Home Statistics and Visualization R-bloggers “You should always (always) have a Naive model. It’s the simplest, cleanest, most intuitive way to explain whether your system is at least treading water. And if it is (that’s a big IF), how much better than Naive is it.”