Backtesting

Running a backtest

matchstick run --strategy src/strategy.json --data data/sample_data.csv --capital 100000

Flags

Flag	Required	Default	Description
--strategy	Yes	-	Path to strategy file
--data	No	generated	Path to CSV market data
--mode	No	backtest	Execution mode
--capital	No	100000	Initial capital ($)

Without data

If you omit --data, the engine generates 1440 bars of synthetic data (30 days of 1-minute bars) using deterministic sine/cosine price movement.

Strategy logic

v0.1.0 uses a hardcoded mean reversion strategy:

1. Calculate SMA(20) on the most recent 20 bars
2. Entry: Buy when price drops 1% below SMA
3. Exit: Sell when price returns to SMA
4. Position size determined by FixedRisk method (1% risk per trade)
5. Risk manager validates each trade against limits

Custom strategy loading is planned for a future release.

Data format

CSV with header row:

timestamp,symbol,open,high,low,close,volume
2024-01-02 09:30:00,AAPL,185.50,186.20,185.10,185.80,1500000

Supported timestamp formats:

• 2024-01-02T09:30:00Z (ISO 8601)
• 2024-01-02 09:30:00
• 2024-01-02

Output metrics

Metric	Description
Final P&L	Absolute dollar profit/loss
Return	Percentage return on initial capital
Total Trades	Number of completed round trips
Win Rate	Percentage of profitable trades
Sharpe Ratio	Annualized risk-adjusted return (252 trading days)
Max Drawdown	Largest peak-to-trough decline

Components used

The backtest engine uses all 5 library crates:

• core-types: OhlcvBar, Position, Timeframe
• indicators: IndicatorEngine for SMA calculation
• orders: OrderManager (initialized, used for tracking)
• risk: RiskManager for position sizing and trade validation
• strategy: StrategyManager for strategy lifecycle