It didn't surprise me. You have to look through the reality distortion field of the media and established politicians, because that only represents the views of a tiny and intellectually very non-diverse group of people. Consider the mismatch between what the media was reporting, and the undeniable fact that Trump's rallies attracted a far greater number of people than Clinton's. You also can't rely on the polls when you have a candidate whose voters are underrepresented in those polls because they've never voted before. Many polls sampled people as if this was Obama instead of Clinton. It's was clear that the number previous non-voters who would actually go out to vote for Clinton would be lower than for Obama, whereas from Trump it's the opposite.

> Consider the mismatch between what the media was reporting, and the undeniable fact that Trump's rallies attracted a far greater number of people than Clinton's.

By itself that doesn't mean anything. Sanders got way bigger rallies than Clinton too and still lost.

Interestingly the Trump campaign was complaining about those polls and used that as a basis for their "rigged" claims. So I'd be curious how they produced their numbers internally that showed them ahead in tossup states.