Thesis. Math can detect portfolio-manager behavioral bias from trade ledgers alone. An LLM should only narrate a structured scorecard — in language a risk committee will actually read — not hunt for patterns in raw blotter data.
1. Problem
Portfolio managers exhibit predictable behavioral biases: disposition effect (selling winners too early, holding losers too long), overconfidence (excess turnover), and cost-basis anchoring (refusing to sell below purchase price). These biases destroy risk-adjusted returns, but they are rarely measured systematically. Most risk reports flag P&L and VaR; they do not flag why the P&L looks the way it does.
2. Approach
Synthetic Market → Control & Biased Traders → Deterministic Evaluator → Scorecard JSON → Fine-Tuned Coach (Qwen3-8B QLoRA) → CRO Memo
Every stage is deterministic and auditable except final coach generation. The LLM is not asked to find bias — it receives a structured scorecard and narrates it.
Biases detected (v1)
| Bias | Metric | Detection |
|---|---|---|
| Disposition effect | PGR / PLR | Realized gains ≫ realized losses; paper losses held |
| Overconfidence | Annualized turnover | Excess trading vs baseline; trade count |
| Cost-basis anchoring | Underwater share + sell-below-cost | Positions held underwater; zero sells below cost |
Phase 3 gate (non-negotiable)
Before any LLM step, the evaluator had to pass a hard gate:
- 100% recovery of all three injected biases on the biased trader
- Zero false positives on the control trader
- Detection uses only prices, positions, and trades — never injection tags
Result: PASS. Control clean; biased flagged on all three dimensions.