538 model & poll tracker thread

[wbrocks67]

Biden only has an 84% chance of winnning in Colorado? What type of f**kery is this?

I would not be surprised to find out Nate tinkered with it b/c of 2016 and it's underselling Biden

[Pericles]

My guess is it's underrating Biden. Some of the key factors that made 2016 uncertain and a low estimate of Clinton's chances appropriate-a lot of polling volatility, and high numbers of undecideds and third party voters-are not applicable this year. It seems like the 2012 race is a much better parallel, except Biden's lead is much larger than Obama's.

[VAR]

It’s Way Too Soon To Count Trump Out by Nate Silver

Take what happens if we lie to our model and tell it that the election is going to be held today. It spits out that Biden has a 93 percent chance of winning.

[wbrocks67]

So..

the most recent Arizona polls: Biden +6, Trump +1, Biden +4, Biden +4, Biden +7, Biden +2, Biden +8, Biden +5

and Arizona's recent proclivity to underestimate Dems in polls.

538's forecast? Biden +1

Huh?

[wbrocks67]

It’s Way Too Soon To Count Trump Out by Nate Silver

Take what happens if we lie to our model and tell it that the election is going to be held today. It spits out that Biden has a 93 percent chance of winning.

I think the problem that the pundits are encountering is them all assuming the "race will tighten" towards Trump by Election Day.

Yes, that *COULD* happen. But I'm not sure why that is the sure-fire CW, when it could easily go both ways.

[wbrocks67]

The forecast really thinks GA is only gonna move 2% to the left since 2016, and Trump is gonna have a bigger margin than Kemp in 2018.

Yeah, Nate juiced this b/c he was scared after 2016. For sure.

[Penn_Quaker_Girl]

It’s Way Too Soon To Count Trump Out by Nate Silver

Take what happens if we lie to our model and tell it that the election is going to be held today. It spits out that Biden has a 93 percent chance of winning.

I think the problem that the pundits are encountering is them all assuming the "race will tighten" towards Trump by Election Day.

Yes, that *COULD* happen. But I'm not sure why that is the sure-fire CW, when it could easily go both ways.

I think it's generally based on the statement: "Since and including 1976, polling averages rarely trend in the direction of the candidate leading at this stage in the cycle".  

https://mobile.twitter.com/baseballot/status/1293214167231594496?ref_src=twsrc%5...

In all but two election cycles since 1976 (Obama in 2008 & 2012), the nat'l polling average at this point in the race tightened in the direction of the trailing candidate.  In some cases, the swing from the 11th of August to Election Day was quite dramatic (such as Carter in 1976 - - he experienced around a 20 pt. swing).  In others though - - such as Reagan in 1984 & 1988, it wasn't nearly enough to matter.

It *is* worth noting that the leading candidate at this stage has gone on to win in the general in ~73% (8/11) election cycles.

And before anybody hits the panic button or starts gloating about "the polls are wrong again/shy Trump voters/Trump bump incoming!!!", it's imperative to restate that this is not a normal election cycle (and that also goes for anybody who believes Biden is a shoo-in, too).  Neither candidate has been able to campaign in a conventional manner and a US Presidential election has never been run during a pandemic of this scale.

[Gustaf]

I'm wondering if a reason they waited is that they wanted to know the VP in order to incorporate their (tiny) VP homestate bonus.

[TrendsareUsuallyReal]

What a clown. He expects Trump to do 5% better in Texas than his polling?

[Granite City]

Already finding myself annoyed with the "Baby's First Election Forecast" theme they have gone with.

[DaleCooper]

It's absolutely astounding to me that this model places Biden at a 78% chance to win Maine and just over 70% to win Minnesota.

[Panhandle Progressive]

According to this model: Florida is the #1 tipping state, about a 23% chance of deciding the election, while Penn is #2 about 22%. If you missed this, look for the Tipping Points tab option under The winding path to victory section.

[Person Man]

My guess is that they won't start making serious predictions until the last month or until there are fewer than 5% swing voters?

[Sadader]

Silver is struggling with the fact that he’s lost his (easily imitable) competitive advantage from averaging & adjusting polls. Just wait until ~40 days from the election to release it. It’s utterly useless now.

The model is assuming that the economy will be way better in November! LMAO. Let’s not throw a bunch of garbage uncertainty in to save face and then just imagine it’ll be (comparatively) easy to project where the economy’ll be!

[Granite City]

Shading represents probability: 90% shade = 90%+ chance of winning and so on

That Hendrix poll obviously did some damage as Biden is more likely to win Arkansas than Kansas on the forecast!

[ProgressiveModerate]

Too much uncertainty and too bullish on Trump. Trump doesn't have a 2% chance of winning HI and Biden doesn't have a 2% chance in ND. Also; wish they had a national map and a senate forecast. I'm underwhelmed considering the wait. Biden at 7% in AR!!!??? Biden at 13% in MS!? I'll take JHK over this any day. This forecast relies too heavily on speculation, and is very underwhelming considering the wait. And this model assumes a free and fair election. On the bright side, the graphics are good.

I still really want a high quality senate model for this cycle; literally no one but JHK has made a senate model, and the senate will be close.

[Senator Spark]

Slightly more fair to Trump this time but still biased.

[ultraviolet]

This model is really awful. Biden 72% in MN but 73% in PA? 84% in CO? Trump 10% to win in NM???

Cmon nate you can do better

[BlueSwan]

OK, I have usually been backing Silver since way back, but this time he totally lost the plot. It is very very clear to me that he did not want to release a model that had Biden at 90% of winning or something like that, so that's probably why he has been tinkering with it for sooo long. This model seems to produce radically different kind of results compared to the models of 2016 or 2012. Remember that in 2012 the model was pretty confident in Obama despite very narrow leads everywhere. Biden leads by much healthier margins than Obama in 12 and Clinton in 16. This makes no sense.

[Person Man]

OK, I have usually been backing Silver since way back, but this time he totally lost the plot. It is very very clear to me that he did not want to release a model that had Biden at 90% of winning or something like that, so that's probably why he has been tinkering with it for sooo long. This model seems to produce radically different kind of results compared to the models of 2016 or 2012. Remember that in 2012 the model was pretty confident in Obama despite very narrow leads everywhere. Biden leads by much healthier margins than Obama in 12 and Clinton in 16. This makes no sense.

He might just be making it more fluid this far back in the season or he could be heavily relying on Trump's "charisma".

[afleitch]

I'm actually all for at least one model that perhaps over states Trump's chances.

[ProgressiveModerate]

Here's a full map of the forecast BTW

[ProgressiveModerate]

I'm actually all for at least one model that perhaps over states Trump's chances.

The issue isn't really that it overates Trump's chances, but that the level of uncertainty is pretty high. If there was less uncertainty Trump's chances would be lower. It kind of seems to me that Nate went out of his way to make every possible adjustment to make the model look more fair

[ProgressiveModerate]

OK, I have usually been backing Silver since way back, but this time he totally lost the plot. It is very very clear to me that he did not want to release a model that had Biden at 90% of winning or something like that, so that's probably why he has been tinkering with it for sooo long. This model seems to produce radically different kind of results compared to the models of 2016 or 2012. Remember that in 2012 the model was pretty confident in Obama despite very narrow leads everywhere. Biden leads by much healthier margins than Obama in 12 and Clinton in 16. This makes no sense.

I 100% agree

[ElectionsGuy]

Here is the trend implied in every state and district for his median popular vote estimates. There seems to be a little bit of underestimating of partisan margin effect again but overall very stable from 2016.

AL: D+2
AK: -
AZ: -
AR: D+7
CA: R+5
CO: -
CT: -
DE: D+7
DC: R+7
FL: R+1
GA: R+3
HI: R+4
ID: R+2
IL: R+1
IN: -
IA: -
KS: D+2
KY: D+5
LA: R+1
ME: -
 ME1: R+1
 ME2: D+1
MD: R+4
MA: R+2
MI: D+2
MN: R+1
MS: D+1
MO: D+4
MT: D+3
NE: -
 NE1: D+3
 NE2: R+2
 NE3: D+2
NV: R+1
NH: D+2
NJ: R+2
NM: R+1
NY: R+2
NC: R+1
ND: D+6
OH: D+2
OK: D+3
OR: R+2
PA: D+1
RI: D+3
SC: -
SD: D+4
TN: D+4
TX: R+1
UT: R+4
VT: D+4
VA: D+1
WA: D+3
WV: D+3
WI: D+1
WY: -