538 Model Megathread

[Wells]

Absolutely astounding that fraudster Nate Silver is making Donald Trump not the biggest conman in this election.

So are you still supporting Trump or not?

[Erich Maria Remarque]

Absolutely astounding that fraudster Nate Silver is making Donald Trump not the biggest conman in this election.

[Erich Maria Remarque]

Absolutely astounding that fraudster Nate Silver is making Donald Trump not the biggest conman in this election.

So are you still supporting Trump or not?

Sarcasm is in the air!

[Wells]

Absolutely astounding that fraudster Nate Silver is making Donald Trump not the biggest conman in this election.

So are you still supporting Trump or not?

Sarcasm is in the air!

I know every post he makes is a joke (especially if it involves Trump), but I was very entertained by his Trump support.

[100% pro-life no matter what]

So, if the election were held today, Rhode Island would be closer to the national average than New Missouri Iowa!

[Sprouts Farmers Market ✘]

Absolutely astounding that fraudster Nate Silver is making Donald Trump not the biggest conman in this election.

So are you still supporting Trump or not?

I wouldn't dream of stopping now. It's time for a New New Deal in the White House, and only one candidate is willing to give it to us.

[RFayette]

Absolutely astounding that fraudster Nate Silver is making Donald Trump not the biggest conman in this election.

So are you still supporting Trump or not?

I wouldn't dream of stopping now. It's time for a New New Deal in the White House, and only one candidate is willing to give it to us.

Wait, why did you hate Bernie so much then?

[Dr. Arch]

Absolutely astounding that fraudster Nate Silver is making Donald Trump not the biggest conman in this election.

So are you still supporting Trump or not?

I wouldn't dream of stopping now. It's time for a New New Deal in the White House, and only one candidate is willing to give it to us.

Wait, why did you hate Bernie so much then?

Not enough xenophobia/fear.

[afleitch]

Absolutely astounding that fraudster Nate Silver is making Donald Trump not the biggest conman in this election.

So are you still supporting Trump or not?

I wouldn't dream of stopping now. It's time for a New New Deal in the White House, and only one candidate is willing to give it to us.

Wait, why did you hate Bernie so much then?

Because of his SJW cultists. I supported him tacitly until it became about muh transgender snake people campaigning on equality for the oppressed demipolysexuals rather than anything of substance for real people who actually matter (i.e. not privileged attention-whores).

e: this is not intended to be transphobic. I simply listed that as a group that is deeply over-represented in the cult.

You're making a broad assumption that Donald Trump is a real person.

[Figueira]

So, if the election were held today, Rhode Island would be closer to the national average than New Missouri Iowa!

Uh, no? The nowcast has Trump 3 points ahead in Iowa, and Clinton 66 points ahead in Rhode Island, and 3 points ahead nationwide.

[Figueira]

So, if the election were held today, Rhode Island would be closer to the national average than New Missouri Iowa!

Uh, no? The nowcast has Trump 3 points ahead in Iowa, and Clinton 66 points ahead in Rhode Island, and 3 points ahead nationwide.

The nowcast has Clinton up 9 in RI.

Whoops, I was reading the wrong graph.

[Figueira]

Lots of weird stuff from the recent New England polls.

[Xing]

Nate Silver actually believes that Rhode Island will be a single digit race. Wow. He's really lost it.

[Wiz in Wis]

Nate Silver actually believes that Rhode Island will be a single digit race. Wow. He's really lost it.

He also added the poll showing TX tied. This isn't about "believing" a poll, he just takes every data point into account. Cherry-picking polls would defeat the purpose of his model.

Yes... I'm so tired of people attributing every change in the 538 model to Nate Silver "choosing" to believe X or Y. He's just putting in new data and hitting "run" a few dozen times a day.

[Xing]

My point is that he used to at least weigh "state fundamentals" to at least some degree, meaning that if a poll showed Rhode Island or Texas close, the model would not immediately follow suit. The margins would narrow somewhat in each state, but if previous elections, demographics, and other various "indicators" did not suggest the state should be close, it would take several polls, all consistently showing a similarly close race for the model to also consider it a close race, though state fundamentals would still have some impact on the margin.

I'm sick of people acting like the model is its own sentient entity. It's not. He created the model, so it represents what he thinks the polls suggest about the race. If the results of his model were absurd to him, he'd tweak it.

[Wiz in Wis]

My point is that he used to at least weigh "state fundamentals" to at least some degree, meaning that if a poll showed Rhode Island or Texas close, the model would not immediately follow suit. The margins would narrow somewhat in each state, but if previous elections, demographics, and other various "indicators" did not suggest the state should be close, it would take several polls, all consistently showing a similarly close race for the model to also consider it a close race, though state fundamentals would still have some impact on the margin.

I'm sick of people acting like the model is its own sentient entity. It's not. He created the model, so it represents what he thinks the polls suggest about the race. If the results of his model were absurd to him, he'd tweak it.

Ok, but he has 3 models, which one are you referring to? The Polls-plus model has been almost painfully stable for weeks!

[emailking]

I'm sick of people acting like the model is its own sentient entity. It's not. He created the model, so it represents what he thinks the polls suggest about the race. If the results of his model were absurd to him, he'd tweak it.

Predicting Trump wins California with 90% likelihood would be absurd. Predicting there's a chance Rhode Island could be within single digits is not absurd.

And I don't think he'd tweak the fundamentals of the model for this cycle at this point anyway.

[Mr. Morden]

It would be rather dishonest for 538 to continually tweak the model over the course of the campaign to "correct" predicted state margins that "seem implausible".  If you do that, you're basically replacing a model with your gut.  Just leave the model as is, let it run, and if there are any huge misses on election day, you can work to improve your model for the next election.

[Xing]

I'm sick of people acting like the model is its own sentient entity. It's not. He created the model, so it represents what he thinks the polls suggest about the race. If the results of his model were absurd to him, he'd tweak it.

Predicting Trump wins California with 90% likelihood would be absurd. Predicting there's a chance Rhode Island could be within single digits is not absurd.

And I don't think he'd tweak the fundamentals of the model for this cycle at this point anyway.

He's not just predicting that there's a chance that Rhode Island could be within single digits, that the aggregate margin for all three models. According to all three models, there's also a chance (greater than 10% in all of them) that Trump could win Rhode Island. I'm saying that if his model is so sensitive to a couple of polls that suggest a result that is not at all in line with the state's demographics and voting history, the model is clearly flawed.

What I think is going on is that he was always bearish on Trump in the primaries, and the idea of Trump winning any, let alone most of them, was laughable to him. Then, because he was proven wrong, he's gone on to basically create a model that is a glorified polling map (since the polls were "right" during the primaries), with more weight on recent polls and pollsters which he deems better. Back in 2012, even when polls showed a close race in Michigan, he argued that Michigan was no in play, due to fundamentals, the fact that neither campaign was seriously looking at it, and its voting history relative to the country as a whole. I guess this year, he's decided that he can't do that, which is a shame.

I'm not saying I could create a better model, but it's clear to me that his model has some issues, even the polls plus one.

[emailking]

I'm sick of people acting like the model is its own sentient entity. It's not. He created the model, so it represents what he thinks the polls suggest about the race. If the results of his model were absurd to him, he'd tweak it.

Predicting Trump wins California with 90% likelihood would be absurd. Predicting there's a chance Rhode Island could be within single digits is not absurd.

And I don't think he'd tweak the fundamentals of the model for this cycle at this point anyway.

He's not just predicting that there's a chance that Rhode Island could be within single digits, that the aggregate margin for all three models. According to all three models, there's also a chance (greater than 10% in all of them) that Trump could win Rhode Island. I'm saying that if his model is so sensitive to a couple of polls that suggest a result that is not at all in line with the state's demographics and voting history, the model is clearly flawed.

If you are right, and his model is flawed, and this RI poll is exposing it, then that does not mean he has "lost it" or anything remotely similar, as was your original assertion.

[elcorazon]

I think states not perceived to be in play which are not polled often have fairly large impacts from individual polls showing something outside the previous expectation. I think that kinda makes sense. It's about odds. The model doesn't mean RI WILL be close - it means that there is a chance that Trump could win... and with few data points and an actual poll showing it to be close, there is  some chance that the information in that poll is meaningful and has uncovered something about the race we didn't know before. If we have more polls showing that one to be an outlier, the model will adjust, but for now it can only react to the data fed into it. It may be true that Trump has impacted the model somewhat because he is not a mainstream Republican, but i don't think it's because his model didn't anticipate the primary results.

[Figueira]

I think the main issue is that Trump is such a weird candidate that we never know whether the RI polls for example are a result of polling error or actual high performance by Trump. Nobody said that those weird UT polls were just statistical noise.

[Lachi]

Guys, remember that even though he added all 50 polls, he weighted them in Trumps favor.

Nate Silver wants a horse race confirmed

[eric82oslo]

Guys, remember that even though he added all 50 polls, he weighted them in Trumps favor.

Nate Silver wants a horse race confirmed

The reason why the model weighed them in favour of Trump, was that the polling period was taken over a period of (more than) three weeks, in which period national and state polls have moved sharply in the direction of Trump, we're talking a shift of 4-5% towards Trump over the last 3-4 weeks.

[KingSweden]

A bounce for Clinton in all projections today