With the scale covered in fecal matter, I dare not venture to give a reading.

http://www.democraticunderground.com/1251133566

because I know you are on top of everything and this is an issue I have read about for at least the last 2 presidential election cycles.

Here's how it goes with RCP - They use cherry picked polls all over the place to show their guy (whoever the R is) in front (unless it is undeniable as it was about 3 weeks ago) until about a day before the election, then they adjust them to what is really going on. That way they can use them to drive voters every 4 years who vote for front runners and at the same time be among those who can claim to be "most accurate". They do this every cycle.

And what are they showing?

I don't know if it's "most accurate" or not, but the past 2 presidential cycles they pretty much called it with a few exceptions.

www.electoral-vote.com

But I rarely see any differences between it and RCP, except how they define tied or tossups.

But part of the problem with trying to remember everything is that sometimes shit slips away. I felt like I knew that RCP was a garbage chute years ago, and haven't thought of them since...so I was just making sure.

When Obama had states with 5 points or more, they did not move them to Leaning Obama, they kept them as undecided.

But why is it they don't include Reuters or Rand?

electoral-vote.com (run by Andrew Tanenbaum, who's an American professor of computer science at Vrije University in the Netherlands) has 272 Obama, 239 Romney, 22 tied for the electoral college.

Nate Silver has Obama 287, Romney 250.

Track record: Nate Silver correctly predicted the electoral vote in 49 states in 2008. Andrew Tenenbaum had the electoral vote probability as Obama 364, McCain 171, 3 tied on the day before the election (the final result was Obama 365, McCain 173).

on the 1 to 10 scale.

Basically (using their formula) so many states are in play or within the margin of error that you can't entirely tell who is going to win. If you subtract all the IFFY states, than romney is ahead marginally but still needs a huge amount(from the iffy column) to win the electoral college.

Incidentally RCP admits that if they HAD to pick a definite winner in each state(using their formulae of averaging polls) Obama wins by a substantial margin of electoral votes.


I think RCP analysis is way off base on which states are truely IFFY.

Using some with old data and throwing out some even if they cover the same days - they have an agenda.

Other than that, their map looks pretty accurate. But, I doubt that Pennsylvania and Michigan, in particular are really much of "tossup".

http://www.realclearpolitics.com/epolls/2012/president/2012_elections_electoral_college_map_no_toss_ups.html

They define a toss-up as a lead of 5% or less (PA is Obama +5.0, and a toss-up; Arizona, Romney +5.3, is not). Which ought to be less, by this stage, which is the main problem (eg Ohio is a 'toss-up', but the early voting has gone enough in favour of Obama that it's next to impossible for Romney to take it).

Decrease the toss-up margin to 4%, and it's Obama 237, Romney 206; 3% is O-243, R-206; 2% (or 1%) O-277, R-235.

http://www.realclearpolitics.com/epolls/2012/president/pa/pennsylvania_romney_vs_obama-1891.html
I think a 5.0 lead is by definition leaning Democrat - that's Pennsylvania.

http://www.realclearpolitics.com/epolls/2012/president/mi/michigan_romney_vs_obama-1811.html
Even the notorious GOP-favorite pollster Rasmussen Reports, which had one of the more current poll results (10/11) has Obama by +7 in Michigan.

So clearly RCP has an issue with defining what is a toss-up.

http://electoral-vote.com/
EV site defines toss-up as being EQUAL and has a slightly category for those within 1-4%. Neither Pennsylvania nor Michigan fall in either of those categories - they are bothing leaning which is greater than 5%. That's 36 Electoral votes right there.