Trend Following System for Stock Index Trading

Abstract

There have been wide applications of trend following based trading system in the real world. The traditional methods include 5 days moving average, 20 days moving average, 200 days moving average and so on. If the midterm market trend is the biggest concern, 20 days and 50 days moving averages can provide efficient forecast information about stock trend. However, the solo stock price moving average does not provide sentimental information of the market in a timely manner.

This paper discusses a mathematical formulation of stock trend decision and the related prediction delay problem. An efficient method which utilizes the market sentiment Index has been proposed, aiming to forecast short term and long term market tops and bottoms. This new method leads to specific trading rules on two major indexes SPY and QQQQ. The experimental results show very accurate position estimation and support that our system implements much more reliable trading rules than the bench mark buy and hold strategy in terms of market exposure and account balance.

Keywords: Trend, Moving Average, Market Sentiment Index, prediction, trading rules