
Strategy Backtester
- Python
- Streamlit
- pandas
- pytest
The successor to the earlier swing-trading scanner, rebuilt around a single question: did the strategy actually beat doing nothing? Strategies run against locally cached historical price data and report through a Streamlit dashboard, currently covering Buy & Hold and SMA Crossover.
Every result is benchmarked against buy-and-hold by default, which is the whole point. A strategy returning eight percent sounds like a success until you find the index returned eleven over the same window, and making that comparison automatic rather than optional guards against the standard failure mode of backtesting, which is quietly fooling yourself. Prices are cached locally so runs stay reproducible and do not depend on an API responding identically twice.
Built to test ideas honestly rather than to manage money — there is deliberately no broker integration.