Lecture on Backtesting.py

Introduction to Backtesting.py

• Instructor: Chad
• A lightweight backtesting framework
• Comparison with other libraries like VectorBT or Backtrader:
  • Backtrader and VectorBT are comprehensive but overwhelming
  • Backtesting.py focuses only on essential backtesting
• Key features:
  • Simple documentation
  • Fast performance (event-driven backtester)
  • Built-in parameter optimizer
  • No support for multiple assets or fractional shares

Getting Started

Installation

• Create and activate a virtual environment:

  virtualenv venv
  source venv/bin/activate
  

• Install necessary packages:

  pip install backtesting
  pip install talib (optional)
  

Writing the First Strategy

• Create a new Python file example.py
• Import the essentials from backtesting.py and other necessary libraries:

  from backtesting import Backtest, Strategy
  from backtesting.test import GOOG
  import talib
  

• Define a simple RSI-based strategy:

  class RSI(Strategy):
      def init(self):
          self.rsi = self.I(talib.RSI, self.data.Close, 14)
      def next(self):
          if self.rsi[-1] < 30:
              self.buy()
          elif self.rsi[-1] > 70:
              self.sell()
  

• Running the backtest:

  bt = Backtest(GOOG, RSI, cash=10_000)
  stats = bt.run()
  bt.plot()
  

Strategy Customization

Adding Custom Parameters

• Adding parameters for optimization:

  class RSI(Strategy):
      upper_bound = 70
      lower_bound = 30
      rsi_window = 14
      def init(self):
          self.rsi = self.I(talib.RSI, self.data.Close, self.rsi_window)
      def next(self):
          if self.rsi[-1] < self.lower_bound:
              self.buy()
          elif self.rsi[-1] > self.upper_bound:
              self.sell()
  

Optimization with optimize

• Running an optimizer with specified ranges:

  stats, heatmap = bt.optimize(upper_bound=range(50, 85, 5), lower_bound=range(10, 45, 5), rsi_window=range(10, 30, 2), 
                               maximize='Sharpe Ratio', return_heatmap=True)
  

• Plotting heatmaps:

  import seaborn as sns
  import matplotlib.pyplot as plt
  hm = heatmap.groupby(['upper_bound', 'lower_bound']).mean().unstack()
  sns.heatmap(hm)
  plt.show()
  

Advanced Techniques

Multi-Time Frame Analysis

• Using weekly and daily RSI:

  from backtesting.lib import resample_apply
  class MultiFrameRSI(Strategy):
      def init(self):
          self.daily_rsi = self.I(talib.RSI, self.data.Close, 14)
          self.weekly_rsi = resample_apply('W-FRI', talib.RSI, self.data.Close, 14)
      def next(self):
          if self.daily_rsi[-1] < 30 and self.weekly_rsi[-1] < 30:
              self.buy()
          elif self.daily_rsi[-1] > 70 and self.weekly_rsi[-1] > 70:
              self.sell()
  

Short Selling and Advanced Orders

• Implementing short selling:

  def next(self):
      if self.daily_rsi[-1] > 70:
          if self.position.is_long:
              self.position.close()
          self.sell()
      elif self.daily_rsi[-1] < 30:
          if self.position.is_short:
              self.position.close()
          self.buy()
  

• Implementing stop loss and take profit:

  def next(self):
      price = self.data.Close[-1]
      sl = price * 0.95
      tp = price * 1.15
      if self.daily_rsi[-1] < 30:
          self.buy(sl=sl, tp=tp)
      elif self.daily_rsi[-1] > 70:
          self.sell()
  

Position Sizing

• Using different position sizes:

  def next(self):
      if self.daily_rsi[-1] < 30:
          self.buy(size=0.1)
      elif self.daily_rsi[-1] > 70:
          self.sell()
  

• Alternative method specifying exact number of shares:

  def next(self):
      if self.daily_rsi[-1] < 30:
          self.buy(size=1)
      elif self.daily_rsi[-1] > 70:
          self.sell()
  

Extracting Trades Data

• Extract and analyze trade details:

  trades = stats['_trades']
  print(trades.to_string())
  

Summary

• Backtesting.py is suitable for simple to moderately complex strategies.
• It’s efficient and user-friendly but lacks support for multiple assets and fractional shares.

Use the documentation: With its small size and simplicity, you should be able to master it quickly.

• Incorporate optimizations, multi-time frame analysis, and advanced trade handling into your strategies.
• Happy backtesting!