The timing of the poll overlaps with a weekend in which news about Clinton focused on her health and the "deplorables" comment, but that's almost certainly not what drives the numbers. The pollster notes that the anticipated electorate in this poll looks a lot more like who turned out in 2004 than who turned out in 2012 — meaning that it's more Republican-friendly. (George W. Bush won the state that year, which you knew since we started this thing out by saying every winning Republican president had won Ohio.) "It is very difficult to say today who will and who will not show up to vote on Election Day," pollster Ann Selzer writes. "Our poll suggests more Republicans than Democrats would do that in an Ohio election held today, as they did in 2004 when George W. Bush carried the state by a narrow margin."
This is one key reason we prefer polling averages to individual polls: Different pollsters make different assumptions about electorates, which poll averages can help to smooth out. If we look at the average in Ohio from RealClearPolitics, it's clear that Ohio has tightened. As it stands, the state is basically tied, although Clinton has a slight lead thanks to a recent poll that had her up by seven points. Clinton's big lead after the conventions has gone away.
But let's step back and put this into broader context. I've noted repeatedly that the key to a Trump victory in November is that he has to win at least a few states that Mitt Romney lost. Why? Because Romney lost.
For some time, there were no states that fell into that category. Over the past few weeks, two states have flipped into that category: Iowa and, more recently, Florida. Ohio looks like it could get there at any point. So what does that tell us about the likely outcome of the race? That Trump still has a big hill to climb.
Let's look at the states that are within five points currently in the RealClearPolitics average of recent polls. They are: Arizona, Florida, Georgia, Iowa, Missouri, New Hampshire, Nevada, North Carolina, Ohio and Virginia. Trump leads in five of those 10 states, including Arizona, Georgia and Missouri, which Romney won. Here's how each has trended since the beginning of August, with lines weighted to the number of electoral votes in each state.
Notice that Trump trails in North Carolina. Arizona and Georgia — red states — are both within two points. It's still safe to assume that Trump will win them, just as it's safe to assume that Clinton will win Virginia and New Hampshire (where Clinton's leads are bigger). So what would the effect of Trump holding North Carolina and winning Ohio, Florida and Iowa be? He'd lose by 20 electoral votes. (You can test out the election possibilities on our interactive feature, below.) If he won each of the states on the list above, and nothing else changed, he'd win by 26. Lose Virginia and win everything else, and it's a tie.
Put another way: Trump needs to win 116 electoral votes from the ten closest states. Clinton needs to win 14.
We spent so long looking at a contest in which Trump faced a shockingly steep climb that we somewhat lost sight of how difficult the climb would be for any Republican. The good news for Trump is that he's clawed his way back to that position. The next challenge is to continue that trend and keep picking up states that Romney lost. Having an electorate like the one that Selzer gamed out in Ohio will certainly help.
Game out the election below. Current polling uses either the current RealClearPolitics average in a state or the national average weighted to how the state voted in 2012.