OpenBacktesting. GitHub
v0.2.0Open source. Open possibilities.

Good strategies
start with evidence.

An open-source Python framework for backtesting,
optimization, and strategy research.

Start building Explore a backtest
Python 3.12–3.14 Run locally. Explore freely.
FROM THE FRAMEWORK, TO THE PAGEA real backtest. Yours to inspect. ↙

SMA crossover

GOOG / Daily bars / 2004–2013

Run complete
Strategy equityBuy & hold*USD
$10,000 initial cash0.2% commission / execution95% order sizingLong & short

Precomputed with OpenBacktesting 0.2.0; switch periods to compare two recorded runs. Positions closed on the final bar. Chart rendered by this website. *Buy & hold is normalized closing price, before costs. Historical results do not predict future performance.

At home in your Python workflow.
PythonNumPypandas
Familiar tools. Inspectable results.

THE FRAMEWORK

Follow the idea.
Question the result.

From the first rule to the next experiment,
keep your research in your own hands.

Your logic, in plain Python.

Define strategies bar by bar. Work with built-in indicators, explicit position sizing, and execution costs on a single instrument.

Write a strategy
init()→next()→run()

A parameter is a question.

Explore a grid of possibilities or use optional SAMBO optimization. Compare candidates and save experiments for later analysis.

Explore optimization
fast51020slow2040

Look past the headline.

Inspect orders, trades, equity, and risk. Take pandas results and Portfolio reports into your own notebooks and charting tools.

Understand the results
equitytradesrisk

Purpose-built for historical, single-asset research.

Know the model’s boundaries

AN IDEA, EXPRESSED IN CODE

Small API.
Room to explore.

Two moving averages. A crossover.
A simple place to start asking better questions.

Define your rules, compare parameters, then inspect the full history behind each result. The example uses bundled GOOG data, so you can get started without a data download.

Walk through your first backtest
YOU BRING THE HYPOTHESIS.

OpenBacktesting handles the historical execution, accounting, and results.

sma_cross.py
from openbacktesting import Backtest, Strategyfrom openbacktesting.test import GOOGclass SmaCross(Strategy):    fast = 10    slow = 20    def init(self):        self.fast_sma = self.ta.sma(self.close, length=self.fast)        self.slow_sma = self.ta.sma(self.close, length=self.slow)    def next(self):        if self.ta.is_crossover(self.fast_sma, self.slow_sma):            self.position.close()            self.buy(percent=95)        elif self.ta.is_crossunder(self.fast_sma, self.slow_sma):            self.position.close()            self.sell(percent=95)if __name__ == "__main__":    bt = Backtest(GOOG, SmaCross, cash=10_000, commission=0.002,                  end_policy="close_at_last_close")    result = bt.run()    print(result[["Return [%]", "# Trades", "Max. Drawdown [%]"]])
Python 3.12–3.14Built-in indicators. Plain Python.

YOUR NEXT EXPERIMENT STARTS HERE

Less setup.
More discovery.

Clone the source. Run your first strategy.
Let the evidence lead from there.

Read the documentation 
git clone https://github.com/lucidcog/openbacktesting.gitcd openbacktestinguv sync --lockeduv run --locked python examples/advanced/native_run.py

Source installation · Python 3.12–3.14 · Install uv ↗

No market-data download needed for the bundled example.Browse more examples

A FEW THINGS TO KNOW

Before the
first backtest.

Explore the documentation
What is OpenBacktesting?

OpenBacktesting is an open-source Python framework for historical backtesting, parameter optimization, and strategy research. Write strategy rules, simulate their execution on historical market data, and inspect trades, equity, and risk statistics.

Can I use my own market data?

Yes. Pass a pandas DataFrame with Open, High, Low, and Close columns. Volume is optional unless an indicator needs it. Each backtest models one asset in a single settlement currency. See the market data guide for requirements.

How do I install the Python framework?

Use Python 3.12–3.14, clone the GitHub repository, and run uv sync --locked. You can also install the checked-out source with python -m pip install .. Add the optimize extra for parameter searches. The commands above run a complete bundled example.

Does OpenBacktesting include live trading or charts?

The framework focuses on historical research. It exports results and plotting data for your own renderer; it does not include a chart renderer or live-trading engine. The interactive chart on this website displays precomputed framework results. Read the current scope and limitations.