No. of Recommendations: 5
A June 2025 study “Volatility: A Dead Ringer for Downside Risk” says volatility (the standard deviation of returns) a good measure of the risk that investors care about (compared to alternative downside risk metrics: semi-deviation (SSD), probability of loss (PL), average loss (AL), expected loss (EL), worst loss (WL), maximum drawdown (MD), and Value at Risk (VaR).) It used data sampled from the MSCI database of 47 countries (23 developed and 24 emerging) each considered from its inception in the database through December 2024.
The ranking made by volatility was most similar to those made by SSD, AL, EL, and VaR, with correlations 0.90 or higher.
No. of Recommendations: 2
Agrees with what I’ve found. Two of the highest performing factors for the machine learning features I’ve tried are the volatility of earnings and price. For both tree and linear type algorithms.