Public opinion polls can be infuriatingly misleading (or the coverage of them can) in several different ways, but here's one that you see quite often. The Chron is reporting that four in ten Americans have had their confidence in the president reduced by his handling of Iraq.

Who cares? Does anyone think that a significant number of those were people who had high confidence in the President to start with? All it means is that low confidence got lower--it has zero electoral implications, because few of them would have voted for him regardless.

I agree with your analysis. Assuming that all 40% of these people vote for the Democratic candidate, then Bush wins 60% to 40%, minus a few points for the third party candidate. I'm pretty sure that's a landslide.