I think Matt Cooper (h/t Andrew Sullivan) has a valid point here:
They (Obama's campaign) are now about arguing math instead of change. The tide is with them, still. But Clinton did a lot last night. Pennsylvania's demographics are like Ohio's and just as she had a popular governor in Ted Strickland helping her the past two weeks, she now has Ed Rendell, more bruised, but still well liked and a force of nature. If she wins there, she certainly has a moral claim on the nomination...
If I were Obama, I'd stop arguing it's over and say, "Okay, let's keep this discussion going." (It's gonna keep going anyway.) You're still the candidate of the past. You still supported NAFTA and voted for the war.
I don't agree with Cooper that Obama should deliberately seek out some opportunity over seas to be seen as "tough". Hillary's vote for the war and dependence on "experience" is enough of a target. But I do agree that while the math does favor Obama, falling back on it as a response to last night's losses is a losing strategy. And Pennsylvania both poses similar problems for Obama as Ohio and is a large enough state, late enough in the process that, whatever his delegate-math advantage, losing there will hurt him a lot. This is especially true because while Clinton cannot catch up to Obama's pledged delegate total during the primary season, neither can Obama win the election outright during the primary season. Last night guaranteed that Hillary will keep going all the way to the convention, and once there, all bets are off.