineup mean	0.0350	-0.003	+0.088	-0.001 ok
RELIEVER Cam Booser4.7411 BF	Setup Man	Hits	0.368 → 0.368	+0.001	0.2169	0.1990n=152 Good	0.21718/9 qualified lineup mean	0.1993	-0.085	+0.001	-0.000 ok
		Strikeouts	0.713 → 0.730	+0.017	0.2212	0.2702n=152 Good	0.22688/9 qualified lineup mean	0.2766	+0.232	+0.026	+0.004 ok
		Walks	0.127 → 0.137	+0.010	0.0889	0.0646n=152 Very good	0.09608/9 qualified lineup mean	0.0699	-0.115	+0.033	-0.008 ok
		Home runs	0.068 → 0.079	+0.011	0.0304	0.0310n=152 Average	0.03528/9 qualified lineup mean	0.0360	+0.003	+0.023	+0.001 ok
RELIEVER Bryan Baker4.6834 BF	Closer	Hits	0.540 → 0.541	+0.001	0.2169	0.1836n=244 Very good	0.21718/9 qualified lineup mean	0.1838	-0.156	+0.001	-0.000 ok
		Strikeouts	0.931 → 0.954	+0.022	0.2212	0.2490n=244 Good	0.22688/9 qualified lineup mean	0.2551	+0.130	+0.026	+0.002 ok
		Walks	0.287 → 0.310	+0.023	0.0889	0.0897n=244 Average	0.09608/9 qualified lineup mean	0.0968	+0.004	+0.033	+0.000 ok
		Home runs	0.049 → 0.057	+0.008	0.0304	0.0254n=244 Good	0.03528/9 qualified lineup mean	0.0295	-0.023	+0.023	-0.004 ok
RELIEVER Ian Seymour2.5597 BF	Reliever	Hits	0.538 → 0.539	+0.001	0.2169	0.2104n=551 Average	0.21718/9 qualified lineup mean	0.2106	-0.017	+0.001	0 ok
		Strikeouts	0.713 → 0.729	+0.017	0.2212	0.2784n=551 Very good	0.22688/9 qualified lineup mean	0.2849	+0.146	+0.014	+0.002 ok
		Walks	0.181 → 0.195	+0.015	0.0889	0.0706n=551 Good	0.09608/9 qualified lineup mean	0.0763	-0.047	+0.018	-0.003 ok
		Home runs	0.099 → 0.115	+0.016	0.0304	0.0387n=551 Poor	0.03528/9 qualified lineup mean	0.0448	+0.021	+0.012	+0.003 ok
RELIEVER Steven Matz2.0966 BF	Reliever	Hits	0.454 → 0.455	+0.001	0.2169	0.2166n=342 Average	0.21718/9 qualified lineup mean	0.2169	-0.001	+0.001	0 ok
		Strikeouts	0.391 → 0.401	+0.010	0.2212	0.1865n=342 Below average	0.22688/9 qualified lineup mean	0.1914	-0.073	+0.012	-0.001 ok
		Walks	0.164 → 0.177	+0.013	0.0889	0.0781n=342 Good	0.09608/9 qualified lineup mean	0.0844	-0.023	+0.015	-0.002 ok
		Home runs	0.076 → 0.088	+0.012	0.0304	0.0362n=342 Poor	0.03528/9 qualified lineup mean	0.0420	+0.012	+0.010	+0.002 ok
RELIEVER Cole Sulser1.7947 BF	Middle Reliever	Hits	0.856 → 0.857	+0.001	0.2169	0.2203n=236 Average	0.21718/9 qualified lineup mean	0.2206	+0.006	+0.001	0 ok
		Strikeouts	1.079 → 1.106	+0.027	0.2212	0.2332n=236 Average	0.22688/9 qualified lineup mean	0.2390	+0.021	+0.010	+0.000 ok
		Walks	0.311 → 0.336	+0.025	0.0889	0.0788n=236 Good	0.09608/9 qualified lineup mean	0.0851	-0.018	+0.013	-0.001 ok
		Home runs	0.104 → 0.121	+0.017	0.0304	0.0301n=236 Average	0.03528/9 qualified lineup mean	0.0350	-0.001	+0.009	-0.000 ok
RELIEVER Kevin Kelly0.7345 BF	Reliever	Hits	0.152 → 0.152	+0.000	0.2169	0.2073n=279 Average	0.21718/9 qualified lineup mean	0.2075	-0.007	+0.000	0 ok
		Strikeouts	0.165 → 0.169	+0.004	0.2212	0.2248n=279 Average	0.22688/9 qualified lineup mean	0.2304	+0.003	+0.004	0 ok
		Walks	0.042 → 0.045	+0.004	0.0889	0.0573n=279 Very good	0.09608/9 qualified lineup mean	0.0620	-0.023	+0.005	-0.002 ok
		Home runs	0.020 → 0.023	+0.003	0.0304	0.0270n=279 Good	0.03528/9 qualified lineup mean	0.0313	-0.003	+0.004	-0.000 ok
RELIEVER Tyler Wells0.3545 BF	Reliever	Hits	0.069 → 0.069	+0.000	0.2169	0.1936n=273 Very good	0.21718/9 qualified lineup mean	0.1938	-0.008	+0.000	0 ok
		Strikeouts	0.090 → 0.092	+0.002	0.2212	0.2541n=273 Good	0.22688/9 qualified lineup mean	0.2602	+0.012	+0.002	+0.000 ok
		Walks	0.026 → 0.029	+0.002	0.0889	0.0743n=273 Good	0.09608/9 qualified lineup mean	0.0804	-0.005	+0.003	-0.000 ok
		Home runs	0.011 → 0.012	+0.002	0.0304	0.0300n=273 Average	0.03528/9 qualified lineup mean	0.0348	-0.000	+0.002	0 ok

NYY against Ryan Yarbrough (LHP)

The pipeline applied one club-wide ×1.0499 to every NYY hitter line, identical for the leadoff hitter and the ninth bat.
Opposing staff exposure. These hitters face the starter for 18.0453 of 35.008 projected batters (51.5%) and the 3 projected relievers for the remaining 48.4%. Every metric below is the exposure-weighted blend of both matchups; the starter-only figure is shown beside it. Real 2026 game logs put a starter at 21.75 batters faced on average (median 23, p10-p90 17-27) and a whole staff at 37.9: this starter is inside that band and this staff is inside its. The per-player columns are the measured alternative: each subject moved by his own KPI pairing. Exposure for a hitter is his projected plate appearances; for the starter it is batters faced.

Player	Pos	Metric	Published → KPI	Δ	League	Own rate	Opponent rate	Log5	Own skill	Opponent	Remainder
Trent Grisham	CF	Hitswhole staff	0.812 → 0.812	0	0.2169	0.1996n=474 Poor	0.2316n=698 Below average	0.2135	-0.071	+0.060	-0.004 ok
		split	starter x1.0696 at 0.515 share (log5 0.2135 vs his 0.2316) + bullpen x0.926 at 0.484 share (log5 0.1848 vs pen 0.2011) = x1.0 \| starter alone would have been +0.057
		Strikeoutswhole staff	0.678 → 0.612	-0.066	0.2212	0.1871n=474 Good	0.1539n=698 Poor	0.1284	-0.140	-0.275	+0.035 ok
		split	starter x0.6864 at 0.515 share (log5 0.1284 vs his 0.1539) + bullpen x1.1321 at 0.484 share (log5 0.2118 vs pen 0.2490) = x0.9023 \| starter alone would have been -0.213
		Walkswhole staff	0.405 → 0.276	-0.129	0.0889	0.1190n=474 Very good	0.0456n=698 Very good	0.0620	+0.123	-0.177	-0.056 ok
		split	starter x0.5212 at 0.515 share (log5 0.0620 vs his 0.0456) + bullpen x0.8533 at 0.484 share (log5 0.1016 vs pen 0.0755) = x0.6821 \| starter alone would have been -0.194
		Home runswhole staff	0.165 → 0.167	+0.001	0.0304	0.0339n=474 Average	0.0302n=698 Average	0.0337	+0.014	-0.001	-0.000 ok
		split	starter x0.994 at 0.515 share (log5 0.0337 vs his 0.0302) + bullpen x1.0197 at 0.484 share (log5 0.0346 vs pen 0.0310) = x1.0065 \| starter alone would have been -0.001
Neutralising Trent Grisham to the qualified-population mean moves the club's KPI run forecast by 0.4210 runs, and the weighted stage by 0.1600. Interactions make these non-additive across players, so they do not sum to the published movement.
Ben Rice	DH	Hitswhole staff	0.981 → 0.981	-0.000	0.2169	0.2240n=651 Average	0.2316n=698 Below average	0.2391	+0.031	+0.064	+0.002 ok
		split	starter x1.0673 at 0.515 share (log5 0.2391 vs his 0.2316) + bullpen x0.9281 at 0.484 share (log5 0.2079 vs pen 0.2011) = x0.9999 \| starter alone would have been +0.066
		Strikeoutswhole staff	0.871 → 0.788	-0.083	0.2212	0.2398n=651 Average	0.1539n=698 Poor	0.1680	+0.080	-0.291	-0.019 ok
		split	starter x0.7007 at 0.515 share (log5 0.1680 vs his 0.1539) + bullpen x1.1225 at 0.484 share (log5 0.2691 vs pen 0.2490) = x0.9051 \| starter alone would have been -0.261
		Walkswhole staff	0.562 → 0.384	-0.178	0.0889	0.1228n=651 Very good	0.0456n=698 Very good	0.0641	+0.147	-0.187	-0.066 ok
		split	starter x0.5223 at 0.515 share (log5 0.0641 vs his 0.0456) + bullpen x0.8538 at 0.484 share (log5 0.1049 vs pen 0.0755) = x0.6829 \| starter alone would have been -0.268
		Home runswhole staff	0.264 → 0.266	+0.002	0.0304	0.0536n=651 Very good	0.0302n=698 Average	0.0533	+0.100	-0.001	-0.001 ok
		split	starter x0.9941 at 0.515 share (log5 0.0533 vs his 0.0302) + bullpen x1.0193 at 0.484 share (log5 0.0546 vs pen 0.0310) = x1.0063 \| starter alone would have been -0.002
Neutralising Ben Rice to the qualified-population mean moves the club's KPI run forecast by 0.1739 runs, and the weighted stage by 0.0661. Interactions make these non-additive across players, so they do not sum to the published movement.
Luis García Jr.	1B	Hitswhole staff	1.056 → 1.056	-0.000	0.2169	0.2435n=539 Very good	0.2316n=698 Below average	0.2595	+0.109	+0.060	+0.005 ok
		split	starter x1.0655 at 0.515 share (log5 0.2595 vs his 0.2316) + bullpen x0.9298 at 0.484 share (log5 0.2264 vs pen 0.2011) = x0.9998 \| starter alone would have been +0.069
		Strikeoutswhole staff	0.747 → 0.674	-0.073	0.2212	0.1817n=539 Good	0.1539n=698 Poor	0.1245	-0.161	-0.274	+0.041 ok
		split	starter x0.685 at 0.515 share (log5 0.1245 vs his 0.1539) + bullpen x1.1331 at 0.484 share (log5 0.2059 vs pen 0.2490) = x0.9021 \| starter alone would have been -0.235
		Walkswhole staff	0.205 → 0.137	-0.068	0.0889	0.0541n=539 Poor	0.0456n=698 Very good	0.0272	-0.142	-0.176	+0.067 ok
		split	starter x0.5034 at 0.515 share (log5 0.0272 vs his 0.0456) + bullpen x0.8441 at 0.484 share (log5 0.0457 vs pen 0.0755) = x0.6685 \| starter alone would have been -0.102
		Home runswhole staff	0.234 → 0.235	+0.002	0.0304	0.0473n=539 Very good	0.0302n=698 Average	0.0471	+0.069	-0.001	-0.000 ok
		split	starter x0.9941 at 0.515 share (log5 0.0471 vs his 0.0302) + bullpen x1.0194 at 0.484 share (log5 0.0483 vs pen 0.0310) = x1.0064 \| starter alone would have been -0.001
Neutralising Luis García Jr. to the qualified-population mean moves the club's KPI run forecast by 0.3035 runs, and the weighted stage by 0.1153. Interactions make these non-additive across players, so they do not sum to the published movement.
Cody Bellinger	LF	Hitswhole staff	0.976 → 0.976	-0.000	0.2169	0.2269n=549 Good	0.2316n=698 Below average	0.2421	+0.042	+0.062	+0.002 ok
		split	starter x1.067 at 0.515 share (log5 0.2421 vs his 0.2316) + bullpen x0.9284 at 0.484 share (log5 0.2107 vs pen 0.2011) = x0.9998 \| starter alone would have been +0.065
		Strikeoutswhole staff	0.608 → 0.548	-0.060	0.2212	0.1619n=549 Good	0.1539n=698 Poor	0.1101	-0.248	-0.282	+0.065 ok
		split	starter x0.6798 at 0.515 share (log5 0.1101 vs his 0.1539) + bullpen x1.1367 at 0.484 share (log5 0.1840 vs pen 0.2490) = x0.9012 \| starter alone would have been -0.195
		Walkswhole staff	0.295 → 0.200	-0.095	0.0889	0.1027n=549 Good	0.0456n=698 Very good	0.0530	+0.058	-0.181	-0.026 ok
		split	starter x0.5166 at 0.515 share (log5 0.0530 vs his 0.0456) + bullpen x0.8509 at 0.484 share (log5 0.0874 vs pen 0.0755) = x0.6786 \| starter alone would have been -0.143
		Home runswhole staff	0.175 → 0.176	+0.001	0.0304	0.0295n=549 Average	0.0302n=698 Average	0.0294	-0.004	-0.001	0 ok
		split	starter x0.9939 at 0.515 share (log5 0.0294 vs his 0.0302) + bullpen x1.0198 at 0.484 share (log5 0.0301 vs pen 0.0310) = x1.0064 \| starter alone would have been -0.001
Neutralising Cody Bellinger to the qualified-population mean moves the club's KPI run forecast by 0.3804 runs, and the weighted stage by 0.1446. Interactions make these non-additive across players, so they do not sum to the published movement.
George Lombard Jr.	SS	Hitswhole staff	0.890 → 0.890	-0.000	0.2169	0.2169n=0 averagebelow floor	0.2316n=698 Below average	0.2316	0	+0.060	0 ok
		split	starter x1.0679 at 0.515 share (log5 0.2316 vs his 0.2316) + bullpen x0.9275 at 0.484 share (log5 0.2011 vs pen 0.2011) = x0.9999 \| starter alone would have been +0.060
		Strikeoutswhole staff	0.987 → 0.892	-0.095	0.2212	0.2212n=0 averagebelow floor	0.1539n=698 Poor	0.1539	0	-0.272	0 ok
		split	starter x0.6956 at 0.515 share (log5 0.1539 vs his 0.1539) + bullpen x1.1258 at 0.484 share (log5 0.2490 vs pen 0.2490) = x0.904 \| starter alone would have been -0.300
		Walkswhole staff	0.275 → 0.186	-0.089	0.0889	0.0889n=0 averagebelow floor	0.0456n=698 Very good	0.0456	0	-0.175	0 ok
		split	starter x0.5128 at 0.515 share (log5 0.0456 vs his 0.0456) + bullpen x0.849 at 0.484 share (log5 0.0755 vs pen 0.0755) = x0.6757 \| starter alone would have been -0.134
		Home runswhole staff	0.158 → 0.159	+0.001	0.0304	0.0304n=0 averagebelow floor	0.0302n=698 Average	0.0302	0	-0.001	0 ok
		split	starter x0.9939 at 0.515 share (log5 0.0302 vs his 0.0302) + bullpen x1.0198 at 0.484 share (log5 0.0310 vs pen 0.0310) = x1.0064 \| starter alone would have been -0.001
Neutralising George Lombard Jr. to the qualified-population mean moves the club's KPI run forecast by 0.0316 runs, and the weighted stage by 0.0120. Interactions make these non-additive across players, so they do not sum to the published movement.
Austin Wells	C	Hitswhole staff	0.670 → 0.670	0	0.2169	0.1904n=387 Poor	0.2316n=698 Below average	0.2038	-0.090	+0.050	-0.004 ok
		split	starter x1.0704 at 0.515 share (log5 0.2038 vs his 0.2316) + bullpen x0.9252 at 0.484 share (log5 0.1761 vs pen 0.2011) = x1.0001 \| starter alone would have been +0.047
		Strikeoutswhole staff	0.799 → 0.724	-0.075	0.2212	0.2588n=387 Below average	0.1539n=698 Poor	0.1827	+0.127	-0.228	-0.030 ok
		split	starter x0.706 at 0.515 share (log5 0.1827 vs his 0.1539) + bullpen x1.119 at 0.484 share (log5 0.2896 vs pen 0.2490) = x0.9061 \| starter alone would have been -0.235
		Walkswhole staff	0.310 → 0.211	-0.099	0.0889	0.1158n=387 Good	0.0456n=698 Very good	0.0602	+0.091	-0.146	-0.041 ok
		split	starter x0.5203 at 0.515 share (log5 0.0602 vs his 0.0456) + bullpen x0.8528 at 0.484 share (log5 0.0987 vs pen 0.0755) = x0.6814 \| starter alone would have been -0.149
		Home runswhole staff	0.158 → 0.159	+0.001	0.0304	0.0338n=387 Average	0.0302n=698 Average	0.0336	+0.012	-0.001	-0.000 ok
		split	starter x0.994 at 0.515 share (log5 0.0336 vs his 0.0302) + bullpen x1.0197 at 0.484 share (log5 0.0345 vs pen 0.0310) = x1.0065 \| starter alone would have been -0.001
Neutralising Austin Wells to the qualified-population mean moves the club's KPI run forecast by 0.2797 runs, and the weighted stage by 0.1063. Interactions make these non-additive across players, so they do not sum to the published movement.
Anthony Volpe	2B	Hitswhole staff	0.807 → 0.807	-0.000	0.2169	0.2178n=237 Average	0.2316n=698 Below average	0.2325	+0.003	+0.056	+0.000 ok
		split	starter x1.0679 at 0.515 share (log5 0.2325 vs his 0.2316) + bullpen x0.9276 at 0.484 share (log5 0.2020 vs pen 0.2011) = x0.9999 \| starter alone would have been +0.055
		Strikeoutswhole staff	0.852 → 0.770	-0.082	0.2212	0.2197n=237 Average	0.1539n=698 Poor	0.1527	-0.006	-0.255	+0.001 ok
		split	starter x0.6952 at 0.515 share (log5 0.1527 vs his 0.1539) + bullpen x1.1261 at 0.484 share (log5 0.2474 vs pen 0.2490) = x0.904 \| starter alone would have been -0.260
		Walkswhole staff	0.279 → 0.189	-0.090	0.0889	0.0999n=237 Average	0.0456n=698 Very good	0.0516	+0.042	-0.164	-0.019 ok
		split	starter x0.5158 at 0.515 share (log5 0.0516 vs his 0.0456) + bullpen x0.8506 at 0.484 share (log5 0.0850 vs pen 0.0755) = x0.678 \| starter alone would have been -0.135
		Home runswhole staff	0.087 → 0.088	+0.001	0.0304	0.0200n=237 Poor	0.0302n=698 Average	0.0199	-0.039	-0.001	+0.000 ok
		split	starter x0.9939 at 0.515 share (log5 0.0199 vs his 0.0302) + bullpen x1.02 at 0.484 share (log5 0.0204 vs pen 0.0310) = x1.0065 \| starter alone would have been -0.001
Neutralising Anthony Volpe to the qualified-population mean moves the club's KPI run forecast by 0.0978 runs, and the weighted stage by 0.0372. Interactions make these non-additive across players, so they do not sum to the published movement.
Spencer Jones	RF	Hitswhole staff	0.837 → 0.837	-0.000	0.2169	0.2177n=283 Average	0.2316n=698 Below average	0.2325	+0.003	+0.056	+0.000 ok
		split	starter x1.0679 at 0.515 share (log5 0.2325 vs his 0.2316) + bullpen x0.9276 at 0.484 share (log5 0.2020 vs pen 0.2011) = x0.9999 \| starter alone would have been +0.057
		Strikeoutswhole staff	1.244 → 1.135	-0.110	0.2212	0.3460n=283 Poor	0.1539n=698 Poor	0.2530	+0.478	-0.258	-0.098 ok
		split	starter x0.7313 at 0.515 share (log5 0.2530 vs his 0.1539) + bullpen x1.1036 at 0.484 share (log5 0.3819 vs pen 0.2490) = x0.9117 \| starter alone would have been -0.334
		Walkswhole staff	0.300 → 0.203	-0.097	0.0889	0.0885n=283 Average	0.0456n=698 Very good	0.0454	-0.002	-0.166	+0.001 ok
		split	starter x0.5127 at 0.515 share (log5 0.0454 vs his 0.0456) + bullpen x0.8489 at 0.484 share (log5 0.0751 vs pen 0.0755) = x0.6756 \| starter alone would have been -0.146
		Home runswhole staff	0.160 → 0.161	+0.001	0.0304	0.0348n=283 Average	0.0302n=698 Average	0.0346	+0.017	-0.001	-0.000 ok
		split	starter x0.994 at 0.515 share (log5 0.0346 vs his 0.0302) + bullpen x1.0197 at 0.484 share (log5 0.0355 vs pen 0.0310) = x1.0065 \| starter alone would have been -0.001
Neutralising Spencer Jones to the qualified-population mean moves the club's KPI run forecast by -0.1656 runs, and the weighted stage by -0.0629. Interactions make these non-additive across players, so they do not sum to the published movement.
José Caballero	3B	Hitswhole staff	0.822 → 0.822	-0.000	0.2169	0.2115n=451 Average	0.2316n=698 Below average	0.2259	-0.019	+0.051	-0.001 ok
		split	starter x1.0684 at 0.515 share (log5 0.2259 vs his 0.2316) + bullpen x0.927 at 0.484 share (log5 0.1960 vs pen 0.2011) = x0.9999 \| starter alone would have been +0.056
		Strikeoutswhole staff	0.826 → 0.747	-0.078	0.2212	0.2397n=451 Average	0.1539n=698 Poor	0.1679	+0.064	-0.234	-0.015 ok
		split	starter x0.7007 at 0.515 share (log5 0.1679 vs his 0.1539) + bullpen x1.1225 at 0.484 share (log5 0.2690 vs pen 0.2490) = x0.9051 \| starter alone would have been -0.247
		Walkswhole staff	0.213 → 0.143	-0.070	0.0889	0.0694n=451 Below average	0.0456n=698 Very good	0.0352	-0.068	-0.151	+0.032 ok
		split	starter x0.5075 at 0.515 share (log5 0.0352 vs his 0.0456) + bullpen x0.8463 at 0.484 share (log5 0.0588 vs pen 0.0755) = x0.6716 \| starter alone would have been -0.105
		Home runswhole staff	0.125 → 0.126	+0.001	0.0304	0.0308n=451 Average	0.0302n=698 Average	0.0306	+0.001	-0.001	0 ok
		split	starter x0.9939 at 0.515 share (log5 0.0306 vs his 0.0302) + bullpen x1.0198 at 0.484 share (log5 0.0314 vs pen 0.0310) = x1.0064 \| starter alone would have been -0.001
Neutralising José Caballero to the qualified-population mean moves the club's KPI run forecast by 0.2890 runs, and the weighted stage by 0.1098. Interactions make these non-additive across players, so they do not sum to the published movement.
STARTER Max Fried20.9702 BF	Starter	Hits	3.792 → 3.962	+0.171	0.2169	0.1991n=418 Good	0.22649/9 qualified lineup mean	0.2081	-0.372	+0.200	-0.012 ok
		Strikeouts	4.580 → 4.022	-0.559	0.2212	0.2303n=418 Average	0.19409/9 qualified lineup mean	0.2022	+0.191	-0.572	-0.017 ok
		Walks	1.414 → 1.490	+0.075	0.0889	0.0799n=418 Average	0.09369/9 qualified lineup mean	0.0842	-0.188	+0.099	-0.009 ok
		Home runs	0.300 → 0.331	+0.031	0.0304	0.0203n=418 Very good	0.03359/9 qualified lineup mean	0.0224	-0.211	+0.065	-0.021 ok
RELIEVER Paul Blackburn2.9202 BF	Middle Reliever (High Leverage)	Hits	0.811 → 0.848	+0.036	0.2169	0.1987n=329 Good	0.22649/9 qualified lineup mean	0.2076	-0.053	+0.028	-0.002 ok
		Strikeouts	0.773 → 0.676	-0.097	0.2212	0.2046n=329 Average	0.19409/9 qualified lineup mean	0.1790	-0.048	-0.080	+0.005 ok
		Walks	0.470 → 0.495	+0.025	0.0889	0.0946n=329 Average	0.09369/9 qualified lineup mean	0.0995	+0.017	+0.014	+0.001 ok
		Home runs	0.068 → 0.075	+0.007	0.0304	0.0265n=329 Good	0.03359/9 qualified lineup mean	0.0292	-0.011	+0.009	-0.001 ok
RELIEVER Ryan Weathers2.7864 BF	Middle Reliever	Hits	0.793 → 0.828	+0.035	0.2169	0.2248n=568 Below average	0.22649/9 qualified lineup mean	0.2346	+0.022	+0.027	+0.001 ok
		Strikeouts	0.732 → 0.645	-0.087	0.2212	0.2480n=568 Good	0.19409/9 qualified lineup mean	0.2183	+0.075	-0.076	-0.007 ok
		Walks	0.415 → 0.438	+0.022	0.0889	0.0717n=568 Good	0.09369/9 qualified lineup mean	0.0756	-0.048	+0.013	-0.002 ok
		Home runs	0.119 → 0.131	+0.012	0.0304	0.0311n=568 Average	0.03359/9 qualified lineup mean	0.0343	+0.002	+0.009	+0.000 ok
RELIEVER David Bednar2.6402 BF	Reliever	Hits	0.535 → 0.559	+0.024	0.2169	0.2027n=277 Good	0.22649/9 qualified lineup mean	0.2118	-0.037	+0.025	-0.001 ok
		Strikeouts	0.693 → 0.612	-0.081	0.2212	0.2624n=277 Good	0.19409/9 qualified lineup mean	0.2316	+0.109	-0.072	-0.009 ok
		Walks	0.207 → 0.218	+0.011	0.0889	0.0783n=277 Good	0.09369/9 qualified lineup mean	0.0825	-0.028	+0.012	-0.001 ok
		Home runs	0.064 → 0.071	+0.007	0.0304	0.0243n=277 Very good	0.03359/9 qualified lineup mean	0.0268	-0.016	+0.008	-0.002 ok
RELIEVER John Schreiber2.6137 BF	Middle Reliever (High Leverage)	Hits	0.765 → 0.799	+0.034	0.2169	0.2088n=246 Average	0.22649/9 qualified lineup mean	0.2181	-0.021	+0.025	-0.001 ok
		Strikeouts	0.657 → 0.573	-0.085	0.2212	0.1770n=246 Below average	0.19409/9 qualified lineup mean	0.1541	-0.116	-0.071	+0.011 ok
		Walks	0.338 → 0.355	+0.018	0.0889	0.0991n=246 Below average	0.09369/9 qualified lineup mean	0.1043	+0.027	+0.012	+0.001 ok
		Home runs	0.090 → 0.099	+0.009	0.0304	0.0297n=246 Average	0.03359/9 qualified lineup mean	0.0327	-0.002	+0.008	-0.000 ok
RELIEVER Brent Headrick2.511 BF	Reliever	Hits	0.534 → 0.557	+0.024	0.2169	0.2125n=326 Average	0.22649/9 qualified lineup mean	0.2219	-0.011	+0.024	-0.000 ok
		Strikeouts	0.716 → 0.634	-0.082	0.2212	0.2849n=326 Very good	0.19409/9 qualified lineup mean	0.2524	+0.160	-0.068	-0.013 ok
		Walks	0.213 → 0.224	+0.011	0.0889	0.0848n=326 Average	0.09369/9 qualified lineup mean	0.0893	-0.010	+0.012	-0.001 ok
		Home runs	0.060 → 0.067	+0.006	0.0304	0.0240n=326 Very good	0.03359/9 qualified lineup mean	0.0265	-0.016	+0.008	-0.002 ok
RELIEVER Will Warren1.7378 BF	Reliever	Hits	0.399 → 0.416	+0.017	0.2169	0.2293n=673 Below average	0.22649/9 qualified lineup mean	0.2392	+0.022	+0.017	+0.001 ok
		Strikeouts	0.403 → 0.354	-0.049	0.2212	0.2320n=673 Average	0.19409/9 qualified lineup mean	0.2038	+0.019	-0.047	-0.002 ok
		Walks	0.140 → 0.148	+0.007	0.0889	0.0807n=673 Average	0.09369/9 qualified lineup mean	0.0850	-0.014	+0.008	-0.001 ok
		Home runs	0.054 → 0.059	+0.005	0.0304	0.0309n=673 Average	0.03359/9 qualified lineup mean	0.0340	+0.001	+0.005	+0.000 ok
RELIEVER Ryan Yarbrough0.3469 BF	Reliever	Hits	0.072 → 0.075	+0.003	0.2169	0.2080n=253 Average	0.22649/9 qualified lineup mean	0.2172	-0.003	+0.003	-0.000 ok
		Strikeouts	0.067 → 0.059	-0.009	0.2212	0.1943n=253 Below average	0.19409/9 qualified lineup mean	0.1696	-0.009	-0.009	+0.001 ok
		Walks	0.030 → 0.031	+0.002	0.0889	0.0853n=253 Average	0.09369/9 qualified lineup mean	0.0898	-0.001	+0.002	-0.000 ok
		Home runs	0.008 → 0.009	+0.001	0.0304	0.0237n=253 Very good	0.03359/9 qualified lineup mean	0.0261	-0.002	+0.001	-0.000 ok

Published pitching, integer draw beside the continuous model line

Pitcher	Role	BF	IP	expected IP	H int/cont	SO int/cont	BB int/cont	ER int/cont	Rate source
Nick Martinez	Starter	18.0	4.1	4.304	4 / 4.238	4 / 3.171	2 / 0.896	2 / 1.939	all_players_stats_recent_season_blend
Cam Booser	Setup Man	2.5	0.2	0.668	0 / 0.368	1 / 0.713	0 / 0.127	0 / 0.121	all_players_stats_recent_season_blend
Bryan Baker	Closer	3.8	1.0	1.002	1 / 0.540	1 / 0.931	0 / 0.287	0 / 0.108	all_players_stats_recent_season_blend
Cole Sulser	Middle Reliever	4.2	2.2	1.002	4 / 0.856	2 / 1.079	0 / 0.311	2 / 0.385	all_players_stats_recent_season_blend
Max Fried	Starter	21.0	5.1	5.255	4 / 3.792	7 / 4.580	3 / 1.414	2 / 1.510	all_players_stats_recent_season_blend
Ryan Weathers	Middle Reliever	4.2	1.0	1.006	0 / 0.793	0 / 0.732	0 / 0.415	0 / 0.414	all_players_stats_recent_season_blend
Paul Blackburn	Middle Reliever (High Leverage)	5.3	1.1	1.341	1 / 0.811	0 / 0.773	0 / 0.470	1 / 0.157	all_players_stats_recent_season_blend
John Schreiber	Middle Reliever (High Leverage)	4.1	1.1	1.006	1 / 0.765	1 / 0.657	1 / 0.338	0 / 0.285	all_players_stats_recent_season_blend

Audit: 1 CHECK(S) FAILED

Invariant	Verdict	Evidence
blend arithmetic away_runs	PASS	3.0000 + 0.34 x (4.3610 - 3.0000) = 3.4627, recorded 3.4627
blend arithmetic home_runs	PASS	4.0000 + 0.38 x (4.5253 - 4.0000) = 4.1996, recorded 4.1996
blend arithmetic home_win	PASS	0.5665 + 0.00 x (0.5817 - 0.5665) = 0.5665, recorded 0.5665
published rounding faithful to the blend	PASS	blended 3.4627-4.1996 rounds to 3-4, layer published 3-4
published probability on the score's side	PASS	3-4 with p(home) 0.566481
Log5 additivity136 checks	PASS	136 of 136 pass
away club fields exactly one starter	PASS	1 starters in a staff of 4
away staff totals a regulation workload	FAIL	26 outs recorded, 24 expected
away starter carries a continuous line	PASS	expected_IP 4.3037
away starter batters faced inside the real band	PASS	Nick Martinez projected 18.0 batters faced; real starters run p01-p99 4-30, median 23, mean 21.75
away reliever batters faced inside the real band3 checks	PASS	3 of 3 pass
home club fields exactly one starter	PASS	1 starters in a staff of 4
home staff totals a regulation workload	PASS	27 outs recorded, 27 expected
home starter carries a continuous line	PASS	expected_IP 5.2547
home starter batters faced inside the real band	PASS	Max Fried projected 21.0 batters faced; real starters run p01-p99 4-30, median 23, mean 21.75
home reliever batters faced inside the real band3 checks	PASS	3 of 3 pass
away hitters satisfy every coherence rule	PASS	9 hitters checked; H<=AB, HR<=H, 2B+3B+HR<=H, SO<=AB, continuous line present
home hitters satisfy every coherence rule	PASS	9 hitters checked; H<=AB, HR<=H, 2B+3B+HR<=H, SO<=AB, continuous line present

Provenance. Read only from published artifacts and stored KPI value documents. Nothing is recomputed, so this document cannot disagree with what was published. Per-player metric adjustments are a measurement of the catalog at player level and are labelled as such throughout; the pipeline itself moved team runs, the win probability, and hitter lines by one club-wide ratio.

Verdict across the slate

How to read this, and what a defect looks like

Slate overview

Per game

Stage chain: every stage that touched the score

KPI team layer: the blend, with its arithmetic

What drove the KPI forecast for away_runs

What drove the KPI forecast for home_runs

Per-player metric movement

CLE against Matt Festa (RHP)

CHW against David Sandlin (RHP)

Published pitching, integer draw beside the continuous model line

Audit: NO DEFECT FOUND

Stage chain: every stage that touched the score

KPI team layer: the blend, with its arithmetic

What drove the KPI forecast for away_runs

What drove the KPI forecast for home_runs

Per-player metric movement

LAD against Edgardo Henriquez (RHP)

ATL against Dylan Lee (LHP)

Published pitching, integer draw beside the continuous model line

Audit: NO DEFECT FOUND

Stage chain: every stage that touched the score

KPI team layer: the blend, with its arithmetic

What drove the KPI forecast for away_runs

What drove the KPI forecast for home_runs

Per-player metric movement

MIL against JoJo Romero (LHP)

SD against Joe Musgrove (RHP)

Published pitching, integer draw beside the continuous model line

Audit: 2 CHECK(S) FAILED

Stage chain: every stage that touched the score

KPI team layer: the blend, with its arithmetic

What drove the KPI forecast for away_runs

What drove the KPI forecast for home_runs

Per-player metric movement

TB against Tyler Wells (RHP)

NYY against Ryan Yarbrough (LHP)

Published pitching, integer draw beside the continuous model line

Audit: 1 CHECK(S) FAILED

What drove the KPI forecast for `away_runs`

What drove the KPI forecast for `home_runs`

What drove the KPI forecast for `away_runs`

What drove the KPI forecast for `home_runs`

What drove the KPI forecast for `away_runs`

What drove the KPI forecast for `home_runs`

What drove the KPI forecast for `away_runs`

What drove the KPI forecast for `home_runs`