Risk Management

The only thing we can control is risks and costs.

Introduction

The framework makes explicit distinctions between controllable and uncontrollable risks. We cannot predict market direction, but we can control our exposure to that uncertainty.

Controllable risk factors

Our position sizing rules directly control our exposure to any single instrument or strategy. Maximum allocation limits prevent excessive concentration. Leverage constraints are similarly explicit. While HyperLiquid allows significant leverage, our framework caps gross leverage at appropriate leverage limits for each portfolio. This prevents catastrophic losses from unexpected volatility spikes.

Transaction costs, while not eliminated, can be minimised through careful execution as discussed above. Funding fees are implicitly optimised for in our carry strategies.

Uncontrollable risk factors

We accept that markets can move against us. Our volatility targeting doesn't prevent losses, it makes losses predictable in magnitude. A 15% volatility target means we expect roughly 1% daily moves, with occasional 3% moves during stress.

Correlations can spike unexpectedly, making diversification temporarily ineffective. Strategy performance can deteriorate as market regimes shift. These are inherent uncertainties we manage through diversification rather than prediction.

Risk metrics and monitoring

We track several key metrics to ensure the system operates within design parameters:

Value at Risk (VaR)

Daily 95% VaR is calculated from the distribution of historical portfolio returns. This tells us the maximum loss we expect to exceed on 5% of trading days. We also use an average version of this metric over multiple observations , called Expected Shortfall (ES)

Drawdown tracking

We monitor rolling maximum drawdown with alert thresholds. If drawdown exceeds thresholds, we trigger a portfolio review. This doesn't necessarily mean intervention, but ensures human oversight during stress periods.

Signal decay analysis

Each strategy's performance is monitored on a rolling basis. If a momentum strategy stops generating positive returns for 60 consecutive days, it signals potential regime change requiring investigation.

Circuit breakers and extreme event handling

Automated safeguards protect against both system errors and market extremes.

Position limit checks

Before executing any trade, the system verifies positions remain within absolute limits. No single position can exceed a certain, pre-defined % of the portfolio value. Per-asset and gross leverage can't exceed x2. If a proposed trade would violate these constraints, it's rejected even if forecast logic suggests it.

Volatility circuit breakers

When measured volatility exceeds 3x the historical average, the system halves all position sizes temporarily. This automatic deleveraging prevents outsized losses during flash crashes or extreme events.

Liquidity monitoring

We track order book depth and recent volume for each instrument. If liquidity drops below minimum liquidity thresholds, position sizing is reduced proportionally. This prevents concentration in illiquid assets that couldn't be exited during stress.