Rubio will win the New Hampshire primary handily. How to read the polls.

If you believed the media, Rubio was on a "surge" in Iowa, but the "polls" missed the extent of it. Factually not true. The one poll held out universally by media as the "gold standard" of Iowa polls, the Des Moines Register poll, did blow it, but there were polls that did not.  One poll, the Emerson College poll, completely nailed Rubio's level of support in Iowa.  They pegged him at 22 percent the weekend prior to the caucuses, and his finally tally was 23 percent.

So how can you accurately evaluate Trump's coming New Hampshire "implosion" to a solid second place finish, and use the polls to accurately predict Rubio's level of winning support in New Hampshire by primary day?

Use a site like RealClearPolitics to look at all the polls, not just the one or two always quoted by commercial media.  Look for two things.  Look for the polls that rate Rubio's level of support the highest.  Then add up the totals for all candidates.  When you do this, you will notice that the totals won't add up to 100 percent.  Subtract the smaller number from 100 and you will get the percentage of those who were "undecided" or refused to express a preference for that poll.  Most polls no longer expressly list "undecided" so you will have to perform this very simple bit of arithmetic to find out the "undecideds."

Then pick the top couple of polls that have have both Rubio at the highest level of support, and also have the lowest percent of "undecided."  Then average (separately) the numbers for Rubio, Trump, and the undecideds from those polls together. You will have three averages.  Take the averaged percentage number for the undecideds, and multiply it by .8 (80 percent).  Then add that number to Rubio's averaged total. After doing that, subtract at least 5 points from Trump's averaged total, and give half of that number to Rubio.

Let's apply this formula to the UMass tracking poll, which, on it's face, has Trump at 36 percent, and Rubio at 15 percent, the highest raw total for Rubio of any poll listed on RealClearPolitics. This poll has 8 percent (after adding candidate support totals and subtracting the total from 100) expressing "no choice."  If you run the formula, Trump settles at 31 percent, and Rubio's total rises to 24 percent (23.9 percent precisely).  Most polls have a built in 3 to 4 point margin of error anyway, so you can see that already, with several days out from the primary, Trump and Rubio are within just a few points of each other if you consider a margin of error weighted up for Rubio and down for Trump, and Trump's will only go down, and Rubio's will only go up.

That will get you about as close as you can get to knowing ahead of time what will actually happen when the "sheep" line up on Tuesday to get sheared.