538 Model Megathread

[elcorazon]

polls only has clinton winning Arizona, but outside of arizona, all of the Clinton states including NC, FL, OH, IA, NH and NV (PA, too) are over 70%.

[elcorazon]

Iowa seems strange to me. 50/50 in nowcast, but in Polls Only Hillary is at 60 and in Polls Plus Hillary is at 53. Seems very strange to say if the election were held today it is a total tossup, yet, somehow JUST USING POLLS, Hillary has a 60% shot on election day??? thoughts?

[elcorazon]

Iowa seems strange to me. 50/50 in nowcast, but in Polls Only Hillary is at 60 and in Polls Plus Hillary is at 53. Seems very strange to say if the election were held today it is a total tossup, yet, somehow JUST USING POLLS, Hillary has a 60% shot on election day??? thoughts?

Polls Only model is not aggresive. On the election you in generelly don't give any weight at all for polls older than say 5-6 days. Simplied.

I'm not following. Polls Only model is based on ONLY POLLS. Are you saying Polls Only uses different polls than Nowcast? I would think if polls show if election were TODAY the results are 50/50, but if we wait 2 months, Hillary is more likely to win, something doesn't really make sense. I could deal with 53% or something like that, but 60% is pretty significantly different, no?

[elcorazon]

I've noticed sometimes the forecast updates without a new poll being inputted; I would assume this is a fundamentals and/or timeline adjustment (fewer days until the election thus less uncertainty).

they also sometimes correct errors in their updates, so that could explain some of the minor little changes that seem odd.

[elcorazon]

polls are tightening for sure. jobs report wasn't that bad. I'm nervous.

[elcorazon]

The three models have broadly converged. How long are they going to hedge their bets

bets?? what do you mean? The models are designed to show the possibilities. Race is still too close to call with Hillary clearly ahead but not by much.

[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.

[elcorazon]

Trump now at 41.3% in the Now-cast, winning NV, IA, OH, FL, NC and ME-02.

And NH is next to swing

CO is actually closer to swinging than NH according to the nowcast.

[elcorazon]

I'm not sure what polls are responsible, but 538's model has Trump still surging.

This has been puzzling me too, since the recent Trump surge seems to have peaked, but his chances keep getting better on 538.  Meanwhile, the Upshot model (which is generally slower to move) has tracked the poll movement pretty closely.  During the surge, Clinton's win probability in Upshot dropped from near 80% to 73%, but it's recovered in the past couple of days to 75%.

I'm starting to wonder if Nate Silver has tried to do too much in his models, and they're becoming less useful as a result.

maybe some of the better Clinton polls are fading in weighting as they get older, and even though some polls are showing a rebound by Clinton, the overall polling is still showing a very close race which keeps pushing the ultimate results closer to 50/50, especially as the more Clinton favoring polls get further back in time.

[elcorazon]

Agreed. Plus I disagree with notion his model has been wildly moving. The Polls Plus in particular, but Polls Only as well, were the steadiest for awhile as it was other models that had Hillary way ahead early on. It's really been that Trump is continually chipping away at Clinton's lead for so long the numbers have ultimately changed dramatically. If you're saying the race really hasn't changed despite the fact that polling has changed, you're talking about something other than a model based primarily on polling.

[elcorazon]

Because if, say, she loses NH (the current 538 tipping point state), then she actually has not just one but five backup plans: NV, FL, NC, OH, & IA, all of which are more likely (according to 538) to flip to Clinton than NH is to flip to Trump.

If she loses NH, that doesn't make it any more likely she'll win one of those 5 states then it already is right now. If anything, it's less likely. Losing New Hampshire is the assumption in this scenario. That's unlikely to happen, but if it does, she is more likely to lose than win (at least according to current polling).

I'm not sure that's true, but it doesn't really matter, because you can make that claim about dozens of states.

What matters in terms of preserving the firewall is how likely your weakest state is to be the potential tipping point state, i.e., to prove decisive in a close election.

Right now, New Hampshire isn't even in the top 10 on 538's Tipping Point state list:

1. Florida (16.8%)
2. Pennsylvania (11.6%)
3. Michigan (9.8%)
4. North Carolina (8.0%)
5. Ohio (7.9%)
6. Virginia (6.7%)
7. Colorado (6.1%)
8. Wisconsin (5.2%)
9. Minnesota (3.9%)
10. Nevada (3.4%)
11. New Hampshire (2.4%)

Note that ALL FOUR of Trump's weakest states (NV, FL, NC, OH)-- i.e., those where his expected margin of victory is lower than Clinton's in NH-- are higher on this list.

By definition, that means Clinton doesn't need NH as much as Trump needs those four states, so she has more robust -- AND less precarious-- paths to victory.

But that forgets the fact that #2 on the list is a Hillary State as is #3, 6, 7, 8 & 9. She needs ALL those states to hold her FREIWAL

[elcorazon]

But that forgets the fact that #2 on the list is a Hillary State as is #3, 6, 7, 8 & 9. She needs ALL those states to hold her FREIWAL

Not if one of states #1,2,4,5 on the list also flip-- and they're more likely to flip.  

That's my point.

Look, if the only way Clinton could recover from a breach in the firewall is to win states like AZ, GA, or MO, then I would agree she's really vulnerable and has a very limited path.

But if she does lose a firewall state, she can easily recover by winning NV, FL, NC, OH, or IA.

Which is why she has many paths to victory.

Why is that so hard to understand?

#2 is PA. She can't lose that one. She has more states near tipping point than he does

[elcorazon]

race is essentially tied.

[elcorazon]

looks to me like 2012 + NC + AZ + NE2 is the single most likely result on here... that seems odd

[elcorazon]

looks to me like 2012 + NC + AZ + NE2 is the single most likely result on here... that seems odd

Only in polls-only.  Now-cast has 2012+NC+NE2 (or other 348s), and polls-plus still has freiwal (or other 272s) as the most likely.

The reason bigger EV results spike is because we have a bunch of states that can move in concert. And the more of them flip, the more likely the next one along will flip.  If Clinton wins 2012+NC, her chances of winning AZ go up tremendously.  In fact, what polls-only is saying is that if Clinton gets at least 348 EVs, she is more likely to win AZ than to lose AZ.

If Clinton loses any of the "swing 5" (FL, NV, NC, IA, OH), her chances in AZ go way down.

Keep in mind as well, the chances of this scenario are only a bit over 1%.  The model has the chances of Clinton winning AZ at 21-28%.

yea. I get it. Part of my surprise is that AZ but not GA swinging to Hillary is more likely than both of them. I would almost have expected that the gap after Hillary takes the dicier states like Iowa and Ohio would be greater than the gap between Az and GA.

[elcorazon]

AK is now more likely to go to Hillary than GA on both polls plus and the nowcast and almost the same on polls only. Did I miss some Alaska polls?

[elcorazon]

I just noticed Arizona is now blue on the nowcast!!!!

[elcorazon]

Alaska is at 30% on Polls Only - higher than GA

[elcorazon]

Didn't Nate think Rhode Island could be a battleground state based off a single shoddy Emerson poll? lol

It's funny how despite how "complex and advanced" his model is that it can fall prey to the most obvious screw up in the book: taking a single poll seriously when there's no corroboration. Doubly so if the poll is from a shoddy firm. Triply so if the poll is completely out of line with what you'd expect.

It's all about probability. If you have a small number of polls, it is true that they, likely, get too much weight. More polls improves the accuracy of the model. As people discussed above, even a state like Alaska, which does have some polls, gives a larger chance of an upset because we can't "know" what's going to happen as well as we might in Georgia, even if Georgia, in all likelihood should be closer than Alaska (or anyway should be to the left of Alaska), because Alaska has more uncertainty, there is statistically a greater chance it could swing in an election about this close...similarly based on a tight race and a close poll, there was a decent chance RI could have been a surprising result. It makes sense to me, anyway.

And it should be noted one national poll won't swing the numbers that much... there needs to be corroboration as this could be an outlier, an odd moment in time, on the edge of the margin of error, etc... not to mention as Trump's chances drop, it is harder for the raw % change to be as great (moving from 22% to 20% is a pretty large move while going from 48% to 46% is a pretty small move).

[elcorazon]

I'm ashamed of this country

[elcorazon]

Silver's model underestimates the risk of a Trump Presidency at this point. I'd put it at best at a coin flip. I hate America.

[elcorazon]

sometimes I think we internalize our expectations before they're fully baked into the model. Perhaps we see a few national polls showing that the race is getting closer and let's say based on that, WI might be expected to be C + 3-5, then a poll comes in at C + 6 and we think that's good for Clinton. But the model may think that reinforces the tightening because before the tightening it might have expected to see C +8-9. It takes many polls to fully  alter the overall picture.

[elcorazon]

Now I think about it, when Silver went on a mini-rant during a recent podcast, emphasising how big he thought Trump's chance of winning was, he sounded genuinely scared. I wonder if he sees it as part of his mission to get Democrats to take the threat seriously and go out and vote. If so, I can't help feeling there are better ways of doing it.

he's a freedom fighter

[elcorazon]

Yeah. 538 and Upshot are the only media I can read in USA, and about USA.

I follow both Upshot and 538 teams on Twitter and they really hate/scared of Trump, their tweets are often emotional (but it is OK, it is their private thoughts) but nevertheless are their analysis always neutral as possible and based on data.

No one can be completely neutral, but the rest of media don't even try.

It is ridiculous and Breibart'ish to think that they tweaked their models towards any of candidates . Doesn't mean, the models can't be wrong.

I agree with this. sometimes the truth hurts. Although the problem with the rest of the media isn't always that they don't try to be neutral, it's that they don't really understand the polls and they are chasing ratings most of the time. Some are biased (Fox News!!) but most are just dumb/seeking ratings.

[elcorazon]

The development over the last few days is super scary. Clinton's win percentage:

Oct 29th: 81,1%
Oct 30th: 78,8%
Oct 31st: 75,2%
Nov 1st: 71,2%
Nov 2nd: 67,7%
Today: 64,7%

That is virtually a linear nosedive, losing an average of 3,28 points PER DAY. As a thought experiment, if the trend continued, on election day the model would have Clinton at 48,3%.

Trump wont win PA or Colorado  unless her chance drop to 42 percent

Harry Enten keeps saying she won't lose NV, which, if true, would add to Hillary's overall chances given the models have NV around 50/50 depending.