The recent presidential election polls have been mixed — but mostly between polls that are good for Hillary Clinton and polls that are fantastic for her.
Clinton now leads Donald Trump in national polling averages by about 7 points, with every recent live interview poll showing her up by between 4 and 12 points.
She continues to hold on to solid leads in a series of states that would put her over the 270 electoral votes she needs to win, and she seems to have taken the lead in even most of the swing states she doesn't need (like Florida, North Carolina, and Nevada).
Read more from Vox:
I'm voting for Jill Stein. It's a moral choice. It reflects who I am as a person.
I trolled my IRS scammers for weeks. I learned something really dark.
Third presidential debate live stream: how to watch Trump vs. Clinton online, TV
New, troubling signs are emerging for Trump in states like Utah, and the Clinton campaign is signaling confidence by making a new push into Arizona and focusing more on down-ballot races.
And every major election forecasting model gives Clinton at least an 87 percent chance of victory.
The election is three weeks away, and there's still time for the race to tighten somewhat. But we're getting to the point where, for Trump to win, we'd need either a truly seismic event to transform the campaign or massive systemic polling failure.
Clinton is leading almost every national poll
There have been more than 24 national polls of the presidential race released in the past two weeks, and Hillary Clinton is leading in nearly all of them.
Recent live-caller polls of a four-way race have shown a Clinton up 4 (ABC/Washington Post), up 7 (Fox), up 8 (GWU/Battleground), up 9 (CBS News), up 11 (NBC/WSJ), and up 12 (Monmouth).
Internet polls and automated phone polls, however, have tended to show somewhat smaller leads for for Clinton — she's up by an average of just 2.4 points in those, according to HuffPost Pollster.
Only one pollster has shown Trump ahead in its newest survey — that's the LA Times/USC tracking poll, which has long had a pro-Trump lean largely for idiosyncratic methodological reasons, as the Upshot's Nate Cohn explains. Rasmussen polls, which often tend to show a pro-GOP "house lean," have also sometimes shown Trump ahead, but Clinton leads in the outlet's latest poll by 1. These are very clear outliers at this point.
Meanwhile, Gary Johnson and Jill Stein both appear to have lost a couple of points' worth of support. Johnson now averages about 6.5 percent support and Stein averages 2.4 percent, according to RealClearPolitics. Clinton appears to be the main beneficiary here — her margin in a four-way race used to be smaller than her margin in a two-way race, but lately it has tended to be about the same.
The state polls look either good or great for Clinton
For months, Hillary Clinton's easiest path to an Electoral College majority has appeared to be through winning six states: Wisconsin, Michigan, Virginia, Colorado, Pennsylvania, and New Hampshire.
And she still appears to be on track to make that happen, with room to spare. The latest HuffPost Pollster averages show her up at least 4.9 points in each of these six states. There have been zero polls in the past two weeks showing Trump ahead in any of them.
Clinton is also up 7.5 points in Maine, a state some speculated Trump could make competitive. It remains possible that Trump could pick up one electoral vote in Maine's Second Congressional District, but if she wins the above states, that wouldn't affect the electoral math — she'd already be at 272 electoral votes.
And it gets worse for Trump from there.
Florida is an absolute must-win state for Trump, since there are 29 electoral votes in play. But though the Sunshine State looked close in September, Clinton appears to be pulling away there now, leading by 3 to 4 points in the five most recent polls.
North Carolina — a state Mitt Romney won in 2012 — now looks to be leaning Clinton. The past 11 polls of the state have shown her ahead.
Nevada polls also now tend to show Clinton ahead; Trump hasn't led any of the past 7 polls in the state. Only six electoral votes are at stake in Nevada, but there is a competitive Senate race there, and the GOP candidate recently unendorsed Trump.
Ohio, a swing state Trump hoped he had put away, now looks close again. (Meanwhile, Trump is at war with the state's Republican Party.)
The Clinton campaign is even making a big push into Arizona, which has been a red state for decades, and is funding voter turnout pushes to help Democratic Senate candidates in Indiana and Missouri, two states Trump is expected to win.
And finally there's the strange case of Utah, a state Romney won by 48 points. Trump is very unpopular among Mormons, who make up about two-thirds of the state's population, and several leading politicians in the state condemned and unendorsed him after his leaked tape scandal. And lately, new polling shows third-party candidate Evan McMullin (who is Mormon) surging, to the point where he may have a shot to actually win (or tip Utah to Clinton).
The upshot of all this is that Trump has gotten almost uniformly bad news in state races in the past few months and will now face a tremendously difficult time winning 270 electoral votes.
The forecasting models view Clinton as an overwhelming favorite
But enough about polls, you say — what are those data wizards with their fancy models telling us?
Well, they are essentially telling us the same thing as the polls — that Hillary Clinton is an overwhelming favorite to win.
FiveThirtyEight is currently the least pro-Clinton of the major forecasters — as of midday Tuesday, its polls-only model put her at a mere 87.4 percent of victory. But every other major model has her in the 90s, as you can see at the Upshot's rundown.
There were bigger discrepancies during the various models at earlier points in the race — mostly about whether Clinton should be viewed as a narrow favorite or a solid favorite. But now they're all basically saying the same thing.
Now, the polling averages are not infallible — they've frequently been off from the final outcome by a few percentage points. Indeed, they undershot Barack Obama's margin of victory by a few points in 2012, and tended to underestimate Republican strength in 2014.
So, yes, the polls could be wrong — but at this point, Clinton appears to be up by enough that a Trump victory would mean a truly massive polling error. Also keep in mind that Trump's ground game and turnout operation are reputed to be dreadful.
Trump's best hope, then, is that some truly major, unpredictable news event scrambles the race in the final weeks. Because the way things are looking now, he's headed to a big defeat.