Algorithmic trading based on fundamentals — my trading platform’s most powerful feature
Make decisions based on the most important data
Price is the most common “indicator”. When I refer to an indicator, I’m talking about anything that can be evaluated as a number and used to inform trading decisions. There are other indicators too like trading volume and the value of the positions in a portfolio. But price is the most popular.
Most of the technical indicators that we’re familiar with, such as simple moving average, rate of change, and moving average convergence divergence (MACD) use some derivation or transformation from the stock’s price. As a result, most trading platforms really only show you these types of indicators.
However, there is a whole world of indicators that have a predictive power on an asset’s future price. For example, inflation, Consumer Price Index (CPI), and fundamental metrics are much more informative than using an asset’s price alone.
That’s why I’m so excited about this new feature for NexusTrade. The feature is so simple, yet so powerful, and is one the majority of platforms don’t offer. This includes brokerages like Robinhood and Fidelity, and algorithmic trading platforms like Composor.trade, and Capitalise.ai.
I’m talking about the ability to test, optimize, and deploy fully automated trading strategies based on fundamental indicators.
NexusTrade’s History with Fundamental Data
Early last month, I began designing a variety of new features involving fundamental data. It started with an idea of an AI-Powered stock screener.
I outlined what an “AI-Powered” stock screener would look like and what were the technical requirements. I figured out that in order to search for these metrics, I had to ingest historical fundamental data for every US stock so that they can be queried in real-time.
After outlining how the feature would look like on February 8th, I implemented the first version of the feature on February 12th.
This feature is incredible; it allows a non-technical trader to find novel investing opportunities easily, without having to worry about configuring specialized tools or learning how to code.
After I launched the feature, I immediately started making it more robust. I started integrating Retrieval Augmented Generation (RAG), few-shot learning, and improving the prompt engineering.
The value of this feature is not to be understated. For example, on February 19th, I used it to identify 5 AI stocks that were poised to make significant gains during the AI revolution.
All 5 of the stocks, particularly NVIDIA (NVDA) had massive short-term gains since publishing that article.
Historical Fundamental Data
After implementing the stock screener, I realized that I had a rich database of fundamental data. The overwhelming majority of trading platforms don’t show users historical fundamental metrics. For example, you can’t log onto Robinhood and figure out what NVIDIA’s free cash flow was this past quarter. Thus, I decided to become the change I wanted to see, and implemented descriptive fundamental charts for premium NexusTrade users.
I discuss this new feature in more detail here. This feature allows users to make decisions based on more than just pure price data. In fact, I used this data to inform my decision of buying Robinhood (HOOD) call options before a major run-up.
It’s clear that fundamental metrics are important for making trading decisions. What would be awesome is if we could incorporate them into our NexusTrade trading strategies.
We now, in fact, can.
Algorithmic Trading using Fundamentals
This was something that I first thought of nearly 2 years ago. This was before NextTrade, the open-source predecessor of NexusTrade, was open-sourced. The challenge in implementing this was that I needed a way to configure any indicator – not just price-based technical indicators, but more abstract indicators. This would’ve been nearly impossible with NextTrade, which was one of the reasons why it was open-sourced.
However, NexusTrade doesn’t have the same limitations as its predecessor. The main selling point between it is that its much faster and more configurable. And today I proved it.
I started with creating a design to figure out what additional work had to be done.
I then implemented the changes as I described and modeled the indicators based on the existing architecture. Now with NexusTrade, it’s possible to make trading strategies based on fundamental data. This means metrics like revenue, net income, gross profit, Price Earnings (PE) ratio, and free cash flow. Moreover, not only can we use these raw values, but we can take averages and measure the rate of change of these indicators.
Again, I can’t understate how powerful this is. Most traders are using the same exact data source — asset price. Incorporating fundamental data allows you to create profitable trading strategies that are based on a company’s financial health. Like all indicators in NexusTrade, these metrics can be optimized using genetic algorithms.
So, if you think something like revenue and free cash flow matters, you can now prove it to yourself objectively.
Next Steps with Nexus
Now that I’ve proven the architecture is flexible enough to handle fundamental data, I can now incorporate other alternative data sources. The two that come to mind immediately are sentiment data from news articles and economic data like the Consumer Price Index (CPI). The only challenge is sourcing the data in a cost-effective way. I’ve already proven that the architecture is flexible enough to do this.
Drawbacks of this new feature
Unfortunately, this feature is not perfect and has a number of drawbacks. The biggest drawback is that the fundamental data is not available in real-time. Specifically, there is a delay between when the data is made available to investors and when it is made available in the platform. This delay comes from the fact a company may announce its earnings publicly, but then publish their earnings to the SEC Edgar Database a couple of days later.
Because of the delay, the stock most likely already moved up (or down) before the trading platform is aware of the data. The good news is that this movement is already accounted for during backtest. The bad news is that this delay does significantly affect the stock’s price, particularly if you plan to make decisions on the day a company reports earnings. Investors relying on this tool MUST be aware of this discrepancy, and understand the consequences of this delay.
Conclusion
In this article, I summarized perhaps the most powerful features of NexusTrade – a strong integration with fundamental data. I detail the history of this integration and showed how incorporating such data into your strategies can improve your investing decisions. I’ve also outlined some next steps with this brand new feature.
Overall, I’m extremely excited about this. Fundamental data is important when it comes to trading and investing, and this integration proves that NexusTrade is capable of executing any strategy the end-user can imagine. Other automated trading platforms, such as Composor.trade and Capitalize.ai are ONLY capable of creating strategies using derivations of the stock’s price. Hopefully, this feature sets a precedent on the type of data other platforms are expected to have when creating algorithmic trading strategies.
Try out the new feature today!
