[Updated] Democratic Nominee WILL be decided tomorrow. Here's Why

Everyone here and elsewhere is suggesting that the democratic nomination election will likely continue on after super tuesday.  I find that to be a highly unlikely scenario.  How do I come to this conclusion?  Call it the "undecided break" theory.  

It's a theory that gets talked about in the general, but for some reason no one is talking about during the primary.  This is in spite of the fact that we've seen this phenomonen play out in the early primaries.  The theory basically goes that at the last minute, the undecided voters will break disproportionately for one candidate or the other.

We saw it happen in Hew Hampshire where they broke for Hillary.

We saw it happen in South Carolina and Florida where they broke for Obama.

We saw it happen in Michigan where they broke for "uncommited".

I'm leaving out the caucus states since "viability" tends to distort the results.  However, if you look there is some evidence that it may have happened.

Join me on the flip for the number proof.

Check out New Hampshire's pollingnumbers.  The last 5-poll average was

Obama: 38.4    Clinton: 31.0    Edwards:18.2

Now let's look at the actual results

Obama: 37      Clinton: 39    Edwards: 17

As you can see the pollster essentially nailed the Obama and Edwards numbers.  What happened was that the remaining undecideds broke disproportionately for Hillary.  Let's look at one where they broke for Obama in South Carolina

Check out South Carolina's polling numbers.  The last 2 polls released were

Obama: 41\44    Clinton: 26\24    Edwards: 19\19

The actual results?

Obama: 55    Clinton: 27    Edwards: 18

Once again, the pollsters essentially nailed the numbers for 2 out of the 3 candidates.  So what happened once again was that the undecided voters in the polls broke disproportionately for Obama.

Now you might argue that these results won't play out tomorrow because the campaigns did more "retail" politics in those 2 states.  So what happens when there are no campaigning in the state.  Thanks to Florida, we can see that it happened again - just to a lesser extent

If we exclude ARG polls(which suck) and look at the last 4 polls done in the state, here are the ranges.

Obama: 25-29    Clinton: 47-52    Edwards: 12-16

The actual results:

Obama:  33    Clinton 50    Edwards: 14

So what we see is that the actual results for Clinton and Edwards were in the median range of the polls, Obama overperformed by a few points.  Presumably, again, because the few remaining undecideds broke disproportionately for Obama.

Just for fun, let's look at Michigan's primary polling range.

Clinton:44-56    Uncommited: 26-33

Actual Results:

Clinton: 55    Uncommited: 40

The polls on the upper range did get it right, but no one was even close on getting the "uncommited" number.  Again, the undecideds broke disproportionately one way.

What does this mean for tomorrow?  Look at all of the state by state polls.  Any state that has a huge "Keith" number(undecided plus margin of error), watch out!  Those undecideds will very likely break disproportionately to one of the 2 candidates.  That break will likely be the same across the country and cause an "unexpectedly" large win for whichever candidate gets it.  The one thing I won't guess is which candidate they end up breaking for.  There is evidence that it could go either way depending on which poll you're looking at on which day.

So tomorrow, as the results start pouring in, check that state's polling vs the actual results.  If the polling is correct for one candidate, but way under for the other, then you'll know early which way those undecideds are breaking.

UPDATE:
The most common thing I'm hearing in the comments is that the undecideds may break differently state-by state.  Or even region by region.  I admit that is possible, but I don't think it's likely from the scant evidence we have.

The last primary that was this close was the republican one in 2000.  On their "Super Tuesday" the undecideds broke very heavily for George W and knocked McCain out in what was supposed to be a very close match.  Those states that broke for Dubya were spread across the country.

I don't have any earlier data to confirm this happening in a primary.  However, this ALWAYS happens in the general election.  That is why I'm surprised no one is applying the theory to this so-called "national" primary.