Theory about New Hampshire: The polls will be wrong in a way that benefits somebody that did not win the Iowa caucus. In 2008, polls were off by 11 points underestimating Clinton. In 2016, polls were off by 9 points underestimating Sanders. Will the same hold true in 2020?
It's probably a hot garbage idea but who knows.
Really depends on what you mean by "winning", especially considering the circumstances around the primary. I would not say that Sanders lost, for example. Anyway, that basically leaves Warren and Biden, both of whom I do not see doing especially well in New Hampshire.
I think Pete is more likely the one who got the perceived victory, the media focused on the SDE count rather than the popular vote and he got way more of a bump in the polls than Sanders did. So if the theory is right, then I would think Sanders would benefit (as in, the polls would underestimate Sanders).