Here’s what classification should look like based upon a sound decision rule supported by rock solid statistical models. Just wondering if real world cases (as opposed to my data simulated example – chart idea courtesy of Hastie, Tibshirani, Friedman) do resemble this at all.