MACD + SMA Backtesting (2015–2024)

🔍 Project Overview

This project applies systematic backtesting to evaluate a well-known technical trading rule: the MACD crossover strategy, enriched with simple moving averages (SMA 10/20) for trend context.

Backtesting means simulating how a portfolio would have evolved if a given strategy had been applied to historical price data.
It allows us to test:

• when trades would have been triggered,
• how capital would have grown or declined,
• and what risk/return characteristics the strategy generates.

📉 The Trading Strategy

The core of the system is the MACD crossover rule:

• MACD line = EMA(12) – EMA(26)
• Signal line = EMA(9) of MACD
• Trading rules:
  • Buy when MACD crosses above the signal line.
  • Sell when MACD crosses below the signal line.

Two versions were tested:

1. Long-only mode → stay in cash when a sell signal appears.
2. Long & short mode → reverse position (switch between long and short) on every crossover.

In addition, SMA(10) and SMA(20) are plotted directly on the price chart to provide extra context on short-term trend direction.

🧰 Methodology

• Assets tested:
  Tesla (TSLA), Nvidia (NVDA), Apple (AAPL), Microsoft (MSFT), Amazon (AMZN)
• Period covered: 2015-01-01 → 2024-12-31
• Data source: Yahoo Finance (yfinance)
• Indicators used: MACD, Signal line, SMA(10), SMA(20)
• Tools:
  • backtesting.py for simulation,
  • pandas, numpy for data manipulation,
  • TA-Lib (when available, otherwise custom NumPy EMAs),
  • bokeh for interactive visualization.

Technical improvements implemented:

• Batch downloading of multiple tickers (reduces Yahoo rate-limit issues).
• Local CSV caching (avoids unnecessary repeated downloads).
• Retry logic with exponential backoff.
• Automatic generation of:
  • interactive HTML reports (equity curve, OHLC, trades, indicators),
  • CSV logs with entry/exit details (time, size, PnL, returns).

📊 Example Visualization

Legend:

• OHLC candles = price movements
• SMA(10) & SMA(20) = short-term moving averages
• Red/green arrows = trades (entries/exits)
• Lower panels = MACD line & signal line

📈 Key Results

Each run produces a performance summary with:

• Final return & CAGR (annualized growth rate)
• Annualized volatility
• Sharpe, Sortino, and Calmar ratios
• Maximum drawdown (loss from peak to trough)
• Win rate, average trade return, and profit factor

Example (TSLA 2015–2024, long+short):

• CAGR ≈ X%
• Sharpe ratio ≈ Y
• Max drawdown ≈ Z%
• Profit factor ≈ ~2.0

(exact values vary depending on execution and commission settings)

🧠 Why It Matters

This project demonstrates how a technical trading rule can be implemented and tested in Python from scratch.

It shows:

• how to integrate market data acquisition,
• how to design and code a strategy class,
• how to produce risk/performance metrics,
• and how to visualize both equity curves and indicator dynamics.

Even if the strategy itself is relatively simple, the exercise builds a reusable framework for testing more advanced trading rules, combinations of indicators, or risk management overlays.

⚠️ Limitations & Next Steps

No backtest is perfect. This study comes with important caveats:

• Transaction costs & slippage: only commissions were modeled; real execution may differ.
• Sideways markets: MACD crossovers can generate many false signals when prices move sideways.
• Overfitting risk: parameters (12, 26, 9) are conventional, but optimizing them excessively risks fitting noise.
• Survivorship bias: only current large-cap stocks were tested; results may not generalize to delisted or smaller firms.
• No risk overlays: position sizing, stop-loss rules, and portfolio constraints were not included.

Next steps could involve:

• Out-of-sample testing (split data into training/test).
• Running the strategy on intraday data.
• Adding volatility filters (e.g., ATR-based position sizing).
• Combining MACD with other signals (RSI, volume indicators).
• Portfolio-level backtesting (dynamic allocation between tickers).

📚 Suggested Readings

For readers interested in technical trading rules and their evaluation:

• Brock, W., Lakonishok, J., & LeBaron, B. (1992). Simple technical trading rules and the stochastic properties of stock returns. The Journal of Finance, 47(5), 1731-1764.
• Park, C. H., & Irwin, S. H. (2007). What do we know about the profitability of technical analysis? Journal of Economic Surveys, 21(4), 786-826.
• Murphy, J. (1999). Technical Analysis of the Financial Markets. (classic practitioner’s reference).
• Prado, M. L. de (2018). Advances in Financial Machine Learning. (for modern extensions).

🏷️ Tags

#Backtesting #QuantFinance #Python #MACD #SMA #TechnicalAnalysis #TradingStrategy #DataScience