AI is powerful

Let’s use it to make better investments

The AI is built on GAAP accounting data from thousands of companies over the last 35 years. The model today evaluates an amalgamation of financial data and key metrics.

To be clear, this is not a trading algorithm predicting short-term trends or timing the market, but rather selecting quality businesses with knowledge of the underlying security.

Today, we use this as a screening tool for underpriced opportunities. We believe that over time, an AI crafted with the right principles can greatly outperform human judgment, while comparing market opportunities around the globe in an instant.

This simulation method selects 20 companies, evenly weighted, and adjusts the portfolio yearly. This simulation is based on a validation set, where the companies in are completely unknown to the AI before running. This is one of the benchmarks we use to track and improve.

The figure above shows the AI’s top picks in a simulation; however, we find that where the AI is most accurate is in identifying if a business is overvalued. Hence, we see the best application is to use it reductively - our adaptive funds will track an index, while removing these overvalued or weak businesses.