Stock Trading Quantified: An Exploration of Algorithmic Trading Principles using QuantConnect

My thesis is an exploration on the principles of algorithmic trading. I was introduced to the world of algorithmic trading in the Summer of 2018 when I got an internship at a startup trading firm called Helios Machine Intelligence. At HeliosMI, my job was to model algorithms for their in-house developed platform (in Java and C#). I learned how to model several different strategies, but I didn’t understand how, or more importantly, why these strategies worked. In the Spring of 2019 when I first began planning my thesis, I initially planned on recreating and optimizing HeliosMI’s trading platform. It was after reading a few books over the summer, namely; The Man Who Solved the Market by Gregory Zuckerman, Algorithmic Trading by Ernie Chan, and A Random Walk Down Wall Street by Burton Gordon Malkiel, that I realized that I was much more interested in learning the fundamentals of algorithmic trading, so I decided to make this the new focus of my thesis. At HeliosMI, we tested strategies against the historical data of stocks using an application called QuantConnect. This application is easy-to-use, cheap (even offering a free tier) and provides plenty of documentation with an active community forum, making it the obvious choice as the platform for my thesis research. Throughout my research I focused on exploring high-frequency trading algorithms, mainly because these are the types of algorithms that are employed at Wall Street hedge funds, and also the type I worked on at HeliosMI. I developed three distinct algorithms throughout my research; a momentum based strategy, a mean reversion based strategy, and a preferred time of day based strategy. In my thesis report, I go in depth on each of these strategies, as well as discuss the history of algorithmic trading, and explore some future research aspirations.