Cheney looking better than bush in a VP debate is not a victory for the Republicans. It's pretty obvious now that the ticket is inverted, and that's not going to help them. And based on demeanor alone, Edwards won handily.
A more rigorous scoring of the debate might give the edge to Cheney, but only if it neglected the truth-value of numerous statements -- e.g., " The first time I ever met you was when you walked on the stage tonight", or " I have not suggested there's a connection between Iraq and 9/11", or "the charges (against Halliburton) are false". Darth Cheney proved that he thinks much quicker on his feet than his disciple, but he also played very loose with the facts and for once the newsmedia called him on some of it.
The online polls strongly over-represent Edwards' (and Kerry's) support. This is another reason never to pay serious attention to them -- although the TV news outlets delight in reporting them, since it makes them appear "high tech". A true opinion survey is not self-selected, samples each respondent only once, and weights sampled responses in comparison to area demographics. Online polls really do none of these things, but both parties have found it advantageous from a PR perspective to "get out the vote" on them as they are touted by the newsmedia. This time around, the Democrats did a much better job hitting the online polls, producing a lopsided result.
That said, anyone who doubts that the debates are swinging the momentum back to K/E in the
real polls should take a look at this site, which tracks a composite of all the published state-by-state surveys to find the electoral college total at any given time:
http://www.electoral-vote.com/info/graph.html
From the graph, it looks very much like the Kerry campaign is opening up for the home stretch, and barring an October Surprise courtesy of Karl Rove or any black-box voting shenanigans, K/E stand a better-than-even chance of winning the election. It's going to be a nail-biter, though, no matter your preference.