US 2020 Presidential Election prediction thread

[Barrister]

This guy predicts a 90% probability of a Trump win:
http://primarymodel.com/

He made his predictions on March 2nd 2020 and he sticks by it even now.

Dude's not completely insane.  Basically he looks at how candidates did in the primaries and then uses that to see who will win in the general election.  And the last two single-term presidents who lost seeking re-election did in fact have significant primary challengers (Pat Buchanan in 1992 against Bush 41, Ted Kennedy against Carter in 1980).

But given that you're drawing from an incredibly small sample size I dare say this is not a particularly powerful predictive model.  Plus it's worth noting that the GOP shut down the primaries in most states.

[DGuller]

This guy predicts a 90% probability of a Trump win:
http://primarymodel.com/

He made his predictions on March 2nd 2020 and he sticks by it even now.

So I read this article where he explains his model:  http://primarymodel.com/2016-forecast-model-full-version.  He seems to tout himself being the only one who correctly predicted Trump win, and with great confidence.  For some reason he doesn't tout that his model gave Trump a 5% edge in popular vote, which is why he was 87% confident. 

The whole write-up sounds like a confused rambling of a dilettante, who seems to have no understanding of the danger of overfitting to the past.  This is a perfect example of how a bad model can look good by getting lucky once.

[HisMajestyBOB]

2000 and 2016 are the free spaces of election modeling.

[The Brain]

What's the Freebird of election modeling?

[alfred russel]

2000 and 2016 are the free spaces of election modeling.

Also 1960.

And you can argue that prior to 1964, none of the elections should really count because most black people were systematically denied voting rights. So 1968 was the first comparable election. Maybe 1972 as that was the first that allowed 18 year olds to vote.

If we start with 1972 and exclude 2000 and 2016 we have all of 10 elections to go off of. 

[alfred russel]

If Predictit means anything, Trump is at 42% to win.

He is over 50% to win states with 259 electoral votes, and at 44% in Pennsylvania which would give him 20 more (270 needed to win).

[Admiral Yi]

Judge Bozo has dismissed the Texas drive through case.

[Valmy]

Judge Bozo has dismissed the Texas drive through case.

Yeah but is that the end of it? Can they still appeal the dismissal?

[Admiral Yi]

Yeah but is that the end of it? Can they still appeal the dismissal?

The plaintiffs say they're going to appeal.  But I'm going to dismiss the story from my mind and suggest you do the same.

[The Minsky Moment]

Judge Bozo has dismissed the Texas drive through case.

Yeah but is that the end of it? Can they still appeal the dismissal?

Yes but it's dead.

[Valmy]

Ok good. Lawyering is a strange art that still confuses me even though I have been working in administrative hearings court for five years now.

[Eddie Teach]

Now there's just the matter of Judge Bozo's impending libel suit.

[DGuller]

Yeah but is that the end of it? Can they still appeal the dismissal?

The plaintiffs say they're going to appeal.  But I'm going to dismiss the story from my mind and suggest you do the same.

Before you dismiss it from your mind, think of all the money you could've gotten had you not 86ed me from your bet.

[FunkMonk]

Not gonna lie. Pretty psyched for tomorrow. Gonna have a good night throwing this moron asshole out of office.

[DGuller]

I just can't make myself be this confident.  This fucker is the definition of Karma Houdini.  I also think that Biden winning without the Manchin-proof Senate majority would merely turn an unmitigated disaster into a mitigated disaster, and that part is far less certain.