If you look at the 2016 election you can find pundits, pollsters and the general public who predict everything from a Hillary Clinton blowout to a Donald Trump blow out (and everything in between).
People on the right, many of them, believe that the polls have oversampled Democrats. That the turnout of Democrats is going to be depressed among African American voters, Bernie Sanders supporters and even to a certain extent Hispanic voters are unenthusiastic and won't turn out in numbers to get Hillary Clinton elected. Meanwhile, Trump has surged among men, blue collar workers, and a number of other traditional Democrat demographics. In addition, they cite things like the pollster misreads of BREXIT, where the Stay vote was expected to beat the Leave vote and it has not happened.
There are polls that support many of their assertions. There's evidence pollsters were re-working their polls and support was converging between Trump and Clinton prior to the initial Comey news that the FBI were re-opening the Clinton email investigation (before slamming it shut again today). But here's the thing - you can't say the pollsters were adjusting their polls AND that Trump was surging. It's one or the other. Additionally, there were a number of polls that showed Romney leading Obama in 2012 that turned out to be wrong. The preponderance of the evidence of polls was that Obama was leading and he actually was.
While there is a case to be made that Trump has made up ground and that there is this mythical Trump "monster vote" that will emerge on election day. Okay, maybe but there's no solid evidence. As much as I'd like to believe it and I've based a lot of my own analyses on skepticism of the mainstream polls, we might be operating in an echo chamber - we hear people saying things, often anecdotally, that we are predisposed to want to believe and hence exhibit a confirmation bias belief in that 'news'.
The same can be said for Democrats, who want to believe the polling general Clinton lead. They dismiss the anecdotal evidence of huge Trump rallies and small Clinton rallies as forcefully as Trump supporters embrace them. The Democrat echo chamber has been around a lot longer (and consequently a lot bigger) but that makes it no more correct.
What matters is a preponderance of polls in the swing states. If there's a hidden Trump monster vote, or true unmeasured momentum for him, I'll be as happy as anyone - Hillary Clinton is unfit to be president. Trump is a coin toss at least. But until that happens, I am dialing myself back on my confidence levels. I think at this point the ceiling for Hillary Clinton is probably about 280 electoral college votes and the floor is about 225. The same is roughly true for Trump. I think Trump's chances of winning are anywhere from 45% to 53% That's not terrible odds (unless you are Hillary Clinton and are working to stack the deck in your favor). As a Trump proponent, I'll take that. A Trump landslide would be wonderful, but it's not necessary.