RCP Projections in 2004

This shows how closely the RCP polling averages predicted results in the 2004 Presidential election:

Hawaii was bad because there was little polling and it stopped mid-October. In the rest of these states, the results matched the predictions within 4.5%, with only three states being more than 3% off. (Note, though, that I've used actual results rounded to the nearest full percent, while the predictions are rounded to the tenth. So the differences may be off by up to almost one percent. The orange ones indicate where the polls predicted the wrong guy as winner.)

Looking at these results reminds me how we Democrats thought, in 2004, that Bush and Republicans would surely understand that he hadn't been given a mandate. Out of about 46,000,000 votes in these states, Kerry lost by just 270,000 votes (about 1/2 of one percent).