Over the last week I spent a lot of time implementing a stock trading algorithm that determines when to buy and sell based on the relationship between moving averages and closing prices. I had no prior experience working on these types of algorithms but I quickly realized that I really enjoy this type of work. I think that part of this enjoyment is motivated by the prospect of actually deploying algorithms like this (and not just backtesting) with the goal of making money.

As I had mentioned in a previous blog post, I have not had much success with my personal investing strategies. The fact that I can take my time to create an algorithm that only acts based on certain rules and that I can even backtest this algorithm to see how it has performed over any historical period of time is amazing to me and feels like a trading cheat code. With that said, historical performance is absolutely no guarantee of future performance, but there is a reason why “Algorithmic trading accounts for around 60-73% of the overall US equity trading” (Mordor Intelligence). This percentage clearly shows that institutions are not leaving their investments in the hands of human decision making and are instead making use of algorithms like the one I worked on, albeit much more complicated ones.

I am looking forward to continuing to work on this project and am excited to see what other algorithms we can come up with and how well they do historically.