Maybe not everybody is cut out to be an algorithmic trader
It takes a certain passion, determination, and personality to succeed
I naively once thought that anybody could become an algorithmic trader.
I was wrong.
Easier than ever before
The reason why I thought this is because algorithmic trading is easier now that it ever has been before.
Especially with the advent of large language models, the amount of skill required to be an algorithmic trader is now at historical lows.
For example, back in the day if you wanted to create your own algorithmic trading strategies, you would have needed to know the following things:
- You needed a deep understanding of financial markets
- You needed to spend hundreds of dollars each month for high-quality financial data
- You needed to be an expert-level programmer, and at the bare minimum extend from an existing open-source platform
- Ideally, you would also have a math background, with an understanding of statistics, probability theory, and other high-level mathematics
Nowadays, that is no longer the case. You just need some determination and willingness to learn. Let me demonstrate this with a before and after comparison.
Coding is no longer a requirement
The biggest advancement in algorithmic trading is that now coding is completely optional.
In fact, it’s now faster to make progress without coding than it is with code.
For example, let’s say you want to see if companies with a higher market cap tend to be even higher years later.
Traditionally, you would need to be wrangling a whole bunch of data, and then you have to code up your experiments to find your conclusions.
More specifically, wanted to test out whether or not the companies in 2023 with the highest market cap tended to perform better than average in 2024, you would have to do the following:
- Identify 10 companies with the highest market cap as of December 31, 2023
- See how well those companies have performed from Jan 1st, 2024 to now
- Compare results to investing in the broader market (e.g, the S&P500)
Of course, this is a simple experiment. In reality, you may want to perform your experiment on all slices within the past decade. But even with this simple experiment, it is still a lot of work.
Nowadays, if you wanted to perform this experiment, all you’d have to do is this:
- Go to NexusTrade.io’s AI Chat
- Ask her the following question
What 10 companies had the highest market cap as of Dec 31st 2023
From here, you can see how well those companies performed from Jan 1st, 2024 to right now by:
- Creating a portfolio with a strategy to buy all 10 of these companies
- Backtesting that portfolio
To create the portfolio, we simply ask Aurora.
Then, we can backtest in the same way.
The initial results are definitely promising. We see that the stocks with the higher market cap achieved better returns in 2024, but had a lower sharpe and sortino ratio, suggesting that they are more volatile. The average drawdown and maximum drawdown are also both double of the broader market, further suggesting the increase in volatility.
From this experiment, we have some evidence to suggest that stocks with a higher market cap tend to perform better the following months. To validate this, we could test this out using sliding windows from 2016 to now, and see if this pattern held true for the past decade.
Keep in mind, we just automated away months (if not YEARS) of work. Performing this experiment from scratch would have taken a Jane Street employee-of-the-month weeks of work. Data-driven insights from natural language are something that wasn’t heard of in the financial industry before the rise of Large Language Models.
And now, its a reality.
And yet, despite how much easier it is to do algorithmic trading, it is still too difficult for the average person.
The average person still can’t do it
Being an algorithmic trader requires a certain personality. Somebody that’s not afraid to figure things out when they don’t know. Someone not afraid of jargon, and willing to go the extra mile to understand what it means.
And that’s simply not true for everybody.
I’ve received countless emails from people asking me how do they make money with my platform. When I send them YouTube videos or Medium articles on how to use the platform, they still don’t get it.
They want me to spoon-feed them an answer.
And I’m not pretending my platform is perfect. There’s bugs, it gets things wrong, and the UI needs some work.
But compared to starting from scratch with nothing, the difference is night and day.
You can perform financial research. You can test out crazy ideas using both technical and fundamental data. And when you’re done, you can launch your portfolios to the market with the click of a button.
And that’s still not easy enough.
Thank you for reading! Do you think you have what it takes to be an algorithmic trader? Create an account on NexusTrade today!
Follow me: LinkedIn | X (Twitter) | TikTok | Instagram | Newsletter
