Donald Trump gestures after his speech during a campaign rally in Austin. (Suzanne Cordeiro/AFP/Getty Images)

I'm getting to the point where I know that I'm being repetitive, and I know that I sort of have to be, but that I feel the need to predicate my repetitiveness with a self-indulgent recognition of that repetitiveness, so that if you, the reader, also noticed that I have been repetitive, you will hopefully, at least, know that I know that I am repeating myself and you will, therefore, give me a pass on the need to once again repeat myself. So:

Donald Trump cannot win this election unless he wins a state that Mitt Romney lost in 2012.

If you haven't heard me say this before: Welcome! Thanks for reading.

I keep coming back to that point because it is the simplest way of thinking about this election. For all of the polling that's sprinkled down over the past year-and-a-half, for all of the demographic breakdowns and discussions of who and what is doing well where and why, that's the only thing that really needs to be tracked. Is Trump winning a state Romney lost? And if the answer is no, then: game over.

For the first time in a while, though, we can answer that question with a yes instead of a no. And then we can hang an asterisk on it.

Instead of immediately indulging your curiosity about the state (or states??) being discussed, let's instead set some context.

Five new polls came out on Thursday from four critical battleground states: Florida, Ohio, North Carolina and Pennsylvania. (North Carolina was polled twice.) Florida, Ohio and Pennsylvania were all polled by the same outfit a month ago, during the peak of Hillary Clinton's recent national poll numbers. By comparing the new polls in those states with the old ones, we can get a sense for what changed — and what trends may be underlying the numbers.

Sadly, though, doing so doesn't offer us much insight.

The polls are from Quinnipiac University, covering early August and early September. So what's changed? In Florida, things stayed mostly the same overall, though Republicans and Democrats both moved a bit to Clinton. In Ohio and Pennsylvania, Trump gained 5 points against Clinton, bringing him into a tie with her in the Buckeye State. Trump saw big gains with nonwhite voters in Pennsylvania and with white women; white men in Florida and Ohio moved toward Trump, but white women didn't.

Especially since we're talking about changes of 6 to 10 percentage points in demographic groups with margins of error that approach those figures, it's at best murky.


This, of course, is why we look at trends. And why I turn to the RealClearPolitics averages of polls in states to get a sense of how things are changing.

I pulled averages for the last 100 days of the campaign in 12 of the 13 closest states from 2012 to try and address the repetitive point I made at the beginning of this article: Is Trump poised to do better than Romney? (Only 12 of 13 because Minnesota hasn't been polled enough to generate a poll average.)

Let's walk through them, from least- to most-competitive in 2012. The yellow lines show the average in 2012; the blue, 2016.


The new Florida poll takes Clinton from a decent lead to a near-tie. This is absolutely critical for Trump; he essentially can't win the presidency without winning the state. Clinton's outperforming President Obama in 2012 at this point in the race (relative to how Obama fared against Romney), but the trend has shifted back toward Trump.


In North Carolina, Clinton maintains a slight lead. This is important because Romney won North Carolina in 2012. Trump should win this state — but right now, he isn't. More on this later.


Ohio looks like Florida: Clinton's once healthy lead is now a small one. No Republican has ever won the presidency without winning Ohio, which isn't a determining factor, but still reiterates the importance of the state. Clinton's doing about as well right now as Obama was.


Clinton's once giant lead in Virginia is now a decent-sized lead in Virginia — but she's still outperforming Obama (and the 2012 result). That said, the last thing Clinton wants to see is plunges like the one in this graph. More of those, and she's in trouble. Especially if they're in states like Virginia and Colorado, which were basically taken off the map a few weeks ago.


Not that Colorado should be put back on the map.


In Pennsylvania, Trump's promise to outperform past Republicans in the state is sort of coming true: Clinton still has a big, big lead, but Obama's lead at this point was bigger. Clinton's lead is also still bigger than Obama's margin of victory four years ago, but at least Trump can say that he's doing better than Romney (right now).


Clinton still has a big lead in New Hampshire, though its four electoral votes don't make much difference anyway.


Which brings us to Iowa.

Iowa is the first state of the 10 closest in which Trump has seen a polling-average lead since the conventions. The problem is that Iowa only has six electoral votes. If he wins every state Romney won and then flips Iowa, Trump would lose by 114 electoral votes instead of 126.

And that's if he holds all of the Romney states. If he wins Iowa but loses North Carolina, he's gaining six electoral votes but losing 15, an 18-vote swing. But as we'll see in a second, North Carolina's not the only red state at risk.


Since we're doing this in order, though, we need to dispatch with Nevada and Wisconsin, first. In Nevada, Trump is doing better than Romney, but still losing. (Why is Trump doing better in Iowa and Nevada? We looked at this last week.) The polling average in the state is a bit out of date, though, so consume a few grains of salt with this one.


In Wisconsin, Clinton's lead looks like her one in Virginia: a big shift toward Trump but still a big lead. She's outperforming where Obama was at this point in the cycle, but underperforming relative to the final result.


So let's talk about those other red states. Right now, Trump leads in Georgia. He's supposed to; Romney won it by nearly 8 points. But Trump's lead is less than 2, and Clinton is running nearly 8 points better than Obama was at this point in 2012. There was less polling then (since no one thought Obama could win), but that's a big shift.


Same in Arizona. Clinton trails here, too, but is outperforming Obama's position at this point by 9 points — in a state that Romney won by 9 points.

Trump is closer in Ohio, Florida and North Carolina than Clinton is in Georgia or Arizona, but this is distracting us from the main point: Trump has to win Ohio or (well, probably and) Florida; for Clinton, Georgia or Arizona is a feather in the cap.

Why? Because Trump needs to win a state Romney lost and hold the ones Romney won. It's a point that bears repeating.