There’s so much information about polling and turnout floating around right now that it’s almost impossible to shape it into a narrative that makes sense. So let me toss what I’ve seen so far today at you and you can choose your own adventure.

We’ll start with the stuff that’s auspicious for Trump — although, for my money, nothing bodes as well for him as what Ed’s been writing about lately, the fact that Democrats appear to be taking Minnesota seriously. That should not be necessary in a national environment where Biden is supposedly eight points ahead and states like Iowa and Ohio are in play. Minnesota should be a comfortable hold for Sleepy Joe if the electorate really has moved left by six points this year or whatever. So what was he doing campaigning there yesterday?

Ed and I each wrote about growing Democratic panic over turnout in Miami-Dade County yesterday but the hits keep on coming in the media. Sixty-three percent of Republicans have voted already in M-D versus just 56 percent of Democrats. Dem strategists aren’t hiding their anxiety about it either, with one telling CNN about the poor turnout, “I don’t have a good answer for why that might be. But that is certainly a factor at the moment that is giving Democrats a lot of worry.” Light conspiracy theorizing has already begun to try to explain it:

Raw footage of mailroom in post office here in Miami Dade. Source revealed “mail in ballots are within these piled up in bins on the floor. Mail has been sitting for over week!.” @AmandiOnAir @PeterSchorschFL @MarcACaputo @GlennaWPLG @CNNPolitics @NewsbySmiley @realDonaldTrump pic.twitter.com/DO8jx1VUnz

— Kionne “Commish” McGhee (@kionnemcghee) October 30, 2020

Maybe 50,000 ballots or so are stuck in a post office somewhere and just haven’t been delivered yet? I’m skeptical. And then there’s this:

Republicans gained another 50,000 today in Florida. Democrats’ lead has been cut from 430,000 to about 110,000. Republicans have 2 days to pick up 25,000 more to outperform their 2016 numbers. Unless Ds stop the slide this weekend, Biden is in a bind in Florida.

— Joe Scarborough (@JoeNBC) October 30, 2020

As for the big picture nationally:

Trump pollster @JMcLghln is privately telling people that Trump is ahead in ALL the swing states – including Penn and Michigan, and tied in Wisconsin.

I don't believe it. But if he's right, he's a genius. If he's wrong, I wonder if he'll ever work again.

— Frank Luntz (@FrankLuntz) October 31, 2020

That’s encouraging, but make sure you read this about McLaughlin’s big misses in the past before you start celebrating. He has a rap as a guy who tells his clients what they want to hear. This very exciting result in RCP’s average of Arizona polls also comes with a caveat:

Winning Arizona would be huge for Trump, and would really put the pressure on Biden in the Rust Belt. But today’s average is being driven by the two Trumpiest pollsters in America, Rasmussen and Trafalgar, both of whom have the president ahead. They could be right! But when a bunch of polls that typically tilt in favor of one side drop, it’s worth discounting them accordingly.

Now for the less auspicious stuff. A few days ago I wrote about how Democrats seemed to be underperforming in early voting in Nevada, potentially giving Trump a huge pick-up opportunity. If he loses Michigan, Wisconsin, and Arizona but flips Nevada, he’d hold on narrowly for a second term. Jon Ralston, who knows this stuff better than anyone else in Nevada, updated his projections last night to say that Dems there are now hitting their targets.

Early voting blog updated with numbers that show 81K Clark Dem firewall, 45K statewide lead. New models show Biden lead holding steady, Trump needs miracle scenario unless GOP can change mix of electorate.
Come for the numbers, stay for the "Vertigo" ref.https://t.co/QaEyq6ojmD pic.twitter.com/6MnI95pgNo

— Jon Ralston (@RalstonReports) October 31, 2020

Enough voters have turned out early in NV that Ralston thinks it’s unlikely (although of course not impossible) that Trump can completely close the gap on Tuesday. Meanwhile, in Texas:

Early vote jumped today in Hidalgo Co (McAllen), as @KamalaHarris @BetoORourke @JulianCastro rallied supporters

County finishes w 187,893 early + mail votes.

There were 173,437 total votes there in 2016

E Day to come.

— Jonathan Martin (@jmartNYT) October 31, 2020

If you read my post on early voting in Texas yesterday, you know that Latino turnout in border counties like Hidalgo had been the one discouraging sign there for Biden so far. If he wants a realistic shot at flipping that state, he needs those voters out in force. Harris’s visit may have helped solve his problem. Texas is still a longshot, but not nearly as long as it usually is.

The truly worrisome data for Trump today comes from polling conducted by *Republicans.* This is from Remington, a GOP firm:

President (Missouri)

Trump (R) 50%
Biden (D) 45%
Jorgensen (L) 2%
Hawkins (G) 1%

10/28-10/29 by Remington Research Group (C-)
1010 LV

Arf arf! Have a nice day!

Poll #135857 #ElectionTwitter

— Stella (@stella2020woof) October 31, 2020

Trump won Missouri by nearly *19 points* four years ago. If Biden has trimmed double digits from him there, odds are that he’s cut a chunk from Trump’s 2016 margins in other states as well. And Remington’s result isn’t a wild outlier. Check the other polls of Missouri over this past month and you’ll find the president reliably ahead by mere single digits. (He saw a 10-point lead in one poll.) Other internal Republican surveys are flashing warning signs as well:

“Well, the president’s losing Arizona. And, you know, we think that he and Martha are very intrinsically tied together,” Kevin McLaughlin, the executive director of the National Republican Senatorial Committee, explained during a 90-minute presentation on the state of the races Thursday. Trump won Arizona by 3.5 percentage points in 2016 and is now trailing Democrat Joe Biden, according to both public and private polling, which Republicans feel is why Sen. Martha McSally (R) is also trailing in her key race.

At one point, NRSC strategists believed Biden hit 50 percent in Georgia — a figure they found “terrifying” as they try to defend two seats in the state, which Trump won by five percentage points in 2016.

Trump won Kansas by more than 20 percentage points in 2016, but now his lead is in the low single digits, according to Republicans, after bleeding support in the Kansas City suburbs. A state that has voted Democratic just once since 1940 could now be considered a relatively competitive fight between Trump and Biden.

The NRSC also recently found him trailing in Alaska. See why it’s hard to believe that Minnesota is really in play? As for Georgia, the threat is real enough that Barack Obama’s headed there on Monday to make a final pitch for voters to turn out. Texas is still a longshot for Democrats but I’m not so sure Georgia is. And if they flip that state, the allegedly all-important state of Pennsylvania would become a sideshow. Georgia + Michigan + Wisconsin is all Biden needs, presuming Trump doesn’t flip any Hillary 2016 states — like Minnesota.

Speaking of which, the last Muhlenberg College poll of PA came out this morning and has Biden ahead by five, which is in line with other recent polling there. The good news for Trump is that he’s made up some ground; last week’s Muhlenberg poll had him down seven. The bad news for him is that 54 percent say he doesn’t deserve reelection. Winning the state when a solid majority believes you deserve to lose would be … surprising.

Let me give you a comparison that might cut through all of this noise a bit. At the end of the day, elections are popularity contests. Here’s where the two candidates stood in 2016 on favorability:

Don’t focus on the fact that Clinton led Trump the whole way over the final three months, focus on the raw numbers. She was viewed more unfavorably than favorably for the entirety of 2016, and her numbers dropped rapidly in the final week of the race after the Comey letter about Emailgate was published and late deciders started breaking for Trump. Voters hated Hillary; they hated Trump a little more, but ultimately her unpopularity seems to have neutralized his. Americans were willing to roll the dice on an intriguing outsider whom they disliked instead of an establishment dinosaur whom they also disliked.

Here are the favorables for Trump and Biden this year by comparison:

Trump is less unpopular now than he was as a candidate four years ago, but look at Biden’s trajectory. From mildly unpopular in August to somewhat *popular* now. He’s gained closed to 10 points on balance in favorability over the past three months, which is remarkable given that Trump has been attacking him relentlessly. He now leads Trump by close to 20 points in net favorability. (Although, importantly, Biden is still unpopular on balance in the Muhlenberg poll of Pennsylvania today, if not quite as unpopular as Trump.) We’re all thinking hard about the race but that puts us at risk of overthinking it. If the public kind of likes one candidate and really doesn’t like the other, and the guy they dislike has been managing a pandemic that’s killed 225,000 people and keeps acting like it’s no big deal, what’s the probable outcome?

I’ll leave you with this, the USC Dornsife “panel” poll. Famously, the final USC poll in 2016 had Trump ahead by three in the popular vote over Clinton, the only national survey to see him winning. That was wrong inasmuch as Trump didn’t win the popular vote, but its numbers may have captured the late break towards him among voters just before Election Day. This year their poll has scarcely budged in months. Today they have it Biden 54, Trump 42. “Catastrophic” doesn’t begin to describe the magnitude of the polling error if Trump pulls it out on Tuesday.