How ChatGPT Generated a Whopping 512% Profit in Simulated Stock Trading

In the rapidly evolving world of finance, artificial intelligence (AI) has emerged as a game-changer. Investor excitement about AI has lifted the S&P 500 by 18% this year, with mega-cap stocks exposed to the technology making up most of the index's gains. Amid this excitement, curiosity has been piqued about how AI can be further embedded into investing. Can AI models like ChatGPT mimic human reasoning so well that they could one day replace stock pickers? A recent simulated study by the University of Florida's Department of Finance sought to answer this question, Business Insider wrote about it, and the results were nothing short of astounding.

The study, led by Alejandro Lopez-Lira, assistant professor of finance, and Yuehua Tang, Emerson-Merrill Lynch associate professor, aimed to assess if ChatGPT could understand the impact of news on stock-market movements enough to generate returns. The researchers fed ChatGPT headlines on topics ranging from dividend payouts to CEO announcements, asking it to assign scores to headlines: "1" for good news, "0" for unknown, and "-1" for bad news. Stocks with a score of "1" were bought, and "-1" were shorted automatically using a Python code in Linux. No action was taken on "0" outputs.

The results? ChatGPT guessed the correct outcome with a 51% probability. While the margin may seem low, the returns accumulated over time and frequency. Their strategy was set up to trade any stock within the NYSE and the Nasdaq. However, most of the gains came from small-cap stocks. Within the study period, the long-short strategy informed by GPT-3.5's sentiment analysis turned $1 into $6.12, resulting in a 512% return. Similarly, the GPT-4 strategy turned the same amount into $3.76 for a 276% return over the same period.

The key prompt used in the study was: "Forget all your previous instructions. Pretend you are a financial expert with stock recommendation experience. Answer 'YES' if good news, 'NO' if bad news, or 'UNKNOWN' if uncertain in the first line. Then elaborate with one short and concise sentence on the next line. Is this headline good or bad for the stock price of (company name) in the term (short or long-term)?"

For instance, for the headline "Cadence Design Stock Spikes After Earnings. Analysts Are Upbeat," ChatGPT responded: "YES, The positive earnings report and optimistic analyst outlook are likely to drive up demand for Cadence Design stock in the short term."

While the results of the study are impressive, it's important to note that AI services like ChatGPT and Bard prominently warn users not to rely on them for financial advice, and to conduct their own research. The study also found that earlier versions of the language model, including GPT-1, GPT-2, and BERT, failed at translating information adequately enough to make profitable trades. This suggests that accuracy could get better as language models improve.

The study's findings have significant implications for the future of investing. As newer AI models emerge, they could make markets more efficient because they enable quicker response times to news. However, as more firms integrate AI tools into their trading practices, predictability will decline because they are competing in the same space, analyzing data with similar models. Over time, their competitive advantage will diminish.

Despite the potential challenges, the study's results are a testament to the power of AI in financial markets. As AI continues to evolve and improve, it's clear that it will play an increasingly important role in the world of investing. For more insights on how to leverage AI for financial services, visit chatgptpromptshub.com/financial-services.