First, the states where Obama is playing offense:
Obama Offense StatesThese are all the Bush ('04) states where our projections presently have Obama finishing within at least 5 points of his national popular vote total. The margin indicated in each state is the difference between the projected finish in that state and the national popular vote projection.
New Mexico +5.4
Iowa and New Mexico still appear to be outside the range where they can really be considered to be swing states (although the new Big Ten poll shows a tie in the former). Obama could still lose one or both of these states -- but he is unlikely to lose the election because of them. Those states alone, however, are not enough to get Obama to 270 (or 269) Electoral Votes. He needs at least one more state, even if he holds all the Kerry states.
Colorado remains the best bet; it's projecting eight-tenths of a point better than Obama's national numbers. What that means is that, theoretically at least, Obama would still be expected to have a winning elecotral map if he lost the popular vote by 0.8 points. That is why Colorado is so essential; it is the state most likely to be involved in a split result between the popular vote and the Electoral College.
Ohio and Virginia constitute the next tier, but both are projected slightly behind Obama's national numbers. After that are Nevada and Florida, which are a bit further behind. Then the long-shots: Indiana, Missouri and -- just missing this list -- West Virginia and perhaps Montana.
Now, performing the same exercise for McCain:
McCain Offense StatesAlthough Obama's lead has bent in a couple of these states -- particularly Pennsylvania and New Hampshire -- it has not really broken, as all Kerry states are polling at least 1.6 points ahead of Obama's national average. However, we may be on the verge of seeing Michigan and Pennsylvania flip places, as Obama's polling has generally held up pretty well in Michigan while faltering in Pennsylvania. Wisconsin and Minnesota have also drawn closer, but with 4-5 points of difference still separating them from the national numbers, it is not clear that they are that are likely to form the tipping point in the election.
New Hampshire -1.6