# H05-E pick justification (one page)

**Date:** 2026-09-13T23:07:01Z UTC · **Hub job:** `77356229-193e-4693-a90c-cae8dcb0bc4a`  
**Ops:** James exploratory-first keep-going; H04-E already PROMOTED with separate confound worker — run a DIFFERENT parallel cheap exploratory.

## Chosen child
**H05-E** — EIA weekly Lower-48 working gas underground storage (`NW2_EPG0_SWO_R48_BCF`) → STEO monthly `NGWGPUS` (US end-of-period working natural gas inventories) revision across adjacent STEO issues.

## Why runnable today
| Leg | Source | Status |
|-----|--------|--------|
| Physical | EIA hist XLS NW2_EPG0_SWO_R48_BCF | HTTP 200; sha `3449b602cf3c7ff5572da87630598cdef278bb9191556f5fda43a7ace40703a7` (revised hist → vintage unverified) |
| Outlook | STEO archives `*_base.xlsx` | **141/141** jan2015–sep2026 already on disk (shared with H04 lane) |

## Why not other lanes this cycle
- **H04-E crude→STEO:** already EXPLORATORY_PROMOTE; confound worker separate — do not redo crude.
- **H01 corn / H03 ESR / H02 FT900:** blocked or killed / archive digs — James: do not dig.
- **EIA-930→STEO:** first-prints UNKNOWN; weekly storage→STEO is the cheaper runnable energy pair that is still *different* from H04.

## Test shape (H01-E / H04-E discipline)
Frozen baseline vs +one feature; RMSE; ≥5% bar; EXPLORATORY / CAPITAL PROHIBITED / CLAIM_STRENGTH LOW / MODEL_CHANGE=NO.

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