For those who want to be made overconfident

How good a guide are the final state-level presidential polls? In 2004, they were mostly fairly good predictors when averaged by state, though with a few exceptions.

TPM just posted a rundown of the close states in 2004 -- comparing what the average polling was (as calculated by Real Clear Politics) just before the elction to what the actual result was.

Those polls got the winner wrong in one state -- Wisconsin, where Bush was expected to win by .9% but actually lost by .4%.

There were a few states where the polls were several points off. Most notably, in Florida, Bush was predicted to win by 0.6% but actually won by 5.0%. Bush outperformed the polling in Missouri (3.0%), Nevada (3.7%) and West Virginia (4.4%). In all of the other battleground states, though, the final polls were within 2% of the actual election results.