Yeah it's essentially a way to reflect parsimonious assumptions so that your output distribution can be characterized as a law.