Fivethirtyeight.com had Clinton with a 70% chance of winning on election night and a 65% chance just a couple of days prior. With many of the states within the margin of poll error. If you want to click on individual state polls, you can see narrowing of Clinton's lead at the run up to the election. If you look at the state by state, only Wisconsin jumps out as the big surprise on election day.... none of the individual polls showed a margin with the error range. The other states showed one or more polls with very thin Clinton leads in the final polling.
That said, 70% is not a sure thing. If the weatherman says there is a 70% chance of rain, and it does not rain, he was not wrong. So, surprise or political upset, yes.... but the possibility of a Trump win existed on election day. Then, of course, we really do not know how much the Russian efforts contributed. With a 76,000 vote margin over three states, getting some combination of getting 76,000 clinton voters to stay home in apathy or 37,000 to change their votes is not only plausible; its not a stretch to say it was likely.
https://projects.fivethirtyeight.com/2016-election-forecast/
We are twisting off into never never land here..... there is a huge difference between a presidential election poll (which is comprised of 51 individual polls that are BOTH opinion and projection of votes) and a simple opinion poll (which is one poll without a projection of voting). My original post was about an opinion poll. The poster was trying to tell us, from his low information platform, the Trump was extremely popular in those states he barely won. He is not. In Michigan, Pennsylvania and Wisconsin, his job approval is well below his already rather dismal national job approval.