Political analysis is often a tug of war between the deep limbic, animal regions of the brain on one day and the cortical areas on the next.*

Obama's had a tough week, as numbers everywhere reflect, including the Gallup daily track (but not the Rasmussen track) and three state polls showing Clinton outperforming Obama against McCain in Ohio, Missouri and even making Kentucky somewhat competitive.

But the Obama campaign has met the challenge of Rev. Wright, perhaps sufficiently, perhaps not. But from the perspective of wavering superdelegates, it's hard to find a level of panic among them. Obama has four weeks to recover until Pennsylvania; assuming that the bad news evens out the good news, the attitudes of these superdelegates will degrees to the mean and they won't be a position to rethink the entire premise of Obama's candidacy.

The see-saw numbers change every week; why would superdelegates put more stock in them now versus last week versus two weeks from now?

And Florida and Michigan aren't going to re-vote. The chances for their Jan. 29 delegations to be seated intact are slim to shred.

Clinton may well win the political argument in Florida and Michigan, but she won't win a single delegate from those two states until at least April, when the DNC's rules and bylaws committee might decide to seat superdelegates based on the appeal of DNC member Jon Ausman.

To put it crudely, the analyst's emotional brain feels momentum for Hillary; the analyst's analytical brain can't quite figure out how Obama loses.

* Yes, cognitive neuroscientists, I am vastly oversimplifying the brain.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.