ml4t-backtest 0.1.0a5

State-of-the-art event-driven backtesting engine for quantitative trading

ml4t-backtest

Event-driven backtesting engine for ML4T quantitative trading strategies.

Features

• Event-Driven Architecture: Point-in-time correctness with no look-ahead bias
• Exit-First Processing: Matches real broker order execution behavior
• VectorBT Validation: Results validated against VectorBT Pro
• Account Policies: Cash and margin account support
• Minimal Core: ~2,800 lines of focused, maintainable code
• 100k+ events/sec: High-performance event processing

Installation

pip install ml4t-backtest

Quick Start

from ml4t.backtest import Engine, Strategy, BacktestConfig

class SimpleMovingAverage(Strategy):
    def __init__(self, fast=10, slow=30):
        self.fast = fast
        self.slow = slow
        
    def on_bar(self, bar):
        fast_ma = bar.close_ma(self.fast)
        slow_ma = bar.close_ma(self.slow)
        
        if fast_ma > slow_ma and self.position == 0:
            self.buy(size=100)
        elif fast_ma < slow_ma and self.position > 0:
            self.close()

config = BacktestConfig(
    initial_cash=100_000,
    commission=0.001,
)

engine = Engine(data, SimpleMovingAverage(), config)
result = engine.run()

print(f"Total Return: {result.total_return:.2%}")
print(f"Sharpe Ratio: {result.metrics['sharpe']:.2f}")

Documentation

• AGENT.md: Comprehensive API reference for agents and developers
• api.yaml: Machine-readable API specification

Part of ML4T

This library is part of the ML4T quantitative trading toolkit:

• ml4t-data: Market data acquisition and storage
• ml4t-engineer: Feature engineering and indicators
• ml4t-diagnostic: Statistical validation and evaluation
• ml4t-backtest: Event-driven backtesting (this library)
• ml4t-live: Live trading platform

License

MIT License - see LICENSE for details.