Donald Trump is publicly accusing Hillary of "rigging the system." 

Push-polling is one way to do that.  It amounts to artificially blowing up the Democratic poll numbers to "push" a psy-ops mission – to encourage one set of voters to think they are winning and to demoralize the opposition voters, who think they are losing.  One Alinsky rule is to do what your supporters enjoy: winning feels better than losing.  Saul Alinsky’s lessons have gone mainstream.

Reflections on Leadership - Page 49 - Google Books Result

Richard A. Couto - 2007 - ‎Business & Economics

You win a war by putting the enemy in a situation where he will not fight and where you triumph. ... He fought it because he said, "I don't want to give the enemy the impression that I ... So just to throw them off balance, we are going to have a battle. ... Saul Alinsky was a wonderful community action organizer, and he wrote ...

The Hillary Clinton campaign, in concert with allied media, has created a sense of doom for Trump supporters.  Here is the five-month average according to the New York Times.  

A 2008 memo revealed by WikiLeaks openly plans for oversampling.  The known pattern for news polls is actual oversampling of Democrats.  The memo goes to prove that it's intentional.  It openly plans for oversampling.  The known pattern for news polls is actual oversampling of Democrats.

Andy Meyer Sent: Thu Jan 10 21:21:07 2008 Subject: Meet and polling design Hey, when can we meet? I also want to get your Atlas folks to recommend oversamples for our polling before we start in February. By market, regions, etc. I want to get this all compiled into one set of recommendations so we can maximize what we get out of our media polling. -Tom

The term "oversampling" presumably means oversampling of Democrats vs. Republicans, because that is in fact what we see in many polling articles.  The headline says, "Hillary up by 5 points," and the body of the text tells us that Dems were oversampled by some percentage.  As you know, the math basis of statistics uses random variables, and whenever possible, statistical design uses equal size sample.  The polling company can say it did not lie, because it reported the oversampling.  The headline writers do the most important lying, because some big percentage of readers never get past the headline.

Since predictive polling is inherently imprecise, it's a guess.

I googled "Hillary ahead in polls" for 15,200,000 hits.