# H06_E_EXPLORATORY_FREEZE_v0_1

**Created:** 2026-09-13T23:18:04Z  
**Artifact:** `H06_E_EXPLORATORY_FREEZE_v0_1.json`  
**sha256:** `c24ae989f20e3c160a8ad08852cda170a92729767f632b1e432d4f798263fad9`  
**Hub job:** `77356229-193e-4693-a90c-cae8dcb0bc4a`  
**James:** post-H04-MECHANICAL / H05-inventory-class — DIFFERENT cheap freeze→run (weather→STEO sales); no inventory redo; 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
NOAA CAG Contiguous-US HDD+CDD CSV HTTP 200 + **141/141** STEO `*_base.xlsx` already fetched. Weather→official electricity sales outlook lag. Avoids inventory-copy class and blocked archive lanes.

## 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/H05)
- Every obs **VINTAGE_UNVERIFIED**

### 2. Predictor
`noaa_dd_surprise_mean_inter_issue` — mean (HDD+CDD) surprise vs training calendar-month norms.  
Release proxy = month_end + 10d (SCHEDULE_PROXY).

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

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

### 5. Target
`sales_dest[M] − sales_origin[M]` with alias **ELTCTWH** else **EXTCPUS**, 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 NOAA DD feature. No rescue.

## Expected sign
**POSITIVE** (HDD+CDD upside surprises → upward retail-sales revision).

## Mechanical risks (if promote)
STEO weather-table mediation; EPM actuals absorption; NOAA timing proxy; EXTCPUS/ELTCTWH alias.

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