538 Model Megathread

[Erich Maria Remarque]

All 3 models now have NV for Clinton, FL NC and ME2 for Trump.  This matches my prediction exactly.

[BlueSwan]

NC just tilted back to Clinton, but they haven't added the Sienna poll yet, which might tip it back to Trump.

[BlueSwan]

As predicted, the Siena poll puts North Carolina back into Trumps column.

[Erich Maria Remarque]

Nice!

If Upshot "trends" are true nationally (whites ---> Trump), it will be a fun night
Both NC and FL showed them.

It is a pity that they didn't poll PA again to be totally sure.

[BlueSwan]

Trashy 50-state Surveymoney poll moves NC back to Clinton again - but by 50,1%. Closer than close!

[BlueSwan]

Sure, he bombed with Trump in the primaries - mainly because he suddenly played pundit instead of sticking to crunching numbers, which is what he is renowned for - but since then he has returned to number crunching and stayed away from over-confident predictions.

The guy is not pro-Trump in any way, like some are acting. Anybody who has listened to the 538 podcast should know this. He has just been careful in not ruling Trump out, which is very understandable after the primaries, and if you ask me, aligned with the actual data, which has shown Clinton ahead but not consistently enough to rule out a Trump win by any means - particularly with the uncertainties surrounding this strange election with a highly "unconventional" candidate and weird polls.

He gets flack for including polls that seem trashy - like the 50 state polls or the LA times poll - but basically he is just staying true to the methodology that predicted the 2012 and 2008 elections pretty accurately. People act as if they want Silver to subjectively throw out every poll that looks like an outlier, but that is exactly when you end up with a biased result.

People complain about the "house effect" calls and that's the one part I am sceptical about as well, but at least they are applying consistent methods in working this out.

Overall, it is just a model and a model that has proven to be pretty damn good in 2008 and 2012. Maybe it isn't so good this time around, maybe it is. That's the way it works. These models are only as good as the numbers they put into them. If the polling is off the models will be off. 538 is doing their stuff in a fairly transparent way.

[BlueSwan]

....aaand it is back to Trump again - 50,1%.

[Erc]

Sure, he bombed with Trump in the primaries - mainly because he suddenly played pundit instead of sticking to crunching numbers, which is what he is renowned for - but since then he has returned to number crunching and stayed away from over-confident predictions.

The guy is not pro-Trump in any way, like some are acting. Anybody who has listened to the 538 podcast should know this. He has just been careful in not ruling Trump out, which is very understandable after the primaries, and if you ask me, aligned with the actual data, which has shown Clinton ahead but not consistently enough to rule out a Trump win by any means - particularly with the uncertainties surrounding this strange election with a highly "unconventional" candidate and weird polls.

He gets flack for including polls that seem trashy - like the 50 state polls or the LA times poll - but basically he is just staying true to the methodology that predicted the 2012 and 2008 elections pretty accurately. People act as if they want Silver to subjectively throw out every poll that looks like an outlier, but that is exactly when you end up with a biased result.

People complain about the "house effect" calls and that's the one part I am sceptical about as well, but at least they are applying consistent methods in working this out.

Overall, it is just a model and a model that has proven to be pretty damn good in 2008 and 2012. Maybe it isn't so good this time around, maybe it is. That's the way it works. These models are only as good as the numbers they put into them. If the polling is off the models will be off. 538 is doing their stuff in a fairly transparent way.

Most of the most important things in the model (how he derives the correlation matrix between states, for example) are not transparent at all.  This makes business sense, of course, but makes it hard to assess it properly.

Otherwise, you're mostly right.  This is a model that worked well for the last couple elections (when we had very consistent and accurate state polling) but is failing now because it can't handle terrible polling and Donald Trump.  It's still doing as well as it is because of its proprietary secret sauce, and it is the most honest about the uncertainty of its forecast.

Its baseline is still way off though, but that's because the polls are.

[DavidB.]

Known fraud Nate is totally going to be discredited after his model proves to be waaaay too Republican-friendly. There is no way Trump has a 50% chance to win FL or NC, and there's also no way the GOP have more than a 50% chance to hold the Senate, as 538 currently predicts.

[Erich Maria Remarque]

Known fraud Nate is totally going to be discredited after his model proves to be waaaay too Republican-friendly. There is no way Trump has a 50% chance to win FL or NC, and there's also no way the GOP have more than a 50% chance to hold the Senate, as 538 currently predicts.

Because of non-representative early-voting?

Had early voting in 2014 not good "indications" for Dems as well?

[DavidB.]

1. The adjusting/unskewing of the polls, often in one direction.
2. It's not so much the model itself (though it is too GOP-friendly even if Silver himself is not; by saying it's close you can't be too wrong in any event) as his sweaty statements on Twitter on how the race is really much closer than people make it out to be. It's not, and pushing the horse race frame is just a cheap way to make people click on your site. That's clickbait, not real journalism. He has done this all the time.

[riceowl]

Uh, is he actively manipulating the model today to put it where he wants it? There's no polls I can see that are being input to make the states swing back to their more favorable HRC sides.

[Confused Democrat]

Uh, is he actively manipulating the model today to put it where he wants it? There's no polls I can see that are being input to make the states swing back to their more favorable HRC sides.

I wouldn't doubt it.

Didn't he flip Florida to Obama at the last second in 2012?

[Mallow]

lol at the conspiracy theories.

[jaichind]

The current models seems to have Clinton's lead at 3% which seems to match the medium of the national polls.  So if this site is off then it is more because of how it distributes the votes and not absolute   vote share.  I personally believe Clinton's lead is likely to be less than 3% and have a more Trump friendly distribution than 538.   

[Erich Maria Remarque]

lol at the conspiracy theories.

[BlueSwan]

Florida and North Carolina are now both back in the Clinton camp. Barely.

[ursulahx]

Enten tweeted that he agrees with Jon Favreau's prediction of C+5. That sounds a good deal more confident than 66% to me.

[Devout Centrist]

Nate flipped Nevada, NC, and Florida all to Clinton.

Fat lady is onstage, waiting for her cue.

[Crumpets]

Here's what gets me about Nate Silver's model this year:

In 2008 and 2012, empirical data about the votes in each state, such as race, age, and voting registration, largely bore out what we were already seeing in the polls at the time. This year is different, and the data from early votes is suggesting that at least in some states, Clinton will overperform her polls. Everyone who has taken any kind of science class knows that when your hypothesis (i.e. this is a close race with Trump slightly favored in NV, FL, and NC) is not supported by hard data, it is NOT because the data is wrong, it is because your hypothesis is wrong. Nate should look at the early vote data and tweak his model, however slightly, for the fact that his model does not agree with empirical voting data with regards to turnout of unlikely voters.

[ApatheticAustrian]

wat?

why?

cohn is influencing!

and the only new thing from NV was...emerson.

[Likely Voter]

As I  noted before, the sole reason the model had Trump up in NC, FL and NV was the "trend line adjustment" which was pushing  every state a bit more than 2 points in Trump's direction.  Today that has been cut in half and that flips those three states back to Clinton.  Now 538 aligns with all other models in terms of state wins, but still is more bullish on Trump in terms of margins and odds .

Was it purely data that changed the trend line in the last few hours?

[Erich Maria Remarque]

Was it purely data that changed the trend line in the last few hours?

Eh.. are you implying Nate did it manually? What else but data?

[elcorazon]

As I  noted before, the sole reason the model had Trump up in NC, FL and NV was the "trend line adjustment" which was pushing  every state a bit more than 2 points in Trump's direction.  Today that has been cut in half and that flips those three states back to Clinton.  Now 538 aligns with all other models in terms of state wins, but still is more bullish on Trump in terms of margins and odds .

Was it purely data that changed the trend line in the last few hours?

I will agree that occasionally the changes in the model seem quite difficult to explain and could be noise/trendlines/ etc. but you should also keep in mind all 3 of those states have been teetering back and forth. Slight changes can switch who is ahead quite easily. The truth is the polling indicates that Hillary is right on the edge between a razor thin electoral victory (if FL, NV and NC all go to Trump) and a fairly solid win (if they all go to Clinton). And if she gets FL and NC, suddenly, OH & IA and AZ seem plausible which is when it starts not to look that close at all.
Seeing NC go from 48% Clinton to 51% Clinton makes sense from the recent polling.

[IceSpear]

Nate "Discredited Fraud" Silver herding to the consensus the day before the election? Sad!