Momentum Pullback System

Systematic long-only swing trading strategy for S&P 500 stocks. Buys top momentum stocks on RSI(2) mean reversion pullbacks with ATR-based risk management.

Setup

# Create virtual environment
python3 -m venv .venv
source .venv/bin/activate

# Install dependencies
pip install -r requirements.txt

Download Data

Historical data must be downloaded before running backtests. Data is cached locally as Parquet files in data/cache/.

python scripts/download_data.py

This downloads daily OHLCV data for all S&P 500 constituents and SPY from Yahoo Finance. The initial download takes several minutes.

Run Backtest

python scripts/run_backtest.py

Options:

• --start YYYY-MM-DD — Backtest start date (default: 2021-01-01)
• --end YYYY-MM-DD — Backtest end date (default: 2025-12-31)
• --rsi-threshold N — Override RSI entry threshold
• --universe sp500|combined — Stock universe (default: sp500)

Results are printed to the console and an HTML report is saved to reports/backtest_report.html.

Run Tests

pytest tests/

Strategy Summary

1. Regime Filter — Only trade when SPY is above SMA-200 and SMA-50 > SMA-200
2. Universe Filter — Liquid stocks above $10, volume > 500k, price > SMA-200
3. Momentum Ranking — Composite relative strength vs SPY (1m/3m/6m), top 25 watchlist
4. Entry — RSI(2) drops below threshold while stock is in uptrend (close > SMA-200, close < SMA-5). Default threshold: 10; configurable per-ticker via RSI_ENTRY_OVERRIDES in config.
5. Exit — RSI(2) recovers above 75, ATR-based stop loss, or 5-day time stop

All parameters are configured in momentum_pullback_system/config.py.