Now I freely admit that I would prefer to be seeing Romney ahead in each case, but when you call this a huge handicap you are completely ignoring the fact that every one of the underlying polls has a margin of error which average at about 3 points.
When you average several similar polls together, the margin of error for the average is a lot less than 3 points. That's the whole point of poll aggregation, and why the sites that do so (RCP, 538, Princeton Election Consortium) have seen such success with their electoral predictions since 2000.
Why are the Dems confident in OH? Take a look at the graph:
Romney has NEVER lead the average in polling there. Hard to believe that Romney is actually going to win a state in which he trailed in polls consistently throughout the year, and in which Obama has a large early voting advantage.
How big an advantage? Well, let's just ask... Rasmussen:
Forty percent (40%) of likely voters in the Buckeye State have already voted. Obama leads 60% to 37% among these voters.
It's going to be hard for Romney to motivate a +10 advantage on election day, given the focus the Dems have put on the state.
It's true that all polling COULD be wrong, and that the populace is simply more anti-Obama than the polls have picked up. But I wouldn't want to be in the position of hoping that's the case.
Re: the other states, Romney has about zero chance in PA (McCain pulled this same stunt at the end), and it's the big four that matter to Obama - WI, OH, MI, and PA. If he gets those he wins the election, and he's leading in all those states. That = confidence for Dems.