Backtesting

Test before you trade

Backtesting lets you run any of the built-in strategies over historical market data before you risk capital. It is a real backtesting engine in the Rust quant core — not a rough approximation — so the same logic that produces live signals is the logic being measured.

A run replays historical candles through the strategy, applying its ATR(14, 2x) stop and 2:1 target on every simulated trade, and records the outcome.

Streaming progress

Long runs report streaming progress over Server-Sent Events (SSE). Instead of waiting for a single result at the end, the dashboard updates as the backtest advances through the history window, so you can watch it work and cancel early if needed.

Metrics

Each backtest returns a performance summary, including:

• PnL — net profit or loss over the tested period.
• Max drawdown — the largest peak-to-trough decline in equity.
• Win rate — the share of trades that closed profitably.
• Trades — the number of positions opened and closed.

Use these together — a high win rate with deep drawdown, or strong PnL from very few trades, tells you more than any single number.

Parameter optimization

Beyond a single run, Aegis can optimize a strategy's parameters by sweeping across values and comparing the resulting metrics, helping you find robust settings rather than a single lucky configuration.

Planned

Walk-forward analysis and Monte Carlo simulation are In development. They will add out-of-sample validation and randomized stress-testing on top of the current single-window backtest and optimization.

Once a strategy performs the way you expect, promote it through the trading modes described in Risk Management — PAPER, then TESTNET, then REAL.