I'm surprised the NY Times says the race is that close. Even if we throw out the Times' bias and the fact that a bunch of inaccurate/biased polls don't make them more accurate, I disagree with the logic behind the article ...
I have no idea who will win, but "national" polls aren't useful because of the Electoral college (skipping the U.S. history lesson) - I'll wholeheartedly agree that Obama has a 98% chance of winning the popular vote, but the election will come down to a couple "swing states" (Colorado, Florida, Iowa, Michigan, Nevada, New Hampshire, North Carolina, Ohio, Pennsylvania, Virginia and Wisconsin). Romney could lose the popular vote big in California and New York, therefore lose the popular vote, but win enough of the other (smaller) states to win the required 270 Electoral college votes.
I am also extremely confident in predicting that neither candidate will win a "landslide" of the popular vote, but either candidate could win an electoral college landslide (define "landslide" as 60% or more).
For example: 1984 the popular vote was 58.8% Reagan over 40.6% Mondale, but Mondale only won 13 Electoral college votes to Reagan's 525 ("landslide").
In 2008 the popular vote was 52.9% Obama over 45.7% McCain, the Electoral College result was 365 to 173 ... (landslide in Electoral college, but not in the popular vote)
the only thing that is 97.7% probable is that it will be very close. The "election experts" have been saying for a long time that it will come down to voter turnout, which is 100% true