Analysis Lifecycle
Full lifecycle of an options analysis from flag detection through trade management. Every data artifact, FK relationship, and training signal.
Click any stage to expand details — data refreshed live from Supabase
Score >= threshold triggers analysis queue
Draft written, human initiates review
Human confirms 'let's trade this'
Ongoing check-ins on open position
Trade closed, expired, or stopped out
Live Data Health
Total Flags
20,452
202 symbols flagged (score 2+)
Latest Scan
2026-04-24
7:1 6:12 5:26 4:30 3:69 2:260
Research
85
7d / 0r
Avg Flag Score
14.3
of drafted tickers
Paper Trades
55
5 open, 50 closed
Win Rate
33%
avg P&L: -19.1%
Conversations
7
41 total turns
Training Pairs
0
41 conv turns, 7 insights
options_flags
options_research
options_paper_trades
conversations
Foreign Key Relationships
Every stage adds to the training corpus. These are the supervised training pairs and fine-tuning datasets available from the lifecycle pipeline.
| Training Use Case | Source Table(s) | Input | Output | Volume |
|---|---|---|---|---|
| Draft quality improvement | options_research | draft_snapshot (MJS output) | Reviewed version (headline, writeup, structures) | 0 |
| RAG retrieval tuning | options_research | rag_citations with citation_useful | Boolean: was citation useful? | 0 |
| Conversation fine-tuning | conversations | turns (human messages) | turns (agent responses) + summary | 41 |
| Domain expertise capture | conversations | Conversation context | summary.human_insights | 7 |
| Entry decision quality | options_paper_trades | Entry snapshot (flags, IV, regime) | outcome, estimated_pnl_pct | 50 |
| Conviction calibration | conversations | conviction_start + context | conviction_end + eventual outcome | 41 |
| Flag predictive value | options_flagsoptions_paper_trades | Flag combo at entry | Trade outcome | 50 |