A couple weeks ago I wandered into a ridiculous Facebook debate about the 2000 Gore campaign, with the central hypothesis of the thread running “Al Gore lost his own state, therefore Nader voters in Florida were wholly irrelevant to the electoral outcome.” This, I will grant, is a stupid argument to become involved in. Nevertheless, I think that XKCD capably deals with the sort of logic present in this argument; broad historical statements about candidate performance (successful candidates tend to win their home states) should yield to analysis of local temporal and geographic factors. Al Gore in 2000 outperformed (relative to the national popular vote) the Clinton-Gore ’96 ticket in Tennessee, and wildly outperformed the ’04 Kerry and ’08 Obama tickets. This isn’t surprising, given that Tennessee was trending hard right from at least the mid-90s on, and has moved safely into the GOP column for the foreseeable future. A two point swing in the national popular vote would likely have resulted in a substantial Gore electoral college victory without nudging Tennessee a bit.
I would also note that if Mitt Romney loses Massachusetts by as much as Al Gore lost Tennessee, he’ll win about 450 electoral votes.