QuantsPlaybook is a collection of quantitative trading strategies and research implementations focused on reproducing financial engineering reports from major Chinese securities firms.
The project addresses the challenge of understanding and validating quantitative investment strategies by reproducing research from leading brokerages. It provides complete implementations of over 100 strategies spanning market timing, factor construction, value investing, and portfolio optimization. The approach combines traditional technical analysis with modern machine learning techniques, including deep learning models like Transformers and LSTMs, signal processing methods such as Hilbert-Huang Transform and empirical mode decomposition, and gradient boosting algorithms. Each strategy includes full code from data acquisition through backtesting, with visualizations and performance analysis reports based on real A-share market data.
The project suits researchers and practitioners working with Chinese equities who want to study and validate published quantitative strategies. It is particularly valuable for those interested in learning how major brokerages approach quantitative investing, as it provides access to implementations of strategies from firms like Everbright, Huatai, CMB, and Guosen. The collection spans diverse methodologies including timing models like RSRS and QRS, factor-based approaches with custom factors like chip distribution and prominence factors, and portfolio optimization techniques. The project depends on JQData and Tushare for market data access and uses standard tools including Pandas, NumPy, Qlib, and Backtrader for implementation and backtesting.
The repository shows active development with recent updates to its documentation and strategy collection. The project maintains organized folder structures with associated research papers and corresponding Python implementations for each strategy. The codebase integrates multiple technical stacks including PyTorch and TensorFlow for deep learning, LightGBM and XGBoost for gradient boosting, and Matplotlib, Seaborn, and Plotly for visualization across Jupyter Notebook environments.
