(edit: Oct 28, 2004 - Okay maybe not in a landslide. That was a bit over-anxious. He will win though, and it will be convincing, if not a "mandate.")
First, I want to make a public commitment. I will only mention politics once a week. Politics is not that important in my life, really. That said, time to flinch:
I can now predict that George W. Bush will win the election by a landslide. This is non-scientific, and you can laugh (even when I am right) but I am now certain of this. Here are my reasons:
1) Kerry is behind in most polls, even though the media is claiming him the winner in all three debates. Personally I think Kerry won big in the first debate and made good showings in the other two. Bush's scowl in the first debate hurt him to no end, but I think he did well in the second two debates. If Kerry were going to make his move on the debates, it would be evident already. He hasn't passed Bush, although the debates got him back into the game. I don't believe in "momentum." That is no more valid a concept in politics than it is in sports.
2) Kerry got zero bump from his convention, while Bush got a significant bump. This is a big deal. It shows that those people who watched the convetion--which is essentially an infomercial on the candidate--already were in his camp. Kerry should have benifitted from some bump, even a few points. My theory on this is that the neo-activist Bush-haters were already with him and were not shy about saying so. If you are a Bush-hater, you are not undecided. So the "bump" of Kerry's convention happened much earlier in the year instead of at the convention. And after all that, Kerry was still way behind Bush. Bush is still changing minds in his favor, and his bump shows. And Bush was ahead of Kerry even before his own convention, after which he got a further bump.
BTW, here is the blog of my favorite Bush-hater: Fog Blog. Read his stuff. He is interesting even when he is wildly off-base (well, one of us has to be and it can't be me).
3) Bush is the incumbent. This election is essentially a recall election on his presidency. This means that as people think about whether to vote for his presidency or not, he will essentially get most real ties (indecisive people that make their final decision in the ballot box, which is where I usually make my decisions). Kerry has to change minds, and if he fails to do that he fails to get the true undecided. Kerry is failing to do that enough to matter in that voters box. The media is pushing the story of how "close" the election will be, but that is what the media does.
4) The bases have been set for months, and there will be little gain from either party from "getting out the vote" of their base voters that didn't vote in the last election. I could be wrong here, and if I am it will be in the realms of the Christian right (for Bush) and the single woman voter (for Kerry), but I think that any movement from the "flaky" sides of each candidate's base will cancel out the movement from the opponent. So I remove this factor out as a constant.
5) America will not have another close election. The people don't want it, and it will show in their vote. The media can't help but play up the reoccurrence of the Florida election scandal of last election. This has caused both parties to mobilize to eat up the attention the media is giving to this story. Jesse Jackson, of course, will be sitting in Florida threatening lawsuits on election day. Lawyers will be on the starting line in all states where they think they have an issue to fight over (which will probably be all the close states). So while there will be a plethora of lawsuits the day after the election, and voter fraud etc. will be in the news for a few weeks, it won't matter.
For these reasons, I see Bush winning by a landslide.