Violence is declining — is the reasoning/model right?

Here’s a pdf or writing by NNT arguing the reasoning of violence is going down ignores the outliers, but that is not an acceptable assumption.

 

UPDATE: On a bit of thought, I realize that’s not explicitly argued, but only implied,  nay inferred by me, due some of historical writing.

UPDATE 2:  The original article can be crudely summarized as, the way Steven Pinker has eliminated the outliers(like wars) is valid only for a distribution that’s known to approximate the normal  distribution, but not for a fat-tailed distribution and violence is closer to being fat-tailed than normal distribution.