Analysis
There's no middle ground; our Presidential elections are either very close or they are landslides. Presidential candidates win by 40 or fewer electoral votes or by a hundred or more.
Note well that Barack Obama can lose. John Kerry was ahead of George W. Bush by about the same amount at the same time in 2004... and Bush squeaked by with quite possibly a little help from his friends in Ohio. That said, the corrupt machine that "delivered" the electoral votes of Ohio to Dubya is no more, and Republicans will have to win Ohio cleanly this time to win. "Kerry 2004" with the addition of Ohio wins the presidency. Although McCain seems a more competent politician than George W. Bush, Obama's campaign seems more competent than that of John Kerry. Even if the Republicans win Ohio honestly, McCain must win back three of five States that Bush won in 2004 that now seem to be in Obama's camp: Iowa, Colorado, Montana, Nevada, and New Mexico. Iowa seems to be out of reach for McCain now.
I could make some comments on Indiana, Missouri, and Virginia: Obama seems unlikely to win any of these unless he wins Ohio.
Many have seen this election as a realignment. Such does not show in the middle of July in 2008; the electoral geometry remains much the same as that of 2000 and 2004 despite an unpopular President and some portents of economic distress not known since the 1930s. In general, Kerry did well in the northeastern quadrant of the United States and on the West Coast in 2004, as did Gore in 2000 -- and both lost the rest of the country. If Obama wins because he takes Colorado, Florida, Missouri, or Ohio, then that indicates more likely a realignment in the politics in the state in question more than in America as a whole. Indiana or Virginia? Much the same.
Much the same could be true in 2008; Obama's strongest gains are likely to be in the northeastern quadrant of the United States should Ohio and possibly Indiana slip into the voting patterns of the northeastern megalopolis, a tendency that began in the 1930s. Virginia would not so much be evidence of any Democratic inroads into the South as the tendency of Virginia to become part of the Northeast.
Addendum: on July 21, 2004 Kerry was close to having a commanding lead over Bush. Kerry even had a slight lead in Arizona -- and there's no way that Obama can win Arizona this year -- as well as in Florida, Iowa, Missouri, and Ohio, all of which Kerry would lose in November. Tennessee was a tie.
After the fact, February 5, 2009:
I should have thought North Carolina a possibility for Obama, and thought that Florida had been drifting toward the GOP. Had I predicted both North Carolina and Florida to go to Obama, then I would have also picked Georgia.
In any event, North Carolina went to Obama by a margin less than those votes cast for Bob Barr, a candidate whose voters were unlikely to vote for Obama, and McCain won Missouri by a margin smaller than the votes for Ralph Nader, few of whose voters would have ever voted for McCain.
It was only one vote, but because Omaha isn't that different from the Quad Cities or Des Moines, I should have reasonably picked it for Obama.