We can now add Gilbert's 25% line, and it becomes clear why the data supports both:
Gilbert's allegation that a fair election requires this graph to be a straight line is a geometric error, and claimed "illegally".
Thanks for putting this example together, Mendel. If someone can tell me how to do parameters (preferably with sliders) in Geogebra, I can put together an interactive version so that it can be verified that whenever one side's voters avoids mail-in voting, the non-mail-in vs. mail-in plot will have the deviation you show from the nonsense y=ax+0 claim. Of course, outliers being rare, and the curve being quite gentle, seeing the curve in real data is rare, so he's not out of line to approximate the data he sees to a straight line. Just not the straight line he claims
The deviaton is pretty clear - as I wrote in post #8 "you'd expect the whole scatter plot to be even higher on the graph, and more to the left."