The design considerations and development of a simulator for the backtesting of investment strategies

Abstract

The skill of accurately predicting the optimal time to buy or sell shares on the stock market is one that has been actively sought by both experienced and novice investors since the advent of the stock exchange in the early 1930s. Since then, the finance industry has employed a plethora of techniques to improve the prediction power of the investor. This thesis is an investigation into one of those techniques and the advancement of this technique through the use of computational power. The technique of portfolio strategy backtesting as a vehicle to achieve improved predictive power is one that has existed within financial services for decades. Portfolio backtesting, as alluded to by its name, is the empirical testing of an investment strategy to determine how the strategy would have performed historically, with a view that past performance may be indicative of future performance.