Last weekend, I re-watched one of my favorite movies: Spider-Man: Into the Spider-Verse.
In the movie, Miles Morales was trying to prove to Peter Parker that he is ready and capable… despite being new to the whole superhero gig.
Miles asks Peter how would he know when he’s ready to be a “real” superhero and able to help Peter and his friends save the world.
Peter’s reply was simple:
“You don’t. It’s a leap of faith.” — Peter Parker from Spider-Man: Into the Spider-Verse
Let’s rewind for a second. What does this have to do with algorithmic trading and procrastinating on important decisions?
The trading platform that could do everything and nothing
I spent over four years building a no code algorithmic trading platform. The platform can do things that no other trading platform can do, such as the ability to facilitate the financial research process. Users can find novel investment opportunities, uncover hidden patterns in the stock market, and even analyze their portfolio based on fundamentals and data.
For example, in this article, I used the AI within NexusTrade to disprove the existence of a Santa Claus rally. While I found that December has a higher risk adjusted returns than other months of the year, I found that this was not statistically significant.
I’ve also used the NexusTrade AI to find fundamentally strong AI stocks earlier this year. This list of 5 stocks, Microsoft, Broadcom, Nvidia, Salesforce, and Meta, are absolutely crushing the broader market, earning 2x the S&P500 since my initial article was published on Feb 20th.
Backtest of these 5 stocks from Feb 20th to nowThe NexusTrade platform could do everything a trader could possibly hope for.
Everything, except execute real trades.
Why didn’t I allow NexusTrade to execute real trades?
When I built NexusTrade, I built it for myself. I did not build this platform for other people to use.
Making an investing platform that helps strangers manage their money is scarier than fighting Kingpin (the man that killed Peter Parker). A simple misplaced conditional statement could wipe out someone’s life savings.
And that is a big deal. I was terrified of the possibility.
While I didn’t mind the risk of losing my own money, I couldn’t bear the possibility of making someone lose theirs.
So I stalled and built other neat features that would be useful for a trading platform. Some of these features include:
- Genetic Optimization: Automatically improving strategies using advanced AI Algorithms
- AI Powered Stock Screening and Financial Analysis: Finding new stocks based on objective historical data
- Creating strategies using LLMs: Creating fully autonomous trading strategies using Aurora, the AI-Powered Agent
Each of these milestones took a lot of effort and had their own unique setbacks. For example, the initial stock screener was far too slow and un-scalable – it was built using MongoDB, and utilized its unique query engine (the aggregation pipeline) to look over years of financial data for thousands of stocks.
This ultimately proved to be too slow, and it had to be re-built from scratch with a faster technology. I rebuilt the entire feature using BigQuery. The point is, this type of unexpected issue surfacing in live-trading could be catastrophic.
So working up the courage to implement live-trading, which could risk people’s hard-earned money, was horrifying. But, the feedback I received from my users was crystal clear: at the end of the day, people use trading platforms to make money in the stock market.
But how would I know when I’m ready to launch such a risky feature?
Back to the beginning: Miles Morales and Making the Hard Decisions
When I watched Miles Morales jump off the skyscrapers of Brooklyn and swing towards Kingpin’s Collider, his new-found confidence had become contagious through my TV.
I realized that I would never know when I’m ready to launch such an important feature.
I would just have to take a leap of faith.
Implementing the feature with safeguards
I decided to launch live-trading to users of NexusTrade. However, I did so in a thoughtful, calculated way.
The live-trading UI in NexusTrade- I first tested with $10,000 of my own money before anybody had a chance to use the feature
- Then, I released a small number of users from the waitlist
- Afterwards, I released more, and eventually opened up the feature for the wider public
- Finally, I ensured that I had automated alerting so that if an error does somehow occur, I know immediately. The user is sent a notification, and algorithmic trading for that portfolio is temporarily disabled
Most importantly, live-trading isn’t fully autonomous, at least not yet. Users can only execute manual and semi-automated trades.
The way it works is similar to paper-trading. NexusTrade generates the orders from signals, but it doesn’t automatically send them to the brokerage. Users have to manually confirm each order generated by their strategies.
In the future, 100% fully autonomous trading will be allowed (in some capacity) within NexusTrade, but I need to introduce it responsibly. While I’m capable of having my personal strategies run autonomously, I’m not yet ready to give such power to all 14,000 of my users.
Some ideas I have when launching this includes:
- Only allowing fully autonomous trading to people with enough experience
- Only allowing them for a subset of simple trading strategies (like Dollar Cost Averaging strategies that buys a single asset at regular intervals)
- Only allow them for people that have been paper-trading a strategy for a certain amount of time
With great power comes great responsibility. While I want nothing more to make fully autonomous trading accessible to everybody, I also have a responsibility to make sure to do so in a safe way.
Concluding Thoughts
Building an app from nothing takes a lot of guts. Imposter syndrome is real.
Allowing people to risk their hard earned money using your software is a scary experience. It takes a lot of guts because the consequences of a malfunction are astronomical.
But I know what my software is capable of. It's tested, and the only software in the world where the average person can become an algorithmic trader.
I started building NexusTrade over four years ago. Throughout the last year, I've implemented a variety of powerful and useful features for traders and investors of all experience levels. But this feature, live trading, is the most important.
Now, everybody, even non-technical investors and aspiring traders have a platform to create their own algorithmic trading strategies. This empowers anyone, regardless of technical expertise, to develop data-driven investing strategies and make more informed financial decisions.
This ultimately translates to making more money in the market.
When watching Into the Spider-Verse, I realize that I can’t just wait until I “feel ready”. I will never really know when it’s “the right time” to launch such an important feature…
And you never will either.
If you’re someone afraid to do something risky, such as launching a high-stakes feature or even asking a cute girl out, remember this: Nobody really knows when they’re ready to make an important decision or take such a risk.
You have to take a leap of faith.
Follow me: LinkedIn | X (Twitter) | TikTok | Instagram | Newsletter
