Even TPM isn't immune from received wisdom of course. Nobody is. But I would, honestly, expect something a little better than this:

As SurveyUSA found in a state-by-state electoral college breakdown back in March, Obama could pull it off, at least in theory. But it's not a tried and true path to the Presidency for a Democrat. That doesn't mean it can't happen. But it is a high stakes gamble, the kind that looks brilliant if it works and, well, probably pretty idiotic if it doesn't.

The idea that the "tried and true" Dem path to electoral victory is "pray for OH, FL, and PA" is simply false. Let's look at the history. We have only two Democratic victories in the past six elections (before that, the electoral map was completely different, so it's not a meaningful comparison; for example, CA voted GOP consistently with the exception of Johnson's annihilation of Goldwater). 
In 1992, Clinton won OH and PA, and lost FL. In 1996, he won all three. But what does this tell us? Not much, as it turns out. In both years, he could have lost all three of those states and still won the election, as long as Perot was still in (with 326 and 310 EVs in '92 and '96, respectively). 
If Perot wasn't around and we divide his votes 55-45 for Bush, which is being exceedingly generous to Clinton, Clinton loses the election even if he wins all three of those states, because he loses so much ground in the South, West, and "rust belt". 
What have we learned? OH/PA/FL has never been a "tried-and-true" path to Democratic victory, or defeat (Kerry and Gore could have won by turning any of a number of other single close states, or dozens of combinations of any two small, close states). This is 100% fantasy and TPM should take the lead in dispelling it, not encouraging it.