A lightweight backtesting framework based on vectorbt focused on statistical robustness, modularity, and seamless strategy integration with custom-implemented models and crypto focused data-loader.
- Simple integration with vectorbt as the backtesting engine (
bt_instance). - Custom model support: native wrappers for Hawkes processes, Kalmanfilter, and other statistical frameworks.
- Built-in data loaders for cryptocurrencies (e.g., Bitcoin, Ethereum).
- Modular architecture: define strategies by inheriting from a base
Strategyclass (preprocess,generate_signals,param_space). - Robust validation: out-of-sample splits, walk-forward optimization, and hyperparameter tuning via
hyperopt. - Statistical analysis tools: Monte Carlo simulations, bootstrapping of trade outcomes, and sensitivity analysis.
- Performance reporting: generate equity curves, heatmaps, and metric summaries with minimal boilerplate.
Install the package via pip:
pip install quantybt