Mitt's got it wrong -- they're NOT tied.

Mitt Romney's been saying lately that the polls show that he and Obama are tied; that's untrue, or at least he's fooling himself.

His premise relies on any single poll showing him within spitting distance -- the margin of error -- of Obama, mind you, some polls show Obama's lead far exceeding their margin of error.

But the margin of error (dependent upon size of sampling compared to size of population, or in this case, the population of eligible voters / likely voters) naturally shrinks as you aggregate all of the polls.

In the aggregate, the polls show that the margin between Obama and Romney exceed the margin of error of a single poll, let alone the polls in aggregate.  If you look at the trend in polls, it clearly shows that Obama has owned September.  If you look at intrade, Romney's quickly losing support.  As most polls have switched over from registered voters to likely voters, these poll numbers seem to be firming up.  While RCP's data does not reflect the daily changes in tracking polls (only a tracking poll's current day numbers), Rasmussen is the only one that has shown Mitt with any hint of a lead in September; again, Rasmussen has a conservative lean and an unusual methodology that differs strongly from all other polls.

I'm sorry if you're a Romney supporter, but his calendar shows that he's scheduled a loss on November 6th.