That's actually pretty common for Gallup.

Last presidential election they were tied with Rasmussen when Rasmussen for "Most accurate final number"  hoever were horrible when it came to consistency in their poll numbers.

They likely have an oversampling of undecideds/people who change their mind.

On the one hand... I think Obama will win.  On the otherhand it's hard to discount Rasmussen whose sample has been in 1st place when figuring out who will be President.  It's why there is a very real chance of a popular vote/electoral vote split.

Interesting. I never really payed much attention to gallup in particular. I just saw their numbers shoot up to around +7 at one point and hover around that area, so it caught my attention. We were discussing it in one of my classes, and one suggestion was that their likely voter model might be suspect. I guess they don't make their screening process public. Personally, I don't know much about it, but the poll numbers were odd.

Actually, i double checked that and I was wrong...

it was Pew that was like that last tine.   Sorry.

In 08, Gallup just like the majority of polls over estimated Obama's poll numbers.   (Gallup by +2%.)

Which is really the warning that may come to bear.  Polling tends to skew democrat... it's just rarely close enough to matter.

Still, it makes me think their problem is largely in overepresenting undecideds.

Either way, if you were to base elections based on previous accuracy,  You'd use a mix of Rasmussen, Ruerters and GWU,  and maybe Pew though their consistency is questionable.

You're braver than me. I might make a prediction if I have some free time tonight and can count up some electoral votes. My first impression is that the third party vote won't be anywhere near 6 percent, but who knows?

Why not.

Popular Vote
Barak Obama -  49.6%
Mitt Romney - 48.8%

Electoral Vote

Obama - 286.

Romney - 248.

Though if turoun numbers are wrong that whole script could be flipped.

we'll see but people don't seem to like either Obama or Romney very much, so i think some people might vote 3rd party

I expect turnout to be lower than it should be in some states because voting seems to be made something of an inconvience

http://rt.com/usa/news/florida-vote-early-voting-017/

I'm not confident enough to make a real prediction, but I think it will either be Obama just barely beating Romney or Romney beating Obama by a fairly large margin ...

At the moment I think that of people who are going to vote about 45% are intending on voting for Romney, about 46% are intending on voting for Obama, and about 9% plan on voting for a third party candidate or are truly undecided. I think the two most likely outcomes are for third party voters to maintain their vote and the undecided voters to split pretty evenly which leads to an Obama win, or for the third party vote to disappear into Romney and for undecided voters to split in favour of Romney leading to a significant win for Romney.

Believe it or not, low turnout numbers would probably be good for Obama.

As he's winning Democrats, but losing Independents and Republicans. 

That said... voter turnout is actually a lot higher then you would think.  It's just calculated stupidly.

It's calculated by counting the number of votes vs the number of people of voting age that appeared on the census... ignoring

A)  felons can't vote

B) People with visas and other legal aliens can't vote.

C) Illegal aliens can't vote.

All three are counted in the US Census.