Reply to thread

Message: <blockquote data-quote="Jay Verkuilen" data-source="post: 7735973" data-attributes="member: 6873517"><p>I think these are very good points, particularly about not differentiating segments of players. I also question how good the analyses they do are. I have no insight into their data or internal processes, of course, but speaking as someone in the profession of statistics I've not been super impressed by what I've seen from their analyses. In particular they don't seem to be too mindful of the issue of selection bias. Data in the hands of non-experts leads to frequent Dunning-Krueger Effect type errors. Of course, I am mindful I'm just speculating without having seen the data myself, but my general experience is that getting good data is much, much harder than getting shaky data.</p></blockquote><p></p>