vectorbt Library

Lightning-fast vectorized backtesting framework. 100x faster than event-driven frameworks using NumPy vectorization and parallel processing.

Difficulty: Advanced

Category: Backtesting

Installation

$ pip install vectorbt

Python

import vectorbt as vbt

import yfinance as yf

# Download data

df = yf.download("BTCUSD", period="2y")

# Define fast and slow windows

fast_ma = vbt.MA.run(df['Close'], 10)

slow_ma = vbt.MA.run(df['Close'], 50)

# Generate entry and exit signals

entries = fast_ma.ma_crossed_above(slow_ma)

exits = fast_ma.ma_crossed_below(slow_ma)

# Run portfolio simulation

pf = vbt.Portfolio.from_signals(df['Close'], entries, exits, init_cash=10000)

# Print results

print(pf.stats())

print(f"Total Return: {pf.total_return():.2%}")

print(f"Sharpe Ratio: {pf.sharpe_ratio():.2f}")

print(f"Max Drawdown: {pf.max_drawdown():.2%}")

Python

import vectorbt as vbt

import numpy as np

# Download data

price = vbt.YFData.download("EURUSD=X", period="2y").get('Close')

# Define parameter ranges

fast_windows = np.arange(10, 50, 5)

slow_windows = np.arange(50, 200, 10)

# Run optimization

fast_ma, slow_ma = vbt.MA.run_combs(price, window=[fast_windows, slow_windows], r=2, short_names=['fast', 'slow'])

entries = fast_ma.ma_crossed_above(slow_ma)

exits = fast_ma.ma_crossed_below(slow_ma)

pf = vbt.Portfolio.from_signals(price, entries, exits)

# Find best parameters

print(pf.total_return().idxmax())

print(f"Best Return: {pf.total_return().max():.2%}")