Megan McArdle has some buyer's remorse about her electoral choice. In an electoral system system with only two viable options, there is always the temptation to view those options on a single dimension spectrum - after all two points define a line, but not a plane, much less a space of higher dimension. This time, for example, Megan looks at the dimension competence vs. "painting a pretty picture for voters." I would argue that, before the actual results are in, the utility of viewing any candidate in American politics through this prism asymptotically approaches zero - or you could just use the simple proxy of "has/doesn't have a campaign strategist."
Moreover, such a strategy - prioritizing one dimension, with maybe a couple more as tie-breakers, without accounting for spreads between positions and for absolute position on an axis - can lock a voter into a pattern of "practical voting" that can turn out very ugly in the long run. Perry de Havilland of Samizdata summarizes it superbly here. An excerpt:
"Guys, you have been voting for the lesser evil for so long you may have lost sight of the fact that you have been voting for evil, just a tiny bit less than the other guy."
PPAA prefers to look at Nolan charts, only of the multi-dimensional kind, along the lines of the ones suggested by the Friesian Institute, and his default position is to not vote - it is up to a candidate to convince him that he or she is good enough, not just better.