The problem with this whole approach is that they still have this idea that natural
variation is a small «error» ontop of a
huge known, and not that natural
variation is almost all the signal of which a small part ought to be human
induced, but you really can't tell because the noise dominates the signal.