Depends on the sample size for each day, the +/- is a lot larger for a one-day sub-sample in a 4-day poll. The margin of error is probably close to 5 points for a given day's sub-sample (for each candidate). Because of that larger margin of error, the 1-point McCain and the 10-point Obama leads on consecutive days aren't necessarily that much different. And, as others have noted, issues of who is at home on Fridays and Saturdays matter too.
I'm not going to pretend to be optimistic, I'm not (I've not given up either, I just can't buy that every single poll is rigged). It ain't over, largely because it will depend on how undecideds break for McCain.
This isn't like Dole in '96 where at this point we were just hoping that the people would wake up (knowing deep down they probably wouldn't). But it's also an uphill battle. Let's hope and pray our side makes it to the top.
Also unlike Dole in 1996, we don’t have a “Perot” in the race to siphon off 5-8% of the GOP vote.
Additionally, unlike 1996, this is a red state-blue state country now. These national polls just aren’t very important anymore. It all comes down to the state by state battlegrounds.
The path is there for McCain to score an electoral college win but easily lose the popular vote by 4-5 million.