[kkoyung]

I agree. Using the terminologies from the author, the DK paper was trying to show that dy/dx < 1 = dx/dx, rather than the correlation of y-x vs x.