In my data set there's still something off about how that calculates. It's not entirely consistent to use his method