Andrew Tanenbaum's http://www.electoral-vote.com website tries to predict the outcome based on awarding each state's electoral votes to whoever is leading in the most recent statewide polls. As of about 1PM today, it's "predicting" K/E 262, B/C 261, and 15 undecided EVs in New Jersey. But, as Tanenbaum notes, the "most recent" polls in Wisconsin (Bush by 3%) and Ohio (Bush by 3%) are from FOX News; yesterday's numbers were Kerry by 7% in Ohio (from Gallup) and in Wisconsin (from Zogby). How likely is a ten-point swing? Similarly, the "tie" in NJ is coming from Strategic Vision (a Republican house poller), while Quinnipiac had Kerry by 5% the day before and Eagleton-Rutgers had Kerry by 4% the day before that.

My personal guess, right now, is that Kerry will win the election. Taking Tanenbaum's numbers and moving NJ, OH and WI to Kerry, the outcome would be Kerry/Edwards winning 312 to 226.

My hope is that he'll win 270+ EVs by more than 3% per state, which should put paid to any claims that the election was stolen.

(By the way, Tanenbaum's site has apparently being the subject of repeated DoS attacks recently. He's got mirrors at wwww.electoral-vote2.com through www.electoral-vote8.com; he's also been using the attack data to help a couple of the other professors at his university with their research on coping with flash crowds... )