I think this is only true if the election results are both fair (irrelevant levels of cheating or human error) and close.<p>Should Trump win by a significant margin e.g. > 4%, that <i>should</i> cause you to ask whether systemic bias exists in polling. I think most of us in Europe concluded something like this when Brexit occurred. Some people were claiming the polling was accurate but the reporting was inaccurate, however when I doubled checked the polling I could see no discrepancy. Most of that weeks polls were very wrong and so were the bookmakers, the people with strong incentives to discern truth from bullshit.<p>Take a look:<p><a href="https://en.wikipedia.org/wiki/Opinion_polling_for_the_United_Kingdom_European_Union_membership_referendum#Standard_polling_on_EU_membership" rel="nofollow">https://en.wikipedia.org/wiki/Opinion_polling_for_the_United...</a><p>It is really remarkable. Internet polls are accurate?!! Who knew! Depending on your reasons for believing in a significant prejudice in polling, that could indeed change the narrative.