Think in matrices. Backtest at scale.
VectorBT takes a radically different approach to backtesting: instead of looping through bars one strategy at a time, it packs thousands of configurations into NumPy arrays, accelerates the hot path with Numba and Rust, and runs them all at once, turning hours of grid search into seconds.
Explore thousands of trading ideas across assets and timeframes, analyze portfolio performance down to individual trades, and visualize results interactively, all in a few lines of code. Built for both human researchers and AI agents, VectorBT combines large-scale experimentation with a mature, battle-tested backtesting stack refined through years of community use.
VectorBT is the open-source community edition of VectorBT PRO. PRO extends the library with parallelization, additional data integrations, portfolio optimization, pattern recognition, event projections, limit orders, leverage, and over 100 other features, with new features added weekly.
See VectorBT vs PRO or browse PRO feature examples.
Features
- 🧮 Fast, vectorized backtesting and strategy research built on pandas, NumPy, and Numba
- 🦾 Optional Rust engine for precompiled speed without JIT overhead
- 🐼 Pandas-native API with custom accessors and high-performance operations
- 🪗 Flexible broadcasting for multi-asset analysis and large-scale parameter sweeps
- 🎛️ Rich indicator ecosystem with custom indicators and integrations for TA-Lib, Pandas TA, and more
- 🧾 Portfolio backtesting with trade, drawdown, and performance analytics, including QuantStats integration
- 🚦 Signal tooling for generation, ranking, mapping, and distribution analysis
- 🚰 Built-in data access with preprocessing and synthetic data generation
- 🪨 Robustness testing with walk-forward optimization and label generation for ML workflows
- 🪞 Interactive visualization with Plotly, Jupyter widgets, and browser-friendly dashboards
- 🕰️ Automation tools for scheduled updates and Telegram notifications
- 🧱 Composable Python API for rapid experimentation and AI agent-driven workflows
Installation
pip install -U vectorbt
To install the optional Rust engine:
pip install -U "vectorbt[rust]"To install all optional integrations (TA-Lib, Pandas TA, etc.):
pip install -U "vectorbt[full]"To install all optional integrations together with the Rust engine:
pip install -U "vectorbt[full,rust]"Examples
Invest $100 in Bitcoin since 2014
import vectorbt as vbt data = vbt.YFData.download("BTC-USD") price = data.get("Close") pf = vbt.Portfolio.from_holding(price, init_cash=100) print(pf.total_profit())
19501.10906763755
Trade a dual-SMA crossover strategy
fast_ma = vbt.MA.run(price, 10) slow_ma = vbt.MA.run(price, 50) entries = fast_ma.ma_crossed_above(slow_ma) exits = fast_ma.ma_crossed_below(slow_ma) pf = vbt.Portfolio.from_signals(price, entries, exits, init_cash=100) print(pf.total_profit())
34417.80960086067
Generate 1,000 random strategies
import numpy as np symbols = ["BTC-USD", "ETH-USD"] data = vbt.YFData.download(symbols, missing_index="drop") price = data.get("Close") n = np.random.randint(10, 101, size=1000).tolist() pf = vbt.Portfolio.from_random_signals(price, n=n, init_cash=100, seed=42) mean_expectancy = pf.trades.expectancy().groupby(["randnx_n", "symbol"]).mean() fig = mean_expectancy.unstack().vbt.scatterplot(xaxis_title="randnx_n", yaxis_title="mean_expectancy") fig.show()
Test 10,000 dual-SMA window combinations
symbols = ["BTC-USD", "ETH-USD", "XRP-USD"] data = vbt.YFData.download(symbols, missing_index="drop") price = data.get("Close") windows = np.arange(2, 101) fast_ma, slow_ma = vbt.MA.run_combs(price, window=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, size=np.inf, fees=0.001, freq="1D") fig = pf.total_return().vbt.heatmap( x_level="fast_window", y_level="slow_window", slider_level="symbol", symmetric=True, trace_kwargs=dict(colorbar=dict(title="Total return", tickformat="%"))) fig.show()
Tip
VectorBT PRO
Use random search to sample a large parameter grid, or conditional parameters to exclude combinations such as a fast MA window longer than the slow MA window. The cross-validation tutorial shows how to select parameters on training data and evaluate them on later periods.
Inspect any strategy configuration
print(pf[(10, 20, "ETH-USD")].stats())
Start 2017-11-09 00:00:00+00:00
End 2026-01-03 00:00:00+00:00
Period 2978 days 00:00:00
Start Value 100.0
End Value 1604.093789
Total Return [%] 1504.093789
Benchmark Return [%] 866.094127
Max Gross Exposure [%] 100.0
Total Fees Paid 204.226289
Max Drawdown [%] 70.734951
Max Drawdown Duration 1095 days 00:00:00
Total Trades 81
Total Closed Trades 80
Total Open Trades 1
Open Trade PnL -14.232533
Win Rate [%] 41.25
Best Trade [%] 120.511071
Worst Trade [%] -27.772271
Avg Winning Trade [%] 27.265519
Avg Losing Trade [%] -9.022864
Avg Winning Trade Duration 32 days 20:21:49.090909091
Avg Losing Trade Duration 8 days 16:51:03.829787234
Profit Factor 1.275515
Expectancy 18.979079
Sharpe Ratio 0.861945
Calmar Ratio 0.572758
Omega Ratio 1.20277
Sortino Ratio 1.301377
Name: (10, 20, ETH-USD), dtype: object
Plot any strategy configuration
pf[(10, 20, "ETH-USD")].plot().show()
Animate Bollinger Bands across multiple symbols
VectorBT goes beyond backtesting, with tools for financial data analysis and visualization:
symbols = ["BTC-USD", "ETH-USD", "XRP-USD"] data = vbt.YFData.download(symbols, period="6mo", missing_index="drop") price = data.get("Close") bbands = vbt.BBANDS.run(price) def plot(index, bbands): bbands = bbands.loc[index] fig = vbt.make_subplots( rows=2, cols=1, shared_xaxes=True, vertical_spacing=0.15, subplot_titles=("%B", "Bandwidth")) fig.update_layout(showlegend=False, width=750, height=400) bbands.percent_b.vbt.ts_heatmap( trace_kwargs=dict(zmin=0, zmid=0.5, zmax=1, colorscale="Spectral", colorbar=dict( y=(fig.layout.yaxis.domain[0] + fig.layout.yaxis.domain[1]) / 2, len=0.5 )), add_trace_kwargs=dict(row=1, col=1), fig=fig) bbands.bandwidth.vbt.ts_heatmap( trace_kwargs=dict(colorbar=dict( y=(fig.layout.yaxis2.domain[0] + fig.layout.yaxis2.domain[1]) / 2, len=0.5 )), add_trace_kwargs=dict(row=2, col=1), fig=fig) return fig vbt.save_animation("bbands.gif", bbands.wrapper.index, plot, bbands, delta=90, step=3, fps=3)
100%|██████████| 31/31 [00:21<00:00, 1.21it/s]
Visit the website for more examples, documentation, and guides.
Example apps
Explore candlestick patterns interactively and backtest their signals with VectorBT.
Links
License
This work is fair-code distributed under the Apache 2.0 with Commons Clause license.
The source code is publicly available, and everyone (individuals and organizations) may use it for free. However, you may not sell products or services that are primarily this software.
If you have questions or want to request a license exception, please contact the author.
Installing optional dependencies may be subject to a more restrictive license.
Star History
Disclaimer
This software is for educational purposes only. Do not risk money you cannot afford to lose.
Use the software at your own risk. The authors and affiliates assume no responsibility for your trading results.