Tom has an interesting post today on Iowa, referring to Jay Carney's piece in Time. But it's worth reviewing what the "second-choice" issue in Iowa is really about and what the polls do and don't reveal.

It's true that one poll shows Edwards tops the list as the leading second-choice of potential caucus participants, which could pay off because of the 15% threshold cutoff. But leading a second-choice list of all caucus participants doesn't mean much because it's likely that -- barring a major collapse of a major candidate -- Clinton, Obama, and Edwards will all meet the 15% threshold virtually everywhere. The key question is who will be the second-choice candidate of the Richardson, Biden, Dodd, and Kucinich supporters on January 3rd. And the likely answer is, who knows? One could certainly make an argument that Richardson, Biden, and Dodd voters have already rejected Hillary and will do so again, if they can't support their first choice. But maybe they'll bind together and go uncommitted. Maybe these voters are more establishment-oriented and Hillary will appeal to them. Or maybe, the most important thing to them will be to go to a corner of the room where they have the most friends.

The point is this issue is virtually unpollable. And, to the extent it is, polling the whole Iowa universe -- rather than the supporters of the likely losers -- doesn't tell you anything.

Besides, someone will likely do an entrance poll that night, just like in 2004. In the internet age, it's that result that will get reported and disseminated to the nation. At this point, it's all about momentum, not delegates. The second-choicers won't even matter.

To read Steven Stark's complete "Presidential Tote Board" blog, go to www.thephoenix.com/toteboard/