calibration-reportlisted

# calibration-report — outcome engine (Analysis C1) Aggregate expectancy (`portfolio.py expectancy`) answers "am I net positive." This answers "**why**" — which entry conditions the wins/losses cluster around — so the 20-trade Phase-2 calibration gate means something more than a raw count. ## Usage ```bash python3 .claude/skills/calibration-report/calibration_report.py report # human-readable python3 .claude/skills/calibration-report/calibration_report.py report --json # machine-readable ``` Read-only, no config, no cron — run it whenever you want a cut of closed-trade outcomes by entry context. ## What it does For every CLOSED journal entry with a realized R-multiple, reads the entry's `### Data snapshot` table (RSI, sector, sentiment flag, RS vs SPY 1m) and `### Structural risk flags` section, then buckets win-rate/avg-R/sum-R by: - **RSI band at entry** (`<30`, `30-50`, `50-70`, `>=70`) - **Sentiment flag at entry** (FADE / BUY / unknown) - **RS vs SPY (1m) at entry** (leader / laggard / flat / unknown) - **Sector** - **Structural-quality flags at entry** (has flags / clean) Buckets with `n < 3` are printed but marked low-confidence — directional only, never hidden (same warn-loudly-never-block spirit as `quality_flags.py`). ## Data-capture dependency Sector / sentiment flag / RS-vs-SPY are only recorded in prospectuses created via `j.py new` **after 2026-07-02** (when `--sector`/`--sentiment-flag`/`--rs-1m` were added, threaded from the Risk Simulato