Price series, volatility and positioning data are aligned across intraday and higher-timeframe horizons. All data is stored point-in-time to ensure backtests remain reproducible and free from look-ahead bias. Regime characterisation - drawing on volatility, skewness, kurtosis and term-structure - precedes any modelling work.

Each strategy originates as a formal hypothesis, subjected to econometric and time-series testing before any capital is committed. Trend, volatility and positioning features are used to classify when conditions are structurally favourable versus fragile. Prototype strategies are evaluated on risk-adjusted returns, drawdown profiles, tail behaviour and regime sensitivity.

Strategies must pass out-of-sample and walk-forward tests under realistic cost and slippage assumptions. Once deployed, all strategies operate within explicit leverage, concentration and drawdown limits. Live performance is continuously tracked against research benchmarks, with material deviations triggering a formal review process.​