The polls point to...

...Mitt still losing the election.  Yes, really.

• That single, national Pew Poll showing a giant swing to Mitt, was largely because they switched their weighting from predominantly identified Democrats, to predominantly identified Republicans.  The only reason why this poll is worth singling out, is because the other polls show that their movement is well within their margins of error and the trailing month's movements.  If you understand that Pew moved its weighting, combined with the fact that the other polls still are within their own margins of errors, then the correct conclusion is that the Pew's number is an outlier.  And for what it's worth, the polls are likely to swing back to Obama after taking in a full week after the Friday jobs report.
• InTrade still shows Obama with a 61% chance of winning, while conversely Mitt's chances are pegged at 38%.  For Mitt, that's still six percentage points below his high, from late August, at 44%.  Note again, that's still below 50% odds.
• The formula for Mitt still hasn't changed; he still needs to win out in all eight swing states, and he only has a clear lead (in the same states that he's generally held a lead in) in two of them -- Missouri and North Carolina.

Now, lest you think that my analysis is biased, I can point to any number of aggregate poll sites that show the same thing, and to the NYT's fivethirtyeight blog that covers analysis of both state-level and national-level odds.  Currently, Nate Silver has given Mitt a 25.2% chance of winning the electoral college -- yes, as I said it's that hard for Mitt to win, no matter what the national poll now says.

However, if you really don't believe me, just look at Karl Rove's latest electoral college map.  He's already marked Ohio and Nevada as Obama-leans, which means that there is NO WAY that Mitt can win.  If you don't believe the data from a Republican pundit and operative who directs two conservative superPACs, then you're hopelessly lost.

It is worth noting however, that for the first time in this contest, Mitt holds a small lead in RCP's poll aggregation.

It does not affect the outcome of the race at this point however, because it does not reflect the swing-state problem for Mitt.  But if the election were held today, it could result in Mitt winning the popular election but losing big in the electoral college -- worth paying attention to, because a split would give Republicans a rhetorical reason (a majority national support) to continue to resist compromise on important legislation.

This goes to show the importance of winning a majority of the national vote in addition to the electoral college.

---

Update: NBC News has New Hampshire and Wisconsin as swing states; HuffPo does not; neither does Karl Rove nor RCP.  Regardless of which states are in play as swing states, the math doesn't change -- Mitt's still down in the polls (post-debate) of all of the states that he must absolutely must win, and has been for some time.