Daily Analysis

Generate a fresh PredictionRun and present the picks the way a trader would read them.

Steps

1. Confirm scope. Default to today; if the user named a date use --date YYYY-MM-DD.
2. Run analysis:

   python tippmix.py analyze            # today
   python tippmix.py analyze --date 2026-06-14
   

3. The run writes data/latest_run.json, data/predictions/<timestamp>.json, data/reports/report_<timestamp>.json, and appends data/archive/runs.jsonl.
4. Read data/latest_run.json to summarize for the user. Report per match: match header, confidence %, recommended bets (selection + confidence + odds + EV when present), and the combination buckets with their tiered HUF stakes.
5. Lead with expected value, not win probability. Flag any pick where confidence is high but EV is negative (odds too short) — that is not a bet.

Notes

• Recommendations below high_risk_confidence_threshold (50) are already discarded by the engine.
• analyze reads data/lessons.json and applies learned per-market calibration offsets — if calibration looks off, run /calibration first.
• Do NOT dump raw stats dicts to the user; the full data lives in the JSON files.
• If the API is unreachable, run python tippmix.py check to confirm before debugging deeper.