Going to this link reveals that this “model” is built after the fact to fit previous events rather than built before the fact and tested by predicting future events. In other words, it works exactly like this:
Perhaps you two could argue this out and the winner can get back to us.
Nope. He created it in 1996, drawing from historical data. It’s been correct in 5 of 6 since then, including 2016.
No conflict at all. “Eager to vote” and “eager to talk about voting to a stranger” are not the same thing.
Some of the same pollsters who record high enthusiasm for Trump also say they do believe other Trump voters are avoiding them or even lying to them.
Unskew. Unskew.
Unskew like the wind.
His model is based almost soley on how candidates do in their party’s primaries, using data going back to 1912. This year the model is predicting 362 electoral votes for Trump. For that to happen, Trump would have to win every state except WA, OR, CA, NY, VT, NJ and DE.
When the Washington football team won their last home game, the incumbent party won re-election from 1940 to 2000. Then the model fell apart.
Pretty soon it is going to be 5 out of 7. 91% chance of 362 electoral votes is absurd to borderline whackadoo. Predicting a Trump win is one thing - that’s possible, though unlikely. But a crushing Trump win? No.
But props to old Helmut, since he has just given me another reason to election-watch on the 4th. I think I’m going to have to factor in likelihood of seeing this voodoo model smashed into my political drinking game strategy.
This model, constructed by Helmut Norpoth, a political science professor at Stony Brook University, is based on three flexible variables: a presidential voting cycle, long-term trends in partisanship, and performance in the primaries. For elections prior to 1952, the primaries variable incorporates all presidential primaries; for the 1952–2004 elections, only the New Hampshire primary is used; and for the 2008, 2012, and 2016 elections, the New Hampshire and South Carolina primaries are used. Norpoth seems blissfully unaware of the dangers of tweaking models to fit past data. We can anticipate more tweaking after the 2020 election.
The Economist model is based on sound theory, specifies the variables used, shows the source code used to make the calculations, and updates the predictions daily. It currently gives Biden a 99% chance of winning the popular vote and a 91% chance of winning the electoral vote.
True, the Economist model is based on assumptions, which may well be wrong, but models based on plausible assumptions are generally more reliable than models based solely on past statistical patterns.
“Drawing from historical data” is exactly what I said.
It’s quite easy to test the “shy Trump voter” model. It predicts that poll numbers for Trump will significantly lag poll numbers for other Republican candidates (unless you add a plethora of epicycles to the model to account for the “shy Collins voter”, “shy Tillis voter”, “shy Ernst voter”, “shy Cornyn voter”, etc).
In Mexico, 3KG bags of Cheetos are (or were) available.
[quote=“Mike_Mabes, post:246, topic:922525, full:true”]This year the model is predicting 362 electoral votes for Trump.
[/quote]Yes, and I don’t think that’s plausible.
My point was, and is, that while so many people are obsessively focused on opinion polling, that there are multiple fact-based indicators that point to a very different outcome. (With yesterday’s GDP numbers, Lichtman’s 13 keys is another, though I’m not a fan).
AFAIK there has never been an election where the two were this much at odds. It’s a fascinating situation, but people are missing it because the media doesn’t want to report on voter registration, etc.
No, it doesn’t predict that at all.
It predicts that Trump has an outsized share among people who don’t take polls. It is by definition unfalsifiable until votes are counted.
Ahh yes, the convenient single-serving package! 
No wonder the diabetes trajectory in Mexico is even scarier than that found one country to the North.
Trump’s going to win isn’t he? I’m not talking about in the courts or through some form of cheating. I just have this feeling that he is going to win Florida and then all he has to do is take Pennsylvania which is basically tied and he wins again.It just seems that we are headed for 2016 all over again. The polls will underrepresent Trump voters, particularly this year when there is so much vocal anti-Trump sentiment, the red states will vote reliable red and he will eke out a win in just enough swing states to win. I also don’t see the Senate flipping. Maybe it’s the rain but the closer we get to the election the further out on the ledge I get. I just can’t see how we can survive another four years.
It’s possible. But unless the polls change in the next 3 days it will be a MUCH more damning indictment of polling quality. Remember, they were only off slightly at the national level and there were only a handful of states that were significantly off (meaning like 5% or so). It just happened to be the key states.
This year, for Trump to win, the national polls have to be off by 5% or so and the states have to be off by even larger margins than in 2016. That’s not impossible, of course, but if it happens you pretty much have to completely write off polling entirely.
I would take issue with your “Trump wins FL and PA and he wins” statement. There is still a path for Biden in that scenario. It’s low-probability, but not lower than Trump’s overall chances right now. It’s Arizona plus one of the ME or NE districts.
Looks like The Donald has made his own prediction… 
He’s got a plane to catch to Moscow.
There was a thread a month-ish ago about how many newspapers would run “You’re Fired!” as their headline on Wed Nov 4. With luck, nearly every paper in the USA will do so and our near brush with tyranny will end with a giant raspberry from the body politic. Which will yugely important to delay / derail the next run by a wannabe tyrant.
Question about pardons:
What if Trump loses, and then in the next week or so, issues a pardon to every single person convicted of a federal crime?
Essentially, what if he empties the prisons simply as a “fuck you” to the country?
We experience a renaissance of fine dining, as hundreds of old Mafia gangsters open restaurants in their retirement?