You are correct; I am using number of horizontal lines rather than computing the actual DPI. But this barely changes my argument, since even when they are at screens of similar size cheap iPads have double the number of horizontal lines. I have updated my post to reflect that.