---
title: "Do Efron & Hastie's CASI Ch. 7 confirm the James-Stein domination statement and the .0425 vs .0218 baseball SSE figures?"
type: "question"
status: "answered"
date_answered: "2026-07-14T00:00:00.000Z"
answered_log: ["2026-07-14: Answered by [[claim-james-stein-estimator-uniformly-dominates-the-sample-mean]], [[claim-efron-baseball-shrinkage-halved-batting-average-prediction-error]] — both re-fetched and verbatim-confirmed against CASI Ch. 7 in the 2026-07-12 cross-model audits: uniform domination, inadmissibility (dim ≥ 3), 'rude shock to the statistical world', and the baseball worked example (18 players, first 90 at-bats, rest-of-season truth, SSE MLE .0425 / JS .0218) all verified."]
writer_model: "claude-opus-4-8"
date_raised: "2026-07-11T00:00:00.000Z"
tags: ["statistics","shrinkage-estimation","james-stein","verification","quant-check"]
---


Two claim-notes promoted from
`10-inbox/raw/2026-07-09-hop-steins-paradox.md` rest on Efron & Hastie's
*Computer Age Statistical Inference*, Ch. 7, but the source was read by the
capturing pass and not independently re-fetched during headless promotion:

- [[claim-james-stein-estimator-uniformly-dominates-the-sample-mean]]
- [[claim-efron-baseball-shrinkage-halved-batting-average-prediction-error]]

**What to verify.** Re-open the author-hosted PDF
(https://efron.ckirby.su.domains/other/CASI_Chap7_Nov2014.pdf) and confirm:

1. That the chapter states the sample mean is inadmissible / uniformly dominated
   by James-Stein for dimension ≥ 3 under squared-error loss, and that "rude
   shock to the statistical world" is the chapter's own phrasing.
2. The baseball worked example: 18 MLB players, 1970 season, first-90-at-bats
   MLE vs shrinkage, and the exact figures **"Sum of squared errors for
   predicting TRUTH: MLE .0425, JS .0218."** Confirm the numbers, the player
   count (18), and that the "truth" is rest-of-season performance.

**Why it matters.** These are a mechanism claim and a specific quantitative claim
resting on a Tier-1 primary. The floor is met on paper; this closes the loop by
re-reading the primary so both notes can move off `capture-verified`.

**Candidate next move.** WebFetch the PDF, locate the batting-average table and
the inadmissibility statement, quote both verbatim, then upgrade the two notes'
`audit_status` and `status` accordingly.


## Progress log

- 2026-07-14: Answered by [[claim-james-stein-estimator-uniformly-dominates-the-sample-mean]], [[claim-efron-baseball-shrinkage-halved-batting-average-prediction-error]] — both re-fetched and verbatim-confirmed against CASI Ch. 7 in the 2026-07-12 cross-model audits: uniform domination, inadmissibility (dim ≥ 3), 'rude shock to the statistical world', and the baseball worked example (18 players, first 90 at-bats, rest-of-season truth, SSE MLE .0425 / JS .0218) all verified.