Backtest Simple strategy on Spy ETF in phyton code that outperform simple buy and hold…
- User
Backtest Simple strategy on Spy ETF in phyton code that outperform simple buy and hold strategy om metrics like sharp.
- Assistant
Certainly! Below is an example of a simple moving average crossover strategy implemented in Python that, when backtested on the SPY ETF, can outperform a simple buy-and-hold strategy in terms of the Sharpe ratio.
This strategy involves buying when the short-term moving average crosses above the long-term moving average (indicating an upward trend) and selling when it crosses below (indicating a downward trend).
import pandas as pd import numpy as np import yfinance as yf import matplotlib.pyplot as plt # Fetch historical data for SPY symbol = 'SPY' start_date = '2010-01-01' end_date = '2023-09-30' df = yf.download(symbol, start=start_date, end=end_date) # Calculate moving averages df['Short_MA'] = df['Adj Close'].rolling(window=50).mean() df['Long_MA'] = df['Adj Close'].rolling(window=200).mean() # Generate signals df['Position'] = 0 df['Position'][50:] = np.where(df['Short_MA'][50:] > df['Long_MA'][50:], 1, -1) # Long or Short # Calculate daily returns df['Market_Return'] = df['Adj Close'].pct_change() df['Strategy_Return'] = df['Position'].shift(1) * df['Market_Return'] # Drop NaN values df.dropna(inplace=True) # Calculate cumulative returns df['Cumulative_Market_Return'] = (1 + df['Market_Return']).cumprod() df['Cumulative_Strategy_Return'] = (1 + df['Strategy_Return']).cumprod() # Plot the results plt.figure(figsize=(12,6)) plt.plot(df['Cumulative_Market_Return'], label='Buy and Hold') plt.plot(df['Cumulative_Strategy_Return'], label='Moving Average Strategy') plt.legend() plt.title('Strategy vs. Buy and Hold Performance') plt.xlabel('Date') plt.ylabel('Cumulative Returns') plt.show() # Calculate Sharpe Ratios risk_free_rate = 0.01 # 1% annual risk-free rate trading_days = 252 # Number of trading days in a year # Function to calculate annualized Sharpe Ratio def calculate_sharpe_ratio(returns, risk_free_rate): excess_returns = returns - risk_free_rate / trading_days annualized_return = excess_returns.mean() * trading_days annualized_volatility = returns.std() * np.sqrt(trading_days) sharpe_ratio = annualized_return / annualized_volatility return sharpe_ratio strategy_sharpe = calculate_sharpe_ratio(df['Strategy_Return'], risk_free_rate) market_sharpe = calculate_sharpe_ratio(df['Market_Return'], risk_free_rate) print(f"Strategy Sharpe Ratio: {strategy_sharpe:.2f}") print(f"Buy and Hold Sharpe Ratio: {market_sharpe:.2f}")Explanation:
-
Data Fetching: We use
yfinanceto download historical adjusted closing prices for SPY from January 1, 2010, to September 30, 2023. -
Moving Averages: We calculate the 50-day (short-term) and 200-day (long-term) simple moving averages of the adjusted closing price.
-
Generating Signals:
- We initiate a position: long (1) when the short-term MA crosses above the long-term MA.
- We initiate a position: short (-1) when the short-term MA crosses below the long-term MA.
- Note: For simplicity, the strategy goes to a short position when not long, but you can modify it to go to cash (position 0) instead.
-
Calculating Returns:
- We calculate the daily market return (
Market_Return) and the strategy return (Strategy_Return), which depends on our position from the previous day.
- We calculate the daily market return (
-
Performance Metrics:
- We compute the cumulative returns for both the market and the strategy.
- We plot these cumulative returns to visualize performance over time.
- We calculate the annualized Sharpe Ratio for both the strategy and the buy-and-hold approach.
Results:
The output will display the Sharpe Ratios for both the strategy and the buy-and-hold method. Typically, moving average strategies can reduce volatility and potentially increase the Sharpe Ratio even if the total returns are similar or slightly lower than buy-and-hold.
Optimizing the Strategy:
You can optimize the moving average window sizes to further improve performance:
# Optimize Moving Average Window Sizes short_windows = [20, 50, 100] long_windows = [100, 200, 250] best_sharpe = -np.inf best_short_window = None best_long_window = None for short_win in short_windows: for long_win in long_windows: if short_win >= long_win: continue df['Short_MA'] = df['Adj Close'].rolling(window=short_win).mean() df['Long_MA'] = df['Adj Close'].rolling(window=long_win).mean() df['Position'][long_win:] = np.where(df['Short_MA'][long_win:] > df['Long_MA'][long_win:], 1, -1) df['Strategy_Return'] = df['Position'].shift(1) * df['Market_Return'] strategy_sharpe = calculate_sharpe_ratio(df['Strategy_Return'], risk_free_rate) if strategy_sharpe > best_sharpe: best_sharpe = strategy_sharpe best_short_window = short_win best_long_window = long_win print(f"Best Sharpe Ratio: {best_sharpe:.2f}") print(f"Optimal Short Window: {best_short_window}") print(f"Optimal Long Window: {best_long_window}")Important Notes:
-
Transaction Costs: This simple backtest does not account for transaction costs, which can affect the actual performance, especially if the strategy results in frequent trading.
-
Slippage and Market Impact: In real trading, large orders can affect the market price. This is not considered in the backtest.
-
Historical Data Limitations: Past performance is not indicative of future results. The strategy may not perform as well in the future.
-
Risk Management: Always incorporate risk management techniques in live trading to protect capital.
Dependencies:
Ensure you have the required libraries installed:
pip install yfinance pandas numpy matplotlibConclusion:
The moving average crossover strategy is one of the simplest trend-following strategies. By focusing on periods when the price shows upward momentum (short-term average above long-term average), the strategy can potentially achieve a higher Sharpe Ratio compared to a buy-and-hold approach due to improved risk-adjusted returns.
-