Backtesting & risk

Turn signals into costed equity curves, and measure drawdown and performance.

Risk & performance

Drawdowns and performance ratios.

Function	Description
Drawdown	Running drawdown from the cumulative peak.
Maximum drawdown	Worst drawdown experienced so far (since reset).
Rolling Calmar ratio	Annualised return divided by the worst rolling drawdown.
Rolling Conditional Value-at-Risk (Expected Shortfall)	Historical Conditional Value-at-Risk (Expected Shortfall): the mean loss in the worst alpha tail over a window.
Rolling Downside Deviation	Trailing-window downside semideviation: the RMS of returns falling below a minimum acceptable return.
Rolling hit rate	Fraction of strictly-positive samples in a trailing window.
Rolling information ratio	Annualised information ratio: Sharpe of active returns against a benchmark.
Rolling maximum drawdown	Worst peak-to-trough drawdown inside a trailing window.
Rolling Omega Ratio	Omega ratio over a window: total gains above a threshold divided by total losses below it.
Rolling Sharpe ratio	Annualised Sharpe ratio over a trailing window of returns.
Rolling Sortino ratio	Annualised Sortino ratio: Sharpe with downside-only deviation.

Backtesting

Simulate a strategy against market data into a costed equity curve: signals, bars, the trade tape, and top-of-book quotes.

Function	Description
backtest_report	Running statistics and a summary for a backtest engine's output.
Backtest a two-sided market maker against L1 quotes	Backtest a two-sided market maker against a top-of-book (L1) quote stream, into a costed equity curve.
Backtest a target-position strategy against the L1 book	Backtest a target-position strategy against the L1 book, taking each quote update as a market fill to reach the target immediately.
Backtest a market maker against L1 quotes and trades	Backtest a two-sided market maker against top-of-book quotes with a trade tape driving the fills, into a costed equity curve.
Backtest a two-sided order poster on OHLC bars	Backtest a two-sided order poster on OHLC bars, posting resting bids and asks that fill when the bar's range reaches them.
Backtest a target-position strategy on OHLC bars	Backtest a target-position strategy on OHLC bars, executing market orders at the next bar's open (causal, no manual lag).
Backtest a target position on a value series	Backtest a target position against a value series (price/mark) into a costed mark-to-market equity curve.
Running report columns for a backtest	Turn a backtest engine's [equity, pnl, position, cost] into running drawdown, cost, turnover, trades, and Sharpe.
Backtest a two-sided order poster against the trade tape	Backtest a two-sided order poster against the trade tape, filling resting quotes when prints cross them, into a costed equity curve.
Backtest a target-position strategy against the trade tape	Backtest a target-position strategy against the trade tape, taking each print as a market fill to reach the target immediately.
Choosing a backtest engine	The 5x2 engine grid by data model and order definition, with fill rules and encoding.