The article compares different ways to predict financial risk by looking at various measures of how prices change over time. By using a new model called the P-Spline Multiplicative Error Model, researchers found that certain volatility measures can help predict risk better than others. Specifically, trends in volatility and using realized kernels can improve predictions, but not as much as simply looking at the daily price range. This means that understanding how prices move can help make better predictions about potential financial losses.