# H04-E pick justification (one page)

**Date:** 2026-09-13T22:05:06Z UTC · **Hub job:** `77356229-193e-4693-a90c-cae8dcb0bc4a`  
**Ops:** After H01-E **KILL** — stop archive-first shopping; freeze→run on fetchable data.

## Chosen child
**H04-E** — EIA weekly commercial crude stocks (`WCESTUS1`) → STEO monthly `COSXPUS` (crude excl. SPR EOP inventories) revision across adjacent STEO issues.

## Why runnable today
| Leg | Source | Status |
|-----|--------|--------|
| Physical | EIA hist XLS WCESTUS1 | HTTP 200; sha `ef57fa201d252ac4…` (revised hist → vintage unverified) |
| Outlook | STEO archives `*_base.xlsx` | **141/141** jan2015–sep2026 downloaded |

## Why not other lanes this cycle
- **H03-D ESR/AAR** and **H02-C FT900**: first-print archive blockers — James: do not dig.
- **H01**: do not reopen.
- **H02-B EIA-930→STEO**: 930 first-prints UNKNOWN; stocks→STEO is the runnable energy pair.
- **NASS**: Quick Stats key 401 here; unnecessary given EIA.

## Test shape (H01-E discipline)
Frozen baseline vs +one feature; RMSE; ≥5% bar; EXPLORATORY / CAPITAL PROHIBITED / CLAIM_STRENGTH LOW / MODEL_CHANGE=NO.

## Mechanism (hypothesis only)
Weekly physical commercial crude stock changes → next STEO official inventory path may revise. Incremental predictiveness only — not capital.