My Chat GPT-Generated Trading Strategies are DEMOLISHING the Market.
To be honest, I didn’t have high expectations when I started my experiment to evaluate the effectiveness of Chat GPT-generated portfolios. A language model called Chat GPT was developed using reinforcement learning, to practically predict the next word in a phrase. Furthermore, ChatGPT produced a very straightforward approach with just one purchasing rule and one selling rule. Imagine my astonishment when I discovered, a month later, that the tactics I created using Chat GPT were outperforming the competition. By a significant amount.
It should be noted that these are real-time outcomes from paper trading, not back test findings. The experiment will be summarized and the current, unmodified findings will be provided in this article. See this page for a comprehensive list of every article in the series: Overview: The Way to Create An Advanced Algorithmic Trading System.
Experimental Overview
This exercise was never intended to demonstrate that ChatGPT is a Wall Street trading virtuoso. It has always been to demonstrate how easy it is to create a portfolio using ChatGPT and instantly launch it to the market. I thought it would be a great experiment to launch the portfolios and see how they fared, based on the comments I received on the first piece. I had no idea they would perform that well.
Three simple steps were involved in the experiment:
Click here to sign up for a free account on NexusTrade.
Using the GPT-Powered Chat interface, create a portfolio.
Use Genetic Algorithms in the UI to optimize the portfolio.
We intended to try a number of different concepts throughout this experiment to determine how well they would do in the market. Among them were:
Something “un-optimized” Portfolio created by ChatGPT
An optimal portfolio created by ChatGPT at the beginning of the trial
Two portfolios created by ChatGPT are frequently reoptimized in marginally different methods.
a portfolio where the default is to “buy and hold SPY”
As of the latest update, the portfolios were only beginning to turn a profit. As of November 15th, every optimized portfolio is exceedingly successful and surpassing the buy-and-hold approach by a significant margin. Let’s go into more information about each of these portfolios.
Aside: The New and Improved AI-Powered Chat
Without a doubt, Nexus Trade’s AI-powered chat is its most potent feature. Instead of putting the consumer in contact with a customer service representative, the chat interface is far more engaging. It enables users to create trading strategies, do financial research, and assess those strategies using historical data from the past. This is how it operates:
A request enters the chat window.
The request has been properly categorized.
In order to reply to the user, the prompt that best fits their aim is utilized.
It has historically been somewhat challenging to keep these prompts going. They necessitate code modifications, which in the business frequently call for lengthy BUILD cycles, SDLC clearance, and configuration and updating by highly technical users. Furthermore, it might be challenging to assess the connection between several prompts because of their interdependencies. For this reason, I’ve begun working on Nexus GenAI, a Next-Generation AI Configuration Platform.
Discussion
The optimized ChatGPT portfoliosare now doing incredibly well, as the data above indicate. Through real-time paper trading, they are significantly exceeding expectations.
We must re-optimize the Expanding Window and Sliding Window Optimized portfolios, just as we did in the earlier articles. My objective is to attempt to strive for minimizing the drawdown of the present portfolio because of how well the portfolios are performing. This seems like this for the optimization settings:
Conclusion
Every sector will be revolutionized by AI, and banking is no different. This exercise demonstrates how AI may, at minimum, support your trading judgments. Although it’s too soon to say whether or not portfolios created solely by AI are better, what’s really amazing about this experiment is how quickly these portfolios were created.
What if we combined AI with a skilled trader? The trader has the ability to create several strategies for hundreds of assets with no correlation. The difference between traders employing AI and those not utilizing it will grow dramatically, and testing, refining, and enhancing their methods will become routine tasks. A super-intelligent agent is faster and more efficient than any human being. The finance industry is about to undergo a permanent shift.
I appreciate you taking the time to read! Await our upcoming update on these portfolios. Are you interested in using AI in finance? Get Aurora’s Insights by subscribing! Would you want to give the AI Chat a try? Open an account right now on NexusTrade!
