I don't usually try and prognosticate in which direction the tracking polls are liable to move, but it looks to me like the debate may have been a pretty significant help for Barack Obama.
The Zogby-Reuters poll is already out. From reading Zogby's write-up, it sounds like Obama must have had about a 9 point lead in Thursday's interviewing and, oh, a 5-6 point lead in Wednesday's interviewing. This compared with a 2-point lead over the Sunday-Tuesday window, before the debate occurred. (Zogby's party ID weightings are screwy, but we'll discuss that at another time; the poll should be fine for inferring trendlines).
Research 2000 had Obama winning Wednesday's daily sample by 12 points, as compared with 9 on Monday and 8 on Tuesday.
Or take a look at the most recent Rasmussen polling. They put five polls out tonight: Obama by 3 in Florida, Obama by 1 in North Carolina, Obama by 16 in Michigan, Obama by 8 in New Jersey, and McCain by 7 in Indiana. From among that set of polling, the Indiana result is poor for Obama, and the New Jersey result is a little below expectations, but the North Carolina and Florida numbers are pretty good and the Michigan number is very good. Obama's two worst polls from that group -- Indiana and New Jersey -- were conducted on Tuesday, principally before the debate. The other three, which were stronger for him, were conducted on Wednesday, after the debate.
So, I don't know. Take this with a grain of salt. But there are some hints that Obama may be on track to post some of his strongest numbers yet as we head into the weekend.