Mean Reversion Stock Trading System, ReversionSys

Made by rawsashimi1604

Introduction

Hey everyone! 👋 This is a stock trading system that I developed, using a trading software called Amibroker, Python and TWS Interactive Brokers API.

Huge thanks to erdiwit for creating ib_insync module to make programming this much more linear!

Required modules

• talib library (use whl to install, file in package)
• pandas
• numpy
• yfinance
• ib_insync
• pprint

Installation

• Install TWS Workstation
• Install required modules
• Run TWS Workstation
• Configure TWS API Settings (ip, socket, enable)
• Run python code
• Leave TWS on
• Use main.py to run get_trade_list or run_bot

Tips and Tricks

• Update S&P 500 Components every month to account for changes in index.
• Get list of trades everyday before the market opens.

How does the trading system work?

Research material and background

For starters, I was looking into some stock trading strategies and methods that have proved profitable so far. I came across this book, Short Term Trading Strategies that Work by Larry Connors.

Larry described some useful pointers backed by statistics such as:

• Buying pullbacks, and using the Relative Strength Index (RSI) to detect them,
• Buy stocks in an uptrend,
• Using intra-day pullbacks to get a better entry price.

As such, I used some of his parameters and tested it out on a backtesting software, Amibroker.

Trading System

Here were the parameters I used for my trading system.

• Moving Average Period = 100
• RSI Period = 2
• RSI Buy Signal = 5
• RSI Exit Period = 5
• RSI Exit Signal = 40
• Max Positions to Hold = 3
• Limit Order = 1.5%

For the stock universe, I used only the S&P500 Index Stocks.

As for position sizing, each time a position was available, it took up 33% of my portfolio.

As for the ranking system, I chose stocks that ranked the highest in the Rate of Change(ROC) = 100 and ROC > 0 parameter. This is to choose stocks with high relative strength only.

Example of the trading system (using PineScript from Tradingview)

Here is an entry example from TradingView

Here is an exit example

For more information, you can refer to the amibroker AFL code.

Getting a trade list everyday.

To get the list of trades today, we use the function get_trade_list.

def get_trade_list(file_type, components_path=r"C:\\Users\\rawsashimi1604\\VisualStudio\\Reversion_Sys\\ReversionSys\\S&P500 Components.csv", export_path="C:\\Users\\rawsashimi1604\\VisualStudio\\Reversion_Sys\\ReversionSys"):
  # some code
  returns None

Parameters:

• file_type : str
  • Use either "csv" or "txt"
    • Specifies what file_type to receive today's list of trades in.
• components_path : str
  • Specifies where to find list of tickers to trade in. (S&P500 Component List)
• export_path : str
  • Specifies where to export today's list of trades.

Returns:

• None

To run the bot we use the function run_bot.

def run_bot(path = "C:\\Users\\rawsashimi1604\\VisualStudio\\Reversion_Sys\\ReversionSys\\2021-01-25 Reversion Trades.csv", market_open_time = "22:30:05", ip='127.0.0.1', socket=7497, clientId=1):
  # some code
  returns None

Parameters:

• path : str
  • Specifies where list of trades csv is.
• market_open_time : str
  • Specifies what time in local time the stock market opens. Default is "22:30:05".
• ip : str
  • Specifies TWS IP Address Settings. Default is "127.0.0.1".
• socket : int
  • Specifies TWS Socket Settings. Default is 7497.
• clientId : int
  • Specifies TWS ClientID Settings. Default is 1.

Returns:

• None