Oxotradex: Autonomous Options Alpha Agent

Oxotradex is an end-to-end, production-ready quantitative algorithmic trading system engineered specifically for the Alpaca AI Trading Agents Hackathon. The Problem It Solves Most autonomous AI trading bots attempt directional equity speculation (e.g., prompting an LLM to predict whether SPY or NVDA will rise or fall). In financial markets, directional price forecasting over short horizons is notoriously low-edge, suffering from slippage, black swan drawdowns, and unpredictable LLM hallucinations. Furthermore, traditional agent frameworks grant LLMs unconstrained authority to place orders, creating immense tail-risk vulnerability. The Quantitative Edge: Monetizing the Volatility Risk Premium Instead of gambling on directional price moves, our agent exploits a structural anomaly known as the Volatility Risk Premium (VRP)—the empirical reality that Implied Volatility (IV) persistently trades higher than Realized Volatility (RV). By underwriting probability through out-of-the-money (OTM) credit spreads (Bull Put Spreads, Bear Call Spreads) and Iron Condors with 21–35 Days to Expiration (DTE), the agent systematically captures accelerated time decay (Θ), achieving an initial 70%–85% statistical probability of expiring out of the money. Every position is strictly defined-risk: long protective wings are placed on every spread to eliminate unbounded tail loss. 8 Inviolable Deterministic Risk Gates: Hardcoded in Python and verified via automated test suites: Gate 1: Emergency Kill-Switch Gate 2: Daily Circuit Breaker (−2.5 daily loss halts all entries) Gate 3: Capacity Limit (max 5 open positions) Gate 4: Defined-Risk Invariant (zero naked options allowed) Gate 5: Short Delta Ceiling (|Δ|≤ 0.30) Gate 6: Minimum Net Credit ($\ge 0.20 / share) Gate 7: Liquidity Filter (bid-ask spread $\le 0.25, open interest ≥50) Gate 8: Dynamic Capital Sizing (maximum loss per trade capped at ≤ 3.0 of total equity)