Vectorised Backtester

This backtester was developed during Inter IIT Tech Meet 14.0 to get quick performance estimates for algorithmic strategies.

Unlike the provided tick-by-tick accurate backtester (which turned out to be slow for our statistic-heavy strategy), this implementation focuses on faster evaluation using vectorised computations and day-wise processing.

Backtesting was originally implemented around evaluating strategy-specific behavior and summary metrics. We also added a Plotly-based visual report that plots strategy performance along with primary and secondary features from the provided feature set.

What It Does

• Runs a fast, estimation-focused intraday backtest.
• Computes core outputs like PnL, NAV, drawdown, trade counts, returns, and risk metrics.
• Exports a text summary (backtest_results.txt).
• Generates an interactive Plotly HTML report (backtest_report.html) with:
  • Price vs portfolio curve
  • Price vs positions/signals
  • Primary and secondary feature overlays
  • Interval-wise bar analysis

How To Use

1. Prepare a CSV with at least these columns:
   • Time
   • Price
   • Signal
2. (Optional) Include any feature columns you want plotted (for example SMA_10, EAMA, HAMA_Slope, etc.).
3. Update the CSV path in backtester.py under the if __name__ == "__main__": block.
4. Install dependencies:

pip install numpy pandas plotly

5. Run:

python backtester.py

Generated outputs:

• backtest_results.txt
• backtest_report.html
• processed.csv