****correction****

He didn't get 1960 right, my bad.

*edit* ugh I meant to say 'every single election'

"Norpoth relies on statistics from candidates' performances in party primaries and patterns in the electoral cycle to forecast results. The model correctly predicted the victor in every presidential election since 1996, according to the Daily Mail.

Running the model on earlier campaigns comes up with the correct outcome for every race since 1912, except the 1960 election."

You misunderstand. The model was almost certainly calibrated to data from the past, the getting those past results correct is not surprising. That's how you validate a model. He's only really gotten it right since 96, which is not nearly the same thing.

Yes, I did post a correction to the article above. The model was wrong about 1960, but Kennedy only won that by about one hundred thousand votes. Even factoring that loss in the model still has a very high accuracy when it comes to predicting who will win.

Here's the primary source:
https://static1.squarespace.com/static/522ead72e4b0a584ece4eab9/t/57a61cbb29687fc5a2708b8b/1470504123571/Vote+2016+PM+PS.pdf

Yes, from the primary source:

"4 It is 87-percent certain that Trump, not Clinton, will be the next
President. Trump benefits from a swing of the electoral pendulum to the Republican side in 2016 and his superior performance in early primaries"

See the conclusion. I am fine with these odds. These are good odds to go on.

Right but the model was *calibrated* with data from 1912 on, so by definition it should match almost perfectly. It's really only since 96 that we see it used. 5 elections are pretty good, but not as strong as you make it sound.

http://skepchick.org/2016/03/stop-panicking-there-isnt-a-98-chance-of-a-president-trump-future/

"According to the Primary Model website, the model predicted that in a 2008 a general election race between McCain and Obama would end in a statistical tie. In fact, Obama won the popular vote by 7 points and McCain didn’t even come close to Obama’s number of electoral votes, so the model failed to predict Obama’s win."

Polls aren't wrong.

"Polls aren't wrong."

-Lol

It's naive to say they're wrong. Brexit polls weren't wrong either. Look at the conglomerate, and you see they were completely inconsistent.

I am sure some polls will be proven right and some will be discredited.

The thing is that this model hasn't been wrong since 1960. I will believe in the model over polls.

538 used the polls in 2008 and was far more accurate in predicting that election.

"The thing is that this model hasn't been wrong since 1960. I will believe in the model over polls."
The model didn't exist in 1960. It was a retroactive prediction where the answer was already known.

No, it hasn't been wrong since 96. It fits data from 1912-92 with a very high R^2 value. Are you interested in learning the difference or do you want to keep believing something that is wrong because it better agrees with you?

ND, you're missing what abge is saying.

When you create a model, you need past data to base it off of. In this case, that past data is a bunch of elections prior to 1996. It must, by definition, predict those elections well, because that's how it was created. The relevant question is how it performs on data that wasn't used to train it. In those elections it is apparently 5 for 5, though of those elections, only two were close, and it apparently nearly got 2008 wrong anyway.

Don't worry ND, The Donald can beat the odds and still lose this election!

Explain to me how 100% accuracy from 96-2012 is going to be wrong in 2016 then.

I'm not sure if you've noticed, but this has been a rather abnormal election cycle.

No, I haven't noticed that.

52%

"Explain to me how 100% accuracy from 96-2012 is going to be wrong in 2016 then."
You've admitted Obama's 2008 victory was a landslide victory. The model didn't predict an Obama victory at all. That's not 100% accuracy. 538 had 100% accuracy from 2008-2012, and you've said you trust 538, but now you actively defy it because it gives Clinton the win?

The model did predict that Obama would win in 2008. So no, it wasn't wrong.

http://www.theonion.com/article/trump-holds-strategy-meeting-campaigns-top-militia-54481

The Onion? Are you for real?

It was funny ND. Its called satire..

"The model did predict that Obama would win in 2008. So no, it wasn't wrong."
It predicted a statistical tie. Just because it didn't call a McCain win doesn't mean it's accurate.