Oh, Nanos, you mercurial scamp:(38/27/21/8/6)

What was that about a Liberal bias in his methodology? Over/undersampling? Just being wrong?

Regionals, analysis, pretty graphics here for the clicking.