Backtesting Engine
At runtime the EventBus orchestrates the data flow between:
1. Data sources → provide market data.
2. Strategy → consumes data & produces orders.
3. Execution simulator → turns orders into fills.
The BacktestRunner wires these pieces and produces a BacktestResult.
from quantex.backtest import BacktestRunner
from quantex.sources import ParquetDataSource
from my_strategy import SMACross
ds = ParquetDataSource("btc.parquet", symbol="BTC-USD")
strategy = SMACross(symbols=["BTC-USD"], initial_cash=10_000)
result = BacktestRunner(strategy, {"btc": ds}).run()
print(result.metrics)
BacktestResult Fields
| Field | Type | Description |
|---|---|---|
nav |
pd.Series |
Net-asset-value over time. |
orders |
list[Order] |
All generated orders. |
fills |
list[Fill] |
Executed trades. |
metrics |
dict |
Performance metrics (see below). |
Performance Metrics
The metrics dictionary contains key performance statistics:
| Metric | Description |
|---|---|
total_return |
Total return as a percentage (e.g., 0.15 = 15% return) |
max_drawdown |
Maximum drawdown as a percentage (e.g., -0.25 = 25% drawdown) |
sharpe_ratio |
Annualized Sharpe ratio |
# Example accessing metrics
result = BacktestRunner(strategy, data_sources).run()
print(f"Total Return: {result.metrics['total_return']:.2%}")
print(f"Max Drawdown: {result.metrics['max_drawdown']:.2%}")
print(f"Sharpe Ratio: {result.metrics['sharpe_ratio']:.4f}")
Extend by post-processing the
navorfillslists.