Discover more from Theory A : Visualize Value Investing
Compressing Data for Holistic Views
Chat GPT inspired Data Compression
One of the best ways of thinking about AI language models such as Chat GPT is of a super smart auto-complete trained on the internet. It tries to fill in the blanks based on the surrounding context. It’s similar to Photoshop’s magic fill but with words.
Ted Chiang has an article where it’s compared to a “Blurry JPEG of the Web”.
Think of ChatGPT as a blurry jpeg of all the text on the Web. It retains much of the information on the Web, in the same way that a jpeg retains much of the information of a higher-resolution image, but, if you’re looking for an exact sequence of bits, you won’t find it; all you will ever get is an approximation
This is a fundamental idea behind the Theory A screener which cross references each data point with every other data point for each stock to create a highly compressed row of data.
In the example below, to calculate the decile for one column, say market cap for AAPL, Apple’s value has to be compared with every other value in our universe.
With 35 columns, and a universe of around 5000 stocks, each row embeds and compresses and information of 175,000 calculations.
Without the ability to visualize what exactly these numbers mean relative to other numbers, their screeners require you to memorize heuristics like “A P/E Ratio of 15 is Value” and know exactly what you are looking for before you even have an intuitive understanding of current market valuations.
The individual views also provide a great deal of information about a single stock, but the information is isolated.
One of the tenets of value investing is taking price into consideration when buying a stock. Even a terrible company can be a great buy if it’s cheap enough and even a great company can be a terrible buy if it’s over priced.
Many investors get carried away by meme stock mania and forget about price. They *need* to own a dogecoin or a share of TSLA or GME and they have $500 dollars in their pocket.