Like many other Medium analyses, this article fails to consider the soft support.

Here's the math. With a 40% approval rating and 250 m voters, Mr. Trump should get 100m votes. He did not get this number in 2016. He won't get it in November. Despite his vociferous base, a lot of his support is soft------and can be convinced not to make the trip to the polls on election day.

Likewise, Mr. Comey's email allegation a week before the election probably cost Ms. Clinton 500,000 of her soft supporters. They just decided not to make the effort to get themselves to the voting station. Such a non-vote is almost as good as a vote for Mr. Trump.

And there's 108 m Americans who did not vote. If just 10% of them are annoyed enough to cast a vote away from Mr. Trump, that would be a solid D victory. 20% would be a D landslide.

To win, political parties need to identify their soft support and get them to the polls on election day.

To win, political parties need to attack the soft support of the other side--and convince them not to vote.

For sure, but the D's and R's will be playing this game. I Just would have thought a journalist from The Atlantic would know about soft support.