How Algogen evaluates a candidate strategy

Strategy and scope are separate

A candidate strategy contains buy rules, sell rules, and risk management. Instrument, venue, timeframe, and date range belong to the backtest scope. This separation makes it explicit whether a change affects the rules or only the data used to test them.

Performance evidence

Performance evidence includes the equity curve, return, drawdown, win rate, profit factor, exposure, position history, and executed-order history. These are historical observations for the selected scope.

Benchmark evidence

Where the required data exists, Algogen compares the candidate strategy with buy and hold, cash, market, and seeded random-entry baselines. Missing comparison data is omitted instead of replaced with an invented figure.

Robustness evidence

Robustness checks can include rolling out-of-sample walk-forward windows, Monte Carlo resampling of real per-order returns, parameter sensitivity, delayed fills, and volatility-noise tests. Each check uses the strategy and market data available to that run.

Diagnostic evidence

Diagnostic evidence examines signal behaviour, seasonality, variance, and correlation where the run contains enough information to support the calculation.

Interpretation limits

Historical validation can expose fragility and weak assumptions, but it cannot prove that a strategy will work in the future. Results depend on data quality, costs, slippage, margin, sizing, conflict handling, and the exact rule definitions.

Read the full risk disclosure and the editorial policy used for technical guides.