Two weeks before Election Day, all signs point to this being a very tight election.  Romney clearly seized the momentum with his debate win two weeks ago – one which Obama failed to stop with his stronger performance last week.  Most national polls show Romney with a 2-3 point lead; however, the state polls show an even tighter race.

One thing has not changed in terms of the Electoral College; the election will still boil down to Ohio, Ohio, Ohio.  However, there is one major development over the past two weeks that has strengthened Romney’s hand in the Electoral College.  The national surge in support for Romney has created such strong momentum in Florida, Virginia, and Colorado – both in the top line numbers and internal numbers – that it’s hard to see him losing any of those states.

So who cares?  Well, once we allow for the assumption that Romney wins those three states, it is absolutely impossible – not just improbable – for Obama to win the election without Ohio.  Even if he were to run the table in the rest of the battleground states (NH, IA, NV, and WI), he would still come up short.  Take a look at how that would work.

Perforce, Obama cannot win without Ohio.

On the other hand, although it is still unlikely that Romney will win without Ohio, he is beginning to open up a legitimate alternative to 270.  Many polls show Romney leading in New Hampshire, a reflection of his surge in support from white voters.  Moreover, he has the momentum in Wisconsin and Iowa.  Unfortunately, he appears to have stalled out in Nevada, polling about 2-3 points behind Obama.  Any realistic alternative to 270 bypassing Ohio must include a victory in Wisconsin.  Once he wins Wisconsin (from his base of 257), he has 267 votes, and needs to win either N.H.

or Iowa.

Of course, this is predicated on the assumption that Obama keeps Nevada, which is a likely result in the event that he wins Wisconsin and Iowa.

To that end, it’s no surprise that Politico is reporting that Obama’s final strategy heading into November is to hold 3 states; Ohio, Wisconsin, and Nevada.  Obama absolutely must win Ohio in order to get to 270 EVs with the current state of the race.  But that alone is not enough.  He must win Wisconsin to shut off Romney’s other pathway to victory.  Additionally, he must ensure that there is no momentum change in Nevada.  If Obama loses Nevada, then Romney could still win with New Hampshire and Iowa, or tie him at 269 by winning Iowa, in which case, Romney would be selected president by the Republican House.

From Romney’s vantage point, his most important states heading into November will be Ohio (of course), Wisconsin, Iowa, and New Hampshire.  It would be nice if he could move the ball a bit in Nevada and really make Obama fight for Pennsylvania.  Then again, in such a scenario, Romney would have already locked up the requisite 270 votes (Pennsylvania will never be the state that brings a Republican over the 270 mark).

So whose shoes would you rather fill at this point?  It’s really a jump ball at this point.  On the one hand, Romney has a marginal, yet legitimate, pathway to victory without Ohio, while Obama doesn’t.  On the other hand, Obama officially holds a slight lead in Ohio in most polls, and Romney has yet to turn over the final few points in Wisconsin.

What about Romney’s slim but steady lead in most national polls?  There is a legitimate case to be made that Romney can win the popular vote by 2 points due to historic margins in the south and a nice showing in Pennsylvania and Michigan, yet still lose Ohio and some of the swing states by the slimmest of margins due to Obama’s effective ground game, early voting shenanigans, and an oversaturation in TV ads.

However, there are two big wild cards that are not reflected in the aggregate of the top line numbers from the polls, but will favor Romney in the event that the final polls show this to be a 50/50 race.  Those two X factors are turnout models and undecided voters.

Most of the polling data in the field reflects samples that are similar to the 2008 turnout model.  In fact, the latest PPP poll which has Obama leading by just 1 in Ohio, reflects a turnout model that is more Democrat than 2008.  If you drill down into the internal numbers of the post-debate polls, you will find that Romney is meeting or surpassing Bush’s 2004 benchmarks for Independents, white voters, suburban voters, and base voters (vs. Obama’s base voters).  More recently, he is dramatically closing the gender gap.

It is clear that any turnout model that is slightly shy of the 2008 Democrat model will result in a Romney victory. And if we see something close to a 2004 turnout model, we might be talking about Romney running the table (something that might be reflected in Gallup’s national poll which shows Romney up 6).  Everyone now agrees that there is a significant enthusiasm gap favoring the GOP.  It’s hard to justify any prediction model that uses 2008 turnout data.  That was during a year when Democrats swamped Republicans in enthusiasm.

The other X factor is the split of undecided voters.  Any 50/50 projection based upon the current polling is predicated on the assumption that undecided voters will split evenly on Election Day.  However, with Romney surging ahead in likeability and crushing Obama on the economy, in conjunction with the fact that Obama is well below 50% in all the key states, even though he is such a known quantity, it’s hard to see an even split.  In fact, it’s very likely that we see a huge break towards Romney is a way that will totally remake the electoral map.  The election is about the economy; the economy stinks, and Romney is crushing Obama on the economy.

The bottom line is that if we only look at the polling data, Romney holds the edge on the popular vote, but the Electoral College is a dog fight.  However, if we look beyond the polls to factor in an alternative turnout model and a strong backlash of undecided voters against the incumbent, Romney will win by a sizable margin.

Cross-posted from The Madison Project