Basics of algo trading

- [Instructor] It's important to understand when we're talking about the financial industry, the industry's evolving. The reality is, whether we like it or not, computers are playing a bigger and bigger role. Algorithmic, or computer-driven trading, now makes up the large majority of trades in the financial markets. More than 90% of orders are algorithm, some metrics and estimates. So you might wonder, where do humans come in? Well, the reality is, humans still play a role in trading. Among other things, humans design algorithms. Computers are great at executing the trades, carrying out the strategy that humans have developed. The role that you should think of for yourself from an algorithmic trading perspective, is you are the general that tells all your computer soldiers what to do. That means that you need to be familiar with how the markets work and how to develop those algorithmic trading strategies. That's what this course is all about. Now, how does algorithmic trading work? Well, essentially, all algorithmic trading relies on kind of a four step process. We're going to start with some sort of pre-trade analysis. This is built around coming up with the strategy that we're going to follow in our algorithmic trading. What is the model that we are using to try and generate positive returns? This could be based on technical factors, it could be based on fundamental research, but whatever we're doing, we need some sort of a quantitative model that a computer can then use to identify attractive trading opportunities. This is where the trading signal comes into play. Once we've figured out what our strategy is, the computer then will evaluate our strategy and look for opportunities in the markets. Here's a dumb example. Let's pretend that we think that stocks that begin with M will go up on Mondays and stocks that begin with T will go up on Tuesdays. Well, the computer would go through and find all the stocks that begin with M on Mondays and say that we should buy those. The pre-trade analysis is figuring out what letter stock we should buy. The trading signal then goes through and identifies those stocks. Then, we execute that trade. The trade gets pushed through. We do any sort of post-trade analysis to make sure that the models working is appropriate, and we start the whole process over again. This relies on two key inputs: data and research to help us understand that data. These factors are critical in putting together an effective algorithmic trading strategy. Now, algorithmic trading, or algo trading for short, essentially comes in two flavors: market making and data mining. Market-making trades are attempting to capitalize on what's called the bid-ask spread. That is the difference between what people will buy or sell a stock for. This is typically associated with high-frequency traders. Data-mining, instead, is based on patterns in data, things like stock prices and then outside information. Essentially, here, we're looking for correlations or relationships between stock prices and other data points and then trying to capitalize on it. And at a high level, that's what algorithmic trading is all about.