Again, let's keep politics out of this. Betting Romney +280 is a great way to guarantee a cash if you've been hitting Obama at the appropriate time, but polls don't "oversample democrats." That's not how polling and statistics work.
More information about oversampling and polling sample sizes in general can be found in this examination of Rasmussen's historical inaccuracy: [url]https://www.electoral-vote.com/evp2012/Info/rasmussen.html[/url].
A winning bettor can't let beliefs and emotion get in the way of math and logic. If your only exposure is Obama -330, you can eat the juice if you're concerned, of course, but you're still not out of the woods. Remember how much the line jumped back and forth on Election night in 2008. It's in-game betting at its finest.
What's also safe to say is that Obama will not be much bigger of a favorite as he is now unless he leads FL/VA/PA/NH by 5 points each or more when they start announcing numbers. Then CO/IA/WI don't matter.
Here is another possibility. Early voting tends to favor the Democratic party. If they load these numbers in and then start adding in those who voted on Tuesday, Obama is going to have a big early lead in states such as FL and VA. People are going to totally overreact to this, and you may find Romney at +400 or higher around 6 p.m. Pacific.
Just limit your exposure so you can't get hurt and try to anticipate what will happen. I don't know how individual states do their votes (although I know in Ohio that big cities will update every 15 minutes, midsize areas every 30, and rural every hour). In California, SoCal reports sooner, so the results tend to start out with more conservative votes, although it's not so relevant here.
Let's cash some tickets!