Options Edge Lab
Live OTM mispricing surface + perfect-foresight replay across the full THv2 signal stack. Composes the prescience GBM forecast, BSM Greeks engine, SVI volatility surface fit, IV term-structure classifier, signal-confluence matrix, mosaic-strength aggregator, factor-library, net-outcome compounder, and the foresight replay engine.
1. Today's Edge Surface
Strike-grid scan: market mid vs BSM-at-prescience-σ fair value. Surfaces highest-conviction OTM edges ranked by |mispricing| × volume.
Top-10 OTM mispricings (|mispricing| × volume)
2. Strategy Sandbox
Build a strategy + see cost / max loss / max gain / breakeven(s) / POP under prescience GBM (real-world prob) + Kelly fraction + expected edge bps + aggregated Greeks + payoff curve over [0.7·S, 1.3·S].
3. Perfect-Foresight Replay
Pick an asof date + horizon; the replay scores every strategy in the menu against the realized price path and surfaces the optimal play. Foresight gap chip = (perfect − THv2-pick) realized return diff in bps.
4. IV Surface
Per-expiry: DTE / ATM IV / skew slope / SVI raw parameters (a, b, ρ, m, σ) / regime chip. Composes THV2VolSurface.fit() + THV2_IVTermStructure.report().
5. Strategy Optimizer
Composes prescience forecast + signal-confluence classifier + IV term-structure regime into a 4-axis bucket (direction × vol × term-structure × confluence) and surfaces top-3 strategies + Kelly-fraction sizing + entry/exit rules per Tastytrade 1SD 45-DTE convention + Diether-Malloy-Scherbina 2002 (disagreement) + Johnson 2017 (term-structure).
6. Foresight Calibration Cohort
Sweep the perfect-foresight replay across the last N≥10 historical days to surface where THv2 systematically misses perfect foresight — per-regime + per-strategy hit-rate + THv2-vs-perfect agreement rate + Brier-calibration proxy. Bar fetches dominate latency; rate-limit 10 req/min/IP; 24h cache per (ticker, range, horizon).
- BSM pricing + Greeks (Black-Scholes 1973 J.Pol.Econ. 81(3):637; Hull 2017 ch.17). Δ_call = e^(−qT)·N(d1); Γ = e^(−qT)·φ(d1)/(S·σ·√T); Vega = S·e^(−qT)·φ(d1)·√T/100.
- SVI raw parametrization (Gatheral-Jacquier 2014 "Arbitrage-free SVI volatility surfaces", Quantitative Finance 14(1):59). w(k) = a + b{ρ(k−m) + √((k−m)² + σ²)}. Fit grid-search + LM refine when flag on.
- Kelly sizing capped at quarter-Kelly (Kelly 1956 Bell Syst. Tech. J. 35:917; Thorp 1962; MacLean-Thorp-Ziemba 2010).
- POP under prescience GBM = real-world probability (NOT BSM risk-neutral N(d2)).
- Foresight replay scores every strategy in the menu against the realized price path. "Perfect" = optimal-within-menu; not omniscient (Johnson 2017 J.Portfolio Management 43:5).
- Tastytrade managed-loss convention for Phase 2 strategies (butterfly 50% debit, iron_condor 2× credit; McMillan 2012 ch.16; Sosnoff-Battista 2014).
- Regime-conditional strategy ranker per Tastytrade 1SD 45-DTE convention + Diether-Malloy-Scherbina 2002 ("Differences of Opinion and the Cross Section of Stock Returns" JF 57(5):2113) + Johnson 2017 J.Portfolio Management 43:5 (IV term-structure). 11-bucket regime → top-3 strategy mapping HEURISTIC, not Brier-validated against THv2 cohort specifically.
- Cohort calibration sweeps last N≥10 days; per-strategy hit-rate; THv2-vs-perfect agreement rate; Brier-calibration. brier_calibration is gap-bps-style proxy, NOT strict Brier loss on probability forecasts. thv2_vs_perfect_agreement_rate is binary match (no partial credit for ranked-#2). Synthetic-BSM-at-prescience-σ chain limits accuracy when actual market σ-skew differs significantly.
Methodology hash: options-edge-lab-render-2026-06-01-v3 · Optimizer: strategy-optimizer-regime-conditional-2026-06-01-v1 · Cohort: api-options-foresight-cohort-2026-06-01-v1