# H05_E_EXPLORATORY_FREEZE_v0_1

**Created:** 2026-09-13T23:07:01Z  
**Artifact:** `H05_E_EXPLORATORY_FREEZE_v0_1.json`  
**sha256:** `2f3fca7925519cf5a743488a89975b195f25c9df311d9ff5e648ac25c15b4292`  
**Hub job:** `77356229-193e-4693-a90c-cae8dcb0bc4a`  
**James:** exploratory-first keep-going — parallel cheap freeze→run DIFFERENT from H04-E PROMOTED crude; no H01/FT900/ESR

## STATUS
| Field | Value |
|-------|-------|
| STATUS | EXPLORATORY |
| VINTAGE_STATUS | UNVERIFIED/INCOMPLETE |
| CLAIM_STRENGTH | LOW |
| CAPITAL_USE | PROHIBITED |
| MODEL_CHANGE | NO |
| REAL_CAPITAL | PROHIBITED |
| TARGETS_OPENED_AT_FREEZE | NO |

## Why runnable today
EIA `NW2_EPG0_SWO_R48_BCF` hist XLS HTTP 200 (sha `3449b602cf3c7ff5…`) + STEO `*_base.xlsx` archives already fetched for H04 (141/141). Physical weekly NG storage → STEO official-outlook inventory lag. **Not** crude/STEO (H04), not H01 corn, not FT900/ESR digs.

## Locked (all 10)

### 1. Eligible reconstructed sample
- **140** adjacent calendar-consecutive STEO pairs (jan2015→sep2026)
- Forecast dates: mix of PARSED + SCHEDULE_PROXY (same issue clock as H04)
- Every obs **VINTAGE_UNVERIFIED**

### 2. Predictor
`nw2_l48_wg_surprise_mean_inter_issue` — mean weekly Δ working-gas (bcf) surprise vs training week-of-year norms (Lower 48).  
Release proxy = week_end + 6d (WNGSR Thursday).

### 3. Predictor window
`origin_forecast_date < release_proxy <= destination_forecast_date`.

### 4. Normalization
Training-only (destination years ≤2019). No holdout peek.

### 5. Target
`NGWGPUS_dest[M] − NGWGPUS_origin[M]` (bcf), M = destination year-month. **Values closed at freeze.**

### 6. Baseline (Model A)
Intercept + destination-month seasonality + prior revision (+ missing flag).

### 7. Chronological split
- Development destination years **2015–2019** (N=59)
- Holdout destination years **2020–2026** (N=81)

### 8. Metric
**RMSE** primary; MAE + directional accuracy secondary.

### 9. Success threshold
**≥ 5.0%** relative holdout RMSE improvement.  
**KILL** (<1% or neg / N<8) · **HOLD** ([1%,5%)) · **EXPLORATORY_PROMOTE** (≥5%; NOT capital)

### 10. Trial budget
ONE spec: Model A baseline; Model B = A + one NW2 L48 WG feature. No rescue.

## Expected sign
**POSITIVE** (storage-build surprises → upward NGWGPUS revision).

## Coverage caveat (frozen)
Physical series = Lower-48; STEO `NGWGPUS` = US total (incl. small AK). Frozen as known limitation.

## Next
1. Post freeze to Hub (/docs + comment + /reply chatgpt+all). Never /james.
2. Only then open NGWGPUS and run once.