I'm in beautiful Nashville, but (for people like me) here's some not so beautiful news:

The best way to track a change in the polls is to look for instances in which the same firm has surveyed the same state (or the national race) multiple times. The FiveThirtyEight forecast model relies on a procedure very much like this to calculate the overall trend in the race. 

Fifteen polls were released on Friday that provided a comparison with another survey conducted between the Democratic convention and last week's debate in Denver. Mr. Romney gained an average of 4.6 percentage points in these surveys...

It is not clear whether Mr. Romney is still gaining ground -- or whether he benefited from a couple of outlying results. The median change in the polls, which will be less sensitive to potential outliers, was a three-point gain for Mr. Romney, more like earlier in the week. 

But unlike earlier, Mr. Romney is now seeing some of his best results in swing state polls. Six of the seven polls published on Friday from such states had him ahead.
I think Joe Biden did really great the other night. But his task wasn't really to convince the sliver of people on the fence, but to reassure the base that there still was a fight afoot. The opportunities afforded a challenger on a debate stage are unique to the position. The door does not swing both ways. 

One last point from Nate Silver:
Although we prefer to describe the race in quantitative rather than qualitative terms, the nomenclature that we use in our Senate forecasts is to describe a race as a "tossup" if each candidate has at least a 40 percent chance of winning. Mr. Romney is on the verge of that threshold.

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