It's tempting to think of two separate contests a week from today.  In reality, there is only one contest--the quest for the Democratic nomination.  And on May 6, there will be a loser and there will be a winner.

How will we know who won or lost?  Well, to figure that out, we need to define 'victory'--something McLame et al still haven't managed regarding Iraq.

A guide by the numbers to picking a winner and loser on Super Tuesday III below the fold.

First, let's get this out of the way:  in a strategic sense, Obama just needs to avoid a complete meltdown to secure a strategic win on May 6.  His lead in the popular vote and the pledged delegate tally is nearly insurmountable, and will become even moreso after May 6.  That Obama will win the pledged delegate race is a metaphysical certainty.

Second, let's note that if either candidate loses both races--if Obama loses NC or if Clinton loses IN--the consequences would be disastrous.

So, victory here means relative success in terms of meda/spin, tactical impact, and immediacy.

So, where do we set the bar?  Simply put, we crunch the demographics according to current trends.  If a candidate performs above what current demographic trends predict, that candidate wins.  If that candidate loses ground, that candidate loses the night.

So, let's do the math:

Indiana

SurveyUSA,which is pretty good at this kind of thing, puts the likely voter model in IN at:

87% white
10% black
3% other.

For sake of simplicity, let's make that 89% white and 11% black.

Over the past few months or so, the trend has been for Clinton to win white voters 60-40 and for Obama to win black voters 90-10.  Since this is the conventional wisdom, this is where the bar is set for winning or losing.  If Obama only loses white voters by 10%, he's made significant progress in that demographic.  If Clinton only loses black voters by 30%, same story applies.

So, if the current trends hold, we should expect to see:

Clinton 54.5 ([.89*.6] + [.11*.1])
Obama 45.5 [.89*.4] + [.11*.9)

In terms of raw votes, it seems fair to assume that since IN has half the population of PA, that turnout will be about half of what it was in PA--1.15 Million.

Crunching the numbers a little further: gives us:

Clinton:  626,750
Obama:    523,250

In other words, Clinton + 103,500.

N.B.  It also bears noting, for those eager to proclaim that Indiana is on the border with Obama's home state of Illinois, that Connecticut proved that demographics trump geography.

North Carolina

We have two fairly consistent voter turnout models for North Carolina:  PPP and SUSA.

Averaging them out, we get:

White 67%
Black 33%

Running the conventional wisdom/current trends through this prism, we get:

Obama:    56.50 ([.67*.4] + [.33*.9])
Clinton:  43.50 ([.67*.6] + [.33 *.1])

Again using PA as a model for total turnout, we arrive at 1.7 Million voters.  In terms of raw votes, that comes out to:

Obama:    960,500
Clinton:  739,500

Total:  Obama +221,000

Overall, for the night, the raw vote differential should be:

Obama+108,000

Of course, none of the underlying assumptions are 100% accurate, but this should provide a fairly close guide to who has gained ground in what demographics, and how the media will play this.

Now, if Obama supporters want to help him win May 6th, you can help Obama CLOSE THE GAP in Indiana by calling or even by making a personal appearance in the Hoosier state.