Academics used AI and machine learning to test 29,000 financial ratios. Here’s the metric they found was the best predictor of stock-market returns

• Alejandro Lopez-Lira, along with colleagues, used machine learning to test 200 investing theories.
• They also set the algorithm up to seek out ratios with better performance returns. 
• The algorithm identified a specific ratio involving sales from acquisitions and rental expenses.

Determining the best predictors of stock-market returns is like seeking the Holy Grail for investors. The problem is that the number of deciding factors can seem limitless, while the human power to test them all is not.