The last prediction is in. I have set the turnout figure to a lower bound of 1.5%, for a net of +3.0%. This is a value for which the median and map are readily reconcilable, which may reduce my email load. Readers, thanks for the dialogue. The box and final summary are updated. Other entries are bumped down. And here is your Election Night brochure (PDF).

Finally, a note to all of you. As this site has grown since July, you have seen me try new assumptions on the fly. Most have been improvements, but at times I do something unusual, at which time I hear from many of you. This feedback, although invisible to most of you, has been extremely valuable. In a sense, this site has many open-source qualities. In the event that the prediction is accurate, it was the most amazing of phenomena, an Internet-based collaboration. Thanks for your ideas, your criticisms, your praise, your strange email, and most of all your readership (nearly 80,000 visitors today – site meter).

The last polls for the 2004 calculation come from Granite State (NH), Ohio Poll (OH), Opinion Dynamics (R) (FL, IA, OH, WI), Rasmussen (FL, MI, MN, OH, OR, PA, VA, WA), Strategic Vision (R) (FL, NJ, OH, PA), Survey USA (AR, MO, NC, PA), and Zogby (AR, MO, NH, OR, TN, WA).