this gets the right idea across, but i think what he's ultimately trying to say and not really saying is that polling is only accurate up to a reasonable margin, an error that media cannot correct itself on.
they want to tell you that a 3% bump in the polls is a headline, and you should buy their paper or click their links to learn more about it - which is nonsense. that is noise. and, most of what we've seen is noise.
so, when the polls have three or four candidates - and i think it's three now, in iowa - in a (15,25) percent interval, 95% of the time, they really don't say anything more specific than that there's a 95% chance that each of the candidates will fall in that range.
a 10% range is very large, and so long as each of the candidates end up in that range, the polling will have been correct. people expecting more from the polling than that are really abusing the notation...
further, it's iowa in february. the weather and exams are both huge turnout factors that are very hard to predict with polling - that is what that extra 5% is about.
that said, i think there's three likely narratives.
1) the actual party gets the vote out, and biden wins via machinery politics. i'd guess this is the most likely outcome.
2) everybody knows that sanders has a dedicated base and that they do better in caucuses because they're so dedicated. he has a benefit in what has sometimes been called the "enthusiasm gap". but, i'm a little skeptical about how enthusiastic they still are. i know my own enthusiasm for sanders has waned rather considerably, as i've heard him speak on more topics. it is still possible that sanders' base could dominate, but i actually think that his pivot towards a more mainstream voting base is going to undermine his ability to do well in these kinds of caucuses, and that key aspects of his coalition will feel marginalized and not show up.
3) there are compelling, if shallow, reasons for buttigieg to do well on the floor. age is a real consideration. and, the religious thing can work as a browbeating in a caucus scenario in buttfuck, iowa.
4) i think that warren has fallen out of contention, maybe a little later than i thought, but reasonably, anyways. and, i don't think klobuchar ever really had a serious chance.
but, the sterile, quantitative analysis is that we don't actually know what's going to happen - that either sanders or biden or buttigieg could win, and that the polling will not be "wrong", regardless of the outcome.
if yang or steyer wins, the polling will have been wrong.