Back To Blog
Stock Screener

Replacing Your Stock Screener with ChatGPT?

April 11, 2023 - 12 min read
Testing Chat GPT Capabilities
The hype around the endless possibilities of AI technology, especially the ChatGPT chatbot has just begun. But is the chatbot also competent in investment topics or could it even make the use of Stock Screeners obsolet? We had to get an answer.

In this article, we present our five questions and ChatGPT's answers to specific investment topics, along with our own insights. To get a straightforward response to this question, we thought, why not ask ChatGPT directly!



Our Question 1: "Can ChatGPT replace the functionality of a stock screener?"


This was the ChatGPT answer:
Question 1 to ChatGPT*******************************************

"ChatGPT can provide general information and guidance on how to select stocks using various filters and criteria, but it cannot replace the functionality of a stock screener. A stock screener is a tool that allows investors to quickly filter and sort through thousands of stocks based on specific criteria, such as market capitalization, dividend yield, earnings growth, and other financial metrics.

While ChatGPT can provide insights and advice on selecting stocks, it cannot perform the same level of analysis as a stock screener. A stock screener allows investors to easily narrow down their options and identify potential investment opportunities based on their individual preferences and investment goals.

It's important to note that both ChatGPT and stock screeners are tools that can be used in conjunction with other research and analysis methods to make informed investment decisions."
*******************************************

BINGO - HURRAY! Thank you so much ChatGPT!

However, is it really that easy though?

We could have easily wrapped up the article here and felt content knowing that screeners are indispensable. But no, we decided to push further and find out just how much knowledge this AI has when it comes to investing... or lack thereof.

We opted to inquire ChatGPT with further queries, starting with some basic ones.

Our Question 2: "Which are the top 10 stock listed companies in terms of market capitalization in the US Stock market?"

This was the ChatGPT answer:
Market Cap: Top 10 in the US by ChatGPT*******************************************
"As of my knowledge cut-off of September 2021, here are the top 10 publicly traded companies by market capitalization in the US stock market, according to data from the Global Financial Database:

  1. Apple Inc. (AAPL) - $2.47 trillion

  2. Microsoft Corporation (MSFT) - $2.14 trillion

  3. Alphabet Inc. (GOOGL) - $1.82 trillion

  4. Amazon.com Inc. (AMZN) - $1.68 trillion

  5. Facebook Inc. (FB) - $1.01 trillion

  6. Tesla, Inc. (TSLA) - $775 billion

  7. Berkshire Hathaway Inc. (BRK.A) - $622 billion

  8. JPMorgan Chase & Co. (JPM) - $464 billion

  9. Visa Inc. (V) - $483 billion

  10. Johnson & Johnson (JNJ) - $446 billion

Please note that these values can change quickly due to ......."
*******************************************

Allright! The list looks quite good so far, but it turns out ChatGPT has a limit, which isn't really its fault - it's the creators. The financial database was only recently updated in September 2021. This could potentially put an end to investment research as we know it. Actually under the ChatGPT FAQ we found a notification about this limitation.
Data cuts off in 2021 on ChatGPT

However, let's cross-check the GPT results with that of SimFin Screener. Fortunately, we can look back in time and compare the outputs of the top 10 market caps from September 2021.

This is the screenshot of the SimFin Screener Table View Tab, which was loaded with the "Market Cap" metric and sorted in descending order. We can see that the Top10 companies from market capitalisation match with the output from ChatGPT. Great job ChatGPT!

Top 10 Market capitalisation US companies in September 2021This is a historical line graph in SimFin Visualization tab with market cap of the top 10 stocks, highlighting the September 2021.

Historic Line Graph with Stocks displaying Biggest Market Cap

Our Question 3: "What was the Equity per Share of the AAPL stock in January 2021?"

This was the ChatGPT answer:
Equity per Share of Apple reported by ChatGPL*******************************************
"The equity per share of a company is calculated by taking the company's total equity and dividing it by the number of outstanding shares. As of January 2021, Apple Inc. (AAPL) had 16,997 million outstanding shares and a total equity of $65,269 million, according to its financial statements.

Dividing the total equity by the number of outstanding shares gives us:

Equity per share = Total equity / Number of outstanding shares Equity per share = $65,269 million / 16,997 million Equity per share = $3.84

Therefore, as of January 2021, the Equity per Share of AAPL stock was $3.84."
*******************************************

A well-structured response, even with a calculation! Let's compare this value with the SimFin Screener. For the day January 29 2021 we get a value of $3.86, which is almost identical to the $3.84 found by ChatGPT. Again a valid answer from the bot.

Equity per Share AAPL

Our Question 4: "Give me a list of filters for picking stocks that had a high ROI in the last 10 years."

This was the ChatGPT answer:
Filters for stocks for high ROI in ChatGPT*******************************************

"Here are some potential filters that could be used to identify stocks that have had a high ROI over the last 10 years:

  1. Revenue Growth: Look for companies that have consistently grown their revenue over the last 10 years. Revenue growth is a strong indicator of a company's ability to increase profits and generate returns for investors.

  2. Earnings Growth: Look for companies that have consistently grown their earnings per share over the last 10 years. Earnings growth is a key metric for measuring a company's profitability and potential for future growth.

  3. Return on Equity (ROE): Look for companies that have consistently achieved high ROE over the last 10 years. ROE measures how effectively a company is using shareholder equity to generate profits and is a key indicator of a company's financial health.

  4. Price-to-Earnings Ratio (P/E Ratio): Look for companies with a relatively low P/E ratio compared to their peers. A low P/E ratio may indicate that a company is undervalued by the market and has potential for future price appreciation.

  5. Dividend Yield: Look for companies that have consistently paid and increased their dividend yield over the last 10 years. Dividend-paying stocks can provide steady income and may indicate that a company has a stable financial position.

  6. Market Capitalization: Look for companies with a large market capitalization. Larger companies may have more resources to weather market downturns and may have a higher potential for long-term growth.

It's important to note that these filters are just one possible way to identify high-performing stocks, and investors should conduct thorough research and consult with a financial advisor before making any investment decisions..."

*******************************************

Again a well-structured answer from Chat GPT. It includes six tips for indicators, which we will now setup in the SimFin Stock Screener. The resulting stock list we will than backtest in SimFin against the price data of the last 10 years to evaluate the profit of this portfolio with suggested filters by ChatGPT.

This is the setup of the six filters suggested by ChatGPT in the SimFin Screener.
Screener Filters suggested by ChatGPTThese are the 37 stocks that were filtered out in SimFin Screener on base of data from the 30 March 2022.

List of filtered Stocks (ChatGPT Strategy)The next step was to check if these filters would have resulted in a portfolio with high ROI since 2013. Therefore, we started with few mouse-clicks a backtesting in the SimFin app, setting the start date to January 2013 and selected a monthly rebalance interval. Next, you see the performance chart of the backtest performed with SimFin, which also shows the S&P500 as benchmark.
Backtest Result of Strategy suggested by ChatGPTChatGPT's suggested portfolio filters would have made a ROI with a 10.2% CAGR, which was 0.6% higher than the S&P500's (9.6%) in the same period. With this investment result in real life, you would probably not fire your financial advisor immediately but you could have also simply bought the S&P500 ETF. Needless to say, this isn't the kind of result that would get investors excited.

What does this mean? Did we already reached the limit of ChatGPT's financial competence? Well, we should consider that ChatGPT suggested some filters but did not gave any filter values. This means, we acted as if we have no own financial understanding and therefore just set the filter values to be bigger or smaller than the median of all companies in the database. We decided to take it a step further and inquire ChatGPT for any additional values for its recommended filters, in order to offer our strategy more configuration potential.


Our Question 5: "Give me a list for good values for the filters, revenue growth, earnings growth, return on equity, price-to-earnings ratio, dividend yield and market capitalisation for building a high growth investment strategy."

This was the ChatGPT answer:
Filter values for investment indicators from the Bot*******************************************

"When building a high growth investment strategy, the values for various filters can depend on a range of factors, including the investor's risk tolerance, investment goals, and time horizon. However, here are some general values for the filters you mentioned that could be suitable for a high-growth investment strategy:

  1. Revenue Growth: Look for companies with a revenue growth rate of at least 10% year over year (YoY). High-growth companies may have even higher revenue growth rates, such as 20% or more YoY.

  2. Earnings Growth: Look for companies with earnings growth rates of at least 15% YoY. High-growth companies may have even higher earnings growth rates, such as 25% or more YoY.

  3. Return on Equity (ROE): Look for companies with an ROE of at least 15%. High-growth companies may have even higher ROEs, such as 20% or more.

  4. Price-to-Earnings (P/E) Ratio: Look for companies with a P/E ratio below the industry average. High-growth companies may have higher P/E ratios due to their growth potential, but it's important to consider whether the P/E ratio is sustainable given the company's financial performance and growth prospects.

  5. Dividend Yield: High-growth companies may not always pay dividends, as they may prefer to reinvest earnings in growth opportunities. However, if you are looking for dividend-paying stocks, look for companies with a yield of at least 2%, or higher depending on your preferences.

  6. Market Capitalization: High-growth companies can be found across a range of market capitalizations, but smaller companies may have more growth potential. Look for companies with a market capitalization of at least $1 billion, or smaller if you are willing to take on higher risk.

It's important to note that these values are general guidelines, and investors should conduct thorough research and analysis on individual companies before making investment decisions. Additionally, investors should consider diversifying their portfolios and not rely too heavily on any one stock or industry."
*******************************************
Appreciate it ChatGPT! This is what we were hoping for - more defined requirements. We took ChatGPT's advice and set criteria on SimFin Screener based on the info it gave us.

Indicators in Screener with defined valuesOf course, we wanted to see how this strategy would have performed in the last decade. Therefore, we started a new backtest with same parameters as in the first backtest.

Second backtest with improved valuesThe result was disappointing. Only 5% CAGR! We did not think that the indicators were so badly selected but we decided to add another indicator. We modified the strategy in the SimFin Screener to be narrowed only for companies from the Technology sector by a seventh filter. In addition, we applied the unique SimFin feature configuring the suggested indicators not as filters but as scores. Moreover, we decided that all companies stay in the portfolio that fulfil at least 4 of the 6 score criteria (Total Score > 3). Means we "softened" the set of filters.

After altering the SimFin's stock screener settings and ran the backtest again.
3rd backtest with additional technology filterBoom! The results were astounding! 15.4% CAGR, a massive 6% more than S&P500 per year - what an incredible achievement! The backtest reveals that the strategy would have provided an overall return of investment of 319% in 10 years. You can verify the portfolio and its backtesting result here: https://app.simfin.com/portfolio/64286a2ba179b0713b8d1e6b

After performing the above tests, it appears that the ChatGPT filters are suitable candidates for further optimization in order to increase potential ROI. However, in the given case, an additional refinement was required to truly maximize efficiency.

We also have to admit that our tests just reflect a small spectrum of the possibilities and variations existing in ChatGPT and our SimFin Stock Screener. We welcome everyone to explore and experience the potential of using these tools for their financial analysis. Seeing the results for yourself can be a great way to figure out how it can benefit your investments.

Conclusion

Our initial question "Can ChatGPT replace a Stock Screener?" can be answered straightforward with "NO". The handicap, that the ChatGPT datasets are not younger than September 2021 ist just a minor reason. Despite ChatGPT giving precise and correct answers for financial metrics in questions 2 and 3, and providing suitable filters for question 4, it failed to provide all the required filters that would be necessary for an investment strategy providing a high return on investment. The values as response to our fifth question were not enough. At the end it took us some effort and manual tweaking of criteria and backtesting to create a potentially effective investing strategy. ChatGPT can make suggestions with regards to filters and values, however, it cannot determine if the stock picking criteria chosen would actually result in a profitable investment.

Generally speaking, ChatGPT can give novice investors some initial ideas to start with, while stock screeners provide users with more in-depth analysis and insights to help them identify potential investment opportunities.

However, we are curious as to how the AI world is evolving, and we will definitely see more and more features in financial analytics taking advantage of this trend in the near future.

If you have already had experience with AI in stock screening, we would be delighted if you share them with us.

For anybody that like to create and test his own strategies or suggestions given by AI, here are the links:
https://app.simfin.com/
https://chat.openai.com/chat


Important: The information provided in this blog is for educational and informational purposes only and should not be construed as investment advice. The author is not a registered investment advisor and does not provide personalized financial or investment advice. All investments involve risk, and past performance is no guarantee of future results. It is recommended to consult with a professional financial advisor before making any investment decisions. The author and publisher shall not be held liable for any losses incurred as a result of the use of the information presented in this blog.

Share this article
Financial values - visualized schematically

Start analyzing for free and pay as you grow.

Analytics tools & financial data for investors, analysts and students to learn and grow.
Related posts
  • Major SimFin Update - Jan 2024
    We just deployed a major updated that happened after we completed the biggest data migration so far at SimFin, where we transferred all our data to a brand new data extraction tool that we have been working on for a long time.
  • Data Extraction
    Extracting Document Data: The Best Tools to Stay Ahead in 2023
    Navigating the world of data extraction in 2023 can feel like a maze. But don't worry, we've got you covered! In this guide, we'll break down the seven top ai document processing tools out there. This article is your handy roadmap to mastering data extraction from PDFs, web pages, e-mails or images. Dive in and discover the tools that can supercharge your work!