The Thesis

Most systematic strategies treat risk management as a constraint applied after the portfolio is constructed. OVRWCH inverts this. The defense layer operates first, independently, and cannot be overridden by the alpha engine.

This is not a philosophical preference. It is a structural decision based on an asymmetric observation: a 50% drawdown requires a 100% gain to recover. A 33% drawdown requires a 50% gain. The math of loss recovery is nonlinear, and it punishes systems that treat defense as optional.

Three Independent Layers

OVRWCH runs three defense mechanisms simultaneously. Each catches a different category of failure. The specific parameters and thresholds are proprietary, but the architecture is documented here because we believe the approach itself is sound and worth understanding.

Layer 1: Position-Level Protection. Each position has an independent trailing stop measured from its high-water mark. If a single holding breaches its threshold, it is liquidated regardless of what the rest of the portfolio is doing. This catches single-name blowups that might not appear in portfolio-level metrics until they compound.

Layer 2: Portfolio-Level Drawdown Control. A state machine monitors the portfolio’s total drawdown from peak NAV. As drawdown deepens, the system progressively reduces gross exposure and restricts the universe to defensive assets. At maximum severity, the portfolio goes to cash. Re-entry requires sustained confirmation of recovery — the system does not rush back in.

Layer 3: Volatility Targeting. The portfolio’s realized volatility is measured continuously and compared to a target. When realized vol exceeds the target, gross exposure is automatically reduced. When vol is below target, exposure can increase. This creates a natural delevering mechanism during stress periods without requiring the drawdown controller to activate.

The Interaction

The three layers are multiplicative. In a severe event, all three may activate simultaneously: individual positions hit stops (Layer 1), portfolio drawdown triggers deleveraging (Layer 2), and vol-targeting reduces exposure further (Layer 3). The combined effect is rapid, automatic risk reduction.

No manual override exists. The defense layer cannot be turned off, loosened, or bypassed by the scoring engine. If the alpha model’s top-ranked ETF triggers a circuit breaker, it is sold. The model’s opinion is irrelevant.

The Cost

This architecture has a measurable cost: reduced returns during recoveries and false alarms. When the regime detector classifies an environment as defensive but the market subsequently rises, the portfolio misses upside. This drag is the explicit premium paid for drawdown protection.

Over full market cycles that include real drawdowns, the tradeoff has historically been favorable. The capital preserved during genuine stress compounds during the recovery that follows. Avoiding a 34% drawdown and re-entering at the bottom produces better terminal wealth than riding the full cycle unprotected — even if some individual months underperform.

This note describes OVRWCH’s defense architecture at a conceptual level. Specific thresholds, state transition rules, and implementation parameters are proprietary.