Pairs Trading Strategy – Indian Equities

Overview

This project implements a statistically driven pairs trading strategy for Indian equities using cointegration and mean reversion.

Methodology

1. Pair selection via correlation + Engle Granger cointegration
2. Structural filtering using half life of mean reversion
3. Z-score calibration on 2020–2024 daily data
4. Out of sample simulation on 2025 data
5. Capital aware execution with stop loss and cooldown rules

Folder Structure

• data/raw: input datasets
• data/generated: reproducible outputs
• src: full research and execution pipeline

Pipeline Overview

1. Pair Selection (2020–2024)

   • Same sector filtering out of Nifty 100
   • Correlation screening
   • Engle Granger cointegration
   • Hedge ratio estimation

2. Structural Validation

   • Mean reversion half life estimation
   • Spread stability checks

3. z-Score Calibration

   • Rolling z-score using daily data
   • Entry / Exit threshold optimization
   • Drawdown-aware scoring

4. Trading Engine (2025 Simulation)

   • Daily execution
   • Symmetric entry / exit / stop logic
   • Integer share sizing (Indian market compliant)
   • INR PnL accounting
   • Cooldown after stop loss

5. Interactive Dashboard

   • Capital adjustable PnL
   • Pair wise trade breakdown
   • Cumulative performance visualization

How to Run

python src/Ppair_selector.py
python src/Pair_analysis 20-24.py
python src/z_backtest.py
python src/Engine_1d.py
streamlit run src/dashboard.py