Philosophy Three Core Laws Systematize everything — Repeatable beats heroic. If we can’t express it as rules, we don’t run it. Breadth over cleverness — Compute is cheap; missed regimes are expensive. Explore wide, simulate hard, then prune. Probabilities over dogma — Size to odds, not opinions. Promote on proof; bench on breach. How we enforce this →
Data Clean, consistent inputs Continuous futures & multi-asset ETFs Macro fundamentals & standard technicals
Notebooks Raw → features → backtests Narrative, reproducible templates Promotion staging & A/B tracking
Regimes Mogul vs. Groomer Ski Run Regime Signal Type Why it works Mogul Low-Vol / Range Mean-reversion / Option writing Smooth “corduroy”—collect theta with controlled risk. Groomer High-Vol / Breakout Trend-following / Option buying Ride gamma/vega in accelerating markets.
Policy & Guardrails Calibration & promotion Flip: prefer prob ↔ bypass based on rolling ECE/Brier thresholds Promote: 20-day realized edge > 0 and hit ≥ 45% Demote: breach a rule → strategy cools off Same rules for portfolios Model promotion/demotion with tracking-error bounds Tax overlay thresholds & exception handling Audit log of changes How this shows up in Advisory