User blog comment:Homuhomu123/Shelling Accuracy Test Results/@comment-24430017-20141217084536/@comment-25779449-20150102185443

As for the variables, the author has continued his study on each variable (formation, stat bonus from equipment, LoS, health state & success of detection) and each one was considered conclusive. I think if the data were collected rigorously, we can at least evaluate the effects from each factors. As for "made up a model and fit his data to it", I don't really agree on that, since he seems to be a person who dedicated enormous efforts on conducting the tests & writting the blog. Despite that, the data (#HIT & #ATK) doesn't really seem to be randomly made up.