Some random thoughts:
* I like how you split up independents into South-based and non-South based; I've always felt there was some continuity there. It makes sense that the most successful third party combination was Breckinridge (maybe the most successful Southern third party candidate ever) with the earliest possible map (when the South's power was at its apogee).
*Between 1856 and 1956, the best GOP map was almost always 1864, when most of the South didn't vote. After 1956, the best Democratic map has often been 1864.
*Kerry's map in 2004 probably would have been enough to win the vast majority of elections. The post-1960 GOP coalition of states would do quite poorly in most eras (or, on the flip side, the current Democratic coalition of states is poorly optimized for today)
*The Federalists always did best with the earliest possible maps since the West had no representation, whereas the Whigs actually had some appeal in the West.
*It's interesting how often 1812 came up as a "best" map. Not sure what's up with that. It's also interesting that, starting in 2004, the GOP's best map suddenly became either 1896 or 1908.
Re: Kerry, it seems that way but only from a recentish standpoint; that map only wins 38% of elections - it wins every election to 1872 bar 1848, then loses every election to 1964, then wins to 1980, then loses to the present.
Re: 1812, it is interesting - I suspect it is because it is the last election before much of what would become the industrial Midwest started to vote, so it is something of a high point for the relative power of the South compared to the entire rest of the nation. That would explain 1976 at minimum. It is also on the tail end of the highest period of average Dem EC advantage.