Statistical Models, Demographics, and Other Forms of Witchcraft: 2020 Election Liveblog, Part II

Aaron Boyd
4 min readNov 3, 2020
Pictured: Nate Silver’s Polling Gurus (Source: Walt Disney Corporation, I guess?)

5:00 PM

Joe Biden will defeat Donald Trump.

It’s strange how that’s considered a ballsy, cursed statement. We’re all so traumatized from 2016 that virtually everyone, from the man on the street to our top statisticians and data analysts, has developed a paranoid superstition about Trump’s magical reality-warping powers.

Not that I blame them. I was as wrong as everyone else. I even registered as a Republican to vote for Trump in the 2015 primaries because I saw him as the perfect wrecking ball for the GOP.

Back then, my biggest fear was Hillary up against Jeb Bush (two remnants of would-be political dynasties that nobody has a taste for) or Marco Rubio (he’s like the Republican Obama! If Obama was really stupid!).

I saw Trump as a massive, expensive distraction that would throw the party into disarray by whipping up its worst elements, thereby scaring off moderates and forcing the others candidates to frantically rework their strategy. After a long, expensive, drawn-out primary, the winner would emerge battered, weakened, and hounded by an obnoxious loudmouth constantly giving terrible advice.

If anything, my biggest fear was that Trump would flame out too quickly; I remember his “I prefer my war heroes not captured” came just a couple days after he announced his candidacy, and I laughingly said “Oh noooo Donald, don’t torpedo yourself right away! We need you!”

Haha, see, it’s funny because I expected Republicans to be outraged at a millionaire draft-dodger calling our wounded veterans pussies. I expected them to have actual principles and to take a stand.

We all know how that turned out.

And I get it, 2016 was horrifying and humiliating. I don’t blame anyone for being traumatized and reluctant to appear overconfident twice in a row.

But counterpoint: Fucking come on.

Yes, Trump can win. Yes, that is statistically possible. But on what grounds are you basing this?

Biden’s led in the polls has been historically stable. At no point has Trump been ahead of him. Not in national polls, not in the Electoral College. Trump began his term having lost the popular vote by 3 million, and since then, he’s done nothing but make enemies and whine about how the quarter-million Americans that died from COVID-19 are making him look bad. His approval rating is below Carter levels and all his half-assed attempts at smearing Biden as an elderly molotov-hurling revolutionary belly-flopped spectacularly.

Biden isn’t Hillary 2.0. He doesn’t have the trust issues, lousy people skills, or vagina that all contributed to Hillary’s failure. Biden is seen as more calm, empathetic, and professional. He’s a safe, reliable alternative that doesn’t rock the boat so much he alienates seniors and moderates. And he never handed Trump the Magical Gaffe he was definitely relying on.

Everything about the Electoral Map favors Biden. Poll after poll has shown him so far ahead in most swing states that we had to redefine which states even qualified as “swing”. Texas is a swing state right now. Fucking Texas.

And yet nobody, myself included, could resist the Sirens song of Trump Math, which commands us to reframe every event in human history as beneficial to Trump. This lead to many bizarre scenarios where Biden would have a 5-point lead in Pennsylvania and people would say “Yes, but if we arbitrarily assign Trump 6 points, he’s ahead!”

No matter how good the numbers were for Biden, somehow they always translated into a Trump victory.

It wasn’t just raw data, either. Even abstract, theoretical events were reflexively assumed to be in Trump’s favor.

Early in the campaign, pollsters would hedge their findings with the stock phrase, “Of course, a lot can happen between now and the election,” with the hilarious implication that whatever happens in the future would damage Biden and not the one-man fuckup factory.

Even normally sober analysts talking about whether or not Biden’s lead was within the margin of error almost always — maybe even subconsciously — framed the issue as if “margin of error” automatically means assigning a bunch of points for Trump. You’d see polls saying he’s 5 points ahead, with a +/- 3-point margin, and people would translate that into “Biden’s lead is now 2 points,” and from there hand Trump another 3 points because fuck it, God is dead.

Call me old-fashioned, but last time I checked, being ahead is being ahead.

That doesn’t mean that I’m claiming with absolute, mathematical certainty that Trump can’t win. Unlike Trump voters, I understand basic probability. But there’s a difference between stating, as a rational fact, that an 11% chance of victory is an 11% chance of victory, and allowing blind fear to tell you that 11 is a bigger number than 89.

Setting aside any form of extralegal shenanigans (we’ll get to that in a moment), I am genuinely struggling to see a sober, rational argument in favor of Trump. Not a thought experiment, not paranoia, not acknowledging that any number greater than 0 means he can win. I mean an honest-to-god strategy.

I can’t find one.

So at the moment, I feel comfortable, if not slightly anxious. It’s like riding a plane through turbulence. You know that a plane is no more likely to flip over in a bumpy sky than a car will flip over on a bumpy dirt road. You keep telling yourself that everything is fine, that there’s no reason to panic…and then you notice how hard your sweaty palms have been gripping the armrest.