Forget everything you think you know about Obama's appeal and focus on electoral geography. He may be wildly popular, but some of his visibly enthusiastic support is concentrated in a few areas (Illinois, California, etc) and amounts to little more than ensuring that he will win those states in a landslide as opposed to a mere 10-point margin. He is also very popular among certain voters in states he won't win (i.e., among black voters in Georgia or college kids in Texas). That does not mean he can't win. It means that the broader perception of his support isn't relevant in a presidential election. What matters is: in how many states can he cobble together enough supporters to constitute a plurality? Hitting 70% in Illinois or losing 52-48 (instead of 60-40) in Georgia doesn't really matter.
He can win. It won't be easy in this environment. One or more of the following will need to happen:
1. Black turnout rates approximate white turnout. In 2000, survey research indicates that 93% of black voters voted for Al Gore (what would it be for Obama, 99?) The problem is that black turnout severely lags other racial groups. John Kerry gave a yeoman's effort to turn out urban black voters in 2004, namely in Cleveland, and to some extent he succeeded. But without being crass, let's say Obama has some advantages over Kerry in motivating black turnout.
2. Young people vote. 18-to-24s are by far the worst at turning out. The demographic seems really excited about Obama but historically they talk a big game and fail to back it up. They could be determinative in tight states. I'd spend some serious money on registration and election day door-knocking on the Ohio State, Wright State, Akron, Oberlin, and Cleveland State campuses just to name a few.
3. McCain's comedy of errors continues, destroying his credibility among all but Republican diehards. This doesn't seem to be working so far, but there's a lot of time left and he's getting worse by the minute. There could be, but certainly doesn't have to be, a "Dukakis in Tank" or "There is no Soviet domination of Eastern Europe" moment that ends it.
4. Obama hits on an issue that speaks to the "all important" angry blue-collar white people demographic, keeping them from being enticed by McCain's promise of 186291 additional tax cuts. He has yet to do so. Come up with a better answer than "taxes" for the only question these people care about: Why am I broke?
5. Hold every state Kerry won in 2004. This is absolutely essential, although Obama can (and probably will) drop a small one like New Hampshire. Too many competitive states are too close for Obama to surrender anything Kerry won and figure "I'll make it up elsewhere." This means holding Pennsylvania, which was the closest state in 2004 (not Ohio. Really.)
6. Fuck Florida. It's gone. Hillary is popular among Floridians, so maybe she can help there. But mentally tabulate it as a loss and be pleasantly surprised if it comes through. Do not base a strategy on taking Florida.
7. Spread McCain thin. He is going to be outspent, the first time in….ever….that the Democrat will raise more money. Do not make the Kerry error of competing only in the 10 states that are in play. Spend money in places like Nebraska, Indiana, Alaska, or Georgia. Obama won't win any of those, but if he can get McCain paranoid it's worth it. This worked phenomenally well in 2006. Creating the impression that people like Scott Kleeb could win House seats in Nebraska A) panicked the GOP and B) created the image of Republicans In Trouble, i.e. "Gee, if they're losing Nebraska they must really be in the shitter." Kleeb lost, of course. Doesn't matter though. It made a point. Obama has already started doing this by claiming that Virginia is in play, which it probably isn't.
8. Win one of Wisconsin or Iowa. Even with Ohio, Obama will need one of these to win (assuming McCain takes NH and Obama takes New Mexico, both of which are likely). They are crucial. Turning out Madison and Milwaukee can seal the deal for Obama.
9. Prepare for losing Ohio. Don't put all of the eggs in the Ohio/Florida basket. If McCain takes OH and FL, Obama can still win with Iowa, Wisconsin, New Mexico (all reasonably likely), Colorado (depends on who turns out), and Montana (Democratic governor, state legislature, and Senators).
10. Thank God for the House majority. If Obama wins all the Kerry states (minus NH) and McCain wins all the Bush states except NM, CO, IA, and WI, the election is 269-269. This is the most plausible tie scenario we've seen in ages. If McCain holds Ohio somehow, the specter of a tie is real. The election could swing on a tiny state like Montana or New Hampshire. Why do we have an even number of electoral votes again? Oh yeah, because the EC is retarded.
This is a very hard race to gauge at this point, and it likely will remain that way until October. One gets the feeling that Obama could cross some tipping point where this becomes a laugher and he wins a Reaganesque 40 state blowout. But right now it feels more like 2004, where we're doing a lot of mental calculation of how a victory can be scraped out vote-by-vote.