Stock Parfait

Statistical Experiments with Stock Market Data

Welcome to Stock Parfait, an open source project for statistical experiments with stock market data.

The purpose of this project is to perform and document a series of statistical experiments with real financial data, at the moment primarily over the US stock market daily price data from 1998 to mid-2022. The text, code and necessary configurations are published on GitHub, so the experiments can be repeated by anyone with access to the data and some basic understanding of how to compile and run a command-line app.

We purposefully do not follow any specific textbook, any established trading or investment approaches, indicators, quantitative trading strategies, or econometric theories. Instead, we start from the ground up with the actual historical data, and see what the data can tell us.

Our experiments follow a few basic principles derived from the scientific method:

  • Clear, intuitive definitions,

  • Leave no assumptions unchallenged,

  • Verify with real data,

  • KISS: keep it simple, stupid.


GitHub