In terms of both style and arguments, this one was much more even, with Bush probably ahead in the closing minutes, which is the part that sticks in people's heads (if they've sat through the whole thing). But in terms of both momentum and the expectations game, a draw is effectively a win for Bush. Kerry needed a decisive victory; we'll probably see Dubya regain at least part of the lead he'd lost after the first debate.
Postscript: Cato's David Boaz offers another perspective:
If Andrew Sullivan is right, which he might be, that people don't want to reelect Bush and are waiting to see if Kerry is ok, then I think a draw is a win for Kerry.
Maybe. I'm inclined to stick with the first interpretation because I figure that bounce—the sour-on-Bush swing voters concluding that Kerry was good enough to vote for—came from the first debate. Assuming those guys are the people in play, a performance that's relatively better for Bush or worse for Kerry should cut the other way.