Which Atlas members, deserve the biggest "I told you so!"

[afleitch]

So far, for me, (FL is out)  VA, IA, NH, OH, CO.

So basically you got every single swing state wrong. I'm very disappointed in you, J.J. I thought you were a political juggernaut on this forum. But I'm beginning to suspect you really don't know the know the first thing about politics.

That was obvious when he claimed that a 60,000 early vote lead for Obama in Iowa was a sign of Romney victory.

I wasn't the only one.  Dave Wasserman made the same mistake.

Admit it J.J, you got it wrong; the polls, the swing states, the demographics, the turnout. Everything.

[afleitch]

J.J., I don't think you know what a margin of error is and how it is calculated.

Nor can you really base how well polls did on the RCP average rather than what the actual result was, especially as RCP were picky as to what pollsters they included in their average.

So how did the pollsters do? Let's look at the predicited margin of victory for each candidate and look at the actual difference

A positive difference meant the result was too favourable for Obama and a negative result meant the result was too unfavourable for Obama

Colorado -  O+4.7.
PPP + 1.3 difference
RAS -7.7 difference

Florida - O+0.6
PPP  +0.4 difference
RAS -2.6 difference

Iowa - O+5.6
PPP - 3.6 difference
RAS -6.6 difference

Nevada - O+6.6
PPP - 2.6 difference
RAS - 4.6 difference

New Hampshire -  O+5.8
PPP -3.8 difference
RAS - 3.8 difference

North Carolina - R+2.2
PPP +2.2 difference
RAS +3.8 difference

Ohio - O+1.9
PPP +3.1 difference
RAS -1.9 difference

Virginia - O+3
PPP +1 difference
RAS -5 difference

Wisconsin - O+6.5
PPP -3.5 difference
RAS -6.5 difference

The Average of the above:

PPP -0.61% difference
RAS -3.88% difference

Rasmussen was way off across the board. PPP's house 'bias' was actually unfavourable to the President. Rasmussen only bested PPP in Ohio, which still has 200k votes to count and called it wrong. It called Colorado, Florida, Iowa, Ohio, Virginia and Wisconsin wrong. Six key states. Even in North Carolina which it called right, it's error was still bigger than PPP's.

Rasmussen got the swing state calls poorly in 2008. It had one of the worst track records in the 2010 Senate polls and it probably has the worst record for any prolific, multi-state poster in 2012. There is, therefore, good reason to ignore them from now on.