BacktestBay

BacktestBay is a Python-based framework for backtesting trading strategies. It leverages yfinance for importing financial data, which requires an active internet connection.

Getting Started

To get started, first clone the repository using the following command:

git clone repository-url

Next, create and activate the environment by navigating to the directory containing environment.yml and running:

mamba env create -f environment.yml
mamba activate backtest_bay

Once the environment is set up, initialize the project by typing the following in the project directory:

pytask

Configuration

BacktestBay enables the configuration of input parameters through config.py.

First, the data source is specified in more detail in the configuration file:

STOCKS = ["AAPL", "MSFT"]
START_DATE = "2019-01-01"
END_DATE = "2025-01-01"
INTERVAL = "1d"

Multiple valid symbols from yfinance can be added to the STOCKS list. START_DATE and END_DATE define the time range for the price history of the selected STOCKS at a specified INTERVAL. The available options for INTERVAL are listed at the end of the README file.

In the next step, multiple trading strategies can be added to the STRATEGIES list:

STRATEGIES = ["bollinger", "macd", "roc", "rsi"]

Each trading strategy is backtested on the stocks specified in the STOCKS list. The following trading strategies are currently implemented:

• 'bollinger': Bollinger Bands
• 'macd': Moving Average Convergence Divergence
• 'roc': Rate of Change
• 'rsi': Relative Strength Index

Finally, additional backtesting parameters can be specified

INITIAL_CASH = 1000000
TAC = 0.005
TRADE_PCT = 0.05

providing users with greater control and flexibility in their backtesting process. INITIAL_CASH specifies the amount of cash available at the start. TAC specifies the transaction costs as a percentage of the traded volume. TRADE_PCT indicates the percentage of the current portfolio allocated for trading.

The initialization of the project with pytask creates a bld/plot folder in the repository, where two analyses are generated for each stock in STOCKS:

• portfolio: Shows the development of the portfolio over time, including performance metrics.
• signals: Provides a detailed analysis of the signals in relation to the price trend.

Project Structure

The project structure in src/backtest_bay is as follows:

• download_data: Contains functions to download STOCKS.
• backtest: Develops functions to generate trading signals and backtest them.
• plot: Plots portfolio performance.

yfinance Download Intervals and Maximum History

INTERVAL	Description	Maximum History Available
1m	1 minute	7 days
2m	2 minutes	60 days
5m	5 minutes	60 days
15m	15 minutes	60 days
30m	30 minutes	60 days
60m	60 minutes	730 days (2 years)
90m	90 minutes	730 days (2 years)
1h	1 hour	730 days (2 years)
1d	1 day	Max available (up to 50+ years)
5d	5 days	Max available (up to 50+ years)
1wk	1 week	Max available (up to 50+ years)
1mo	1 month	Max available (up to 50+ years)
3mo	3 months	Max available (up to 50+ years)

Notes:

• Intraday data (< 1d) is typically only available for the last 60 days, except for 1m, which is limited to the last 7 days.
• Daily, weekly, and monthly data can be downloaded for the maximum history available, which may go back 50+ years depending on the stock.
• For intervals 60m, 90m, and 1h, data is available for up to 2 years.
• These limitations are set by Yahoo Finance, and attempting to request more data than allowed will result in an error or incomplete data.