538 model & poll tracker thread

[Annatar]

So they just added a bunch of those USC polls. Biden's polling lead jumped back over 9 points and Trump's at his lowest support in over a month. Yet the model didn't move in this mornings update. I really don't understand how it works anymore.

Update: The new "social desirability bias" Trafalgar poll of Michigan affected the model more than weeks worth of national polls showing Biden up in the double digits.

The model doesn't work off national polls, it is a state based model.

[Gass3268]

So they just added a bunch of those USC polls. Biden's polling lead jumped back over 9 points and Trump's at his lowest support in over a month. Yet the model didn't move in this mornings update. I really don't understand how it works anymore.

Update: The new "social desirability bias" Trafalgar poll of Michigan affected the model more than weeks worth of national polls showing Biden up in the double digits.

The model doesn't work off national polls, it is a state based model.

That's not entirely true. There have been days where Biden's % has dropped because his national numbers have dropped.

[Annatar]

So they just added a bunch of those USC polls. Biden's polling lead jumped back over 9 points and Trump's at his lowest support in over a month. Yet the model didn't move in this mornings update. I really don't understand how it works anymore.

Update: The new "social desirability bias" Trafalgar poll of Michigan affected the model more than weeks worth of national polls showing Biden up in the double digits.

The model doesn't work off national polls, it is a state based model.

That's not entirely true. There have been days where Biden's % has dropped because his national numbers have dropped.

Well the national polls are used for trend adjustment it's true, but the effect is really minor, almost non-existent, for all intents and purposes the model basically runs on state polls.

[Landslide Lyndon]

So they just added a bunch of those USC polls. Biden's polling lead jumped back over 9 points and Trump's at his lowest support in over a month. Yet the model didn't move in this mornings update. I really don't understand how it works anymore.

Update: The new "social desirability bias" Trafalgar poll of Michigan affected the model more than weeks worth of national polls showing Biden up in the double digits.

The model doesn't work off national polls, it is a state based model.

If the model ignores both national polls and Trump's approval ratings, then the model is sh**t.

[GeorgiaModerate]

https://twitter.com/KSoltisAnderson/status/1299363547730849795

[Fmr. Gov. NickG]

Biden keeps going up in the polls but down in the projections because look at these fonts!

[Crumpets]

Biden keeps going up in the polls but down in the projections because look at these fonts!

At the very least, Silver should recognize that certain long-anticipated events such as convention acceptance speeches will generate large headlines without leading to the same level of newly-introduced uncertainty that something like the Comey letter or Access Hollywood would generate with similar-sized headlines.

[DaleCooper]

I don't understand this model and I haven't been paying any attention to it. I'm not interested in Nate Silver's gut feeling on the trajectory of the race. Just let the data stand on its own.

[Skye]

The model now shows decimals when you toggle the popular vote.

Current popular vote prediction is Biden 52.4 - Trump 46.3

[Landslide Lyndon]

The model now shows decimals when you toggle the popular vote.

Current popular vote prediction is Biden 52.4 - Trump 46.3

Has anyone questioned Silver why his model takes for granted that the race will tighten?
I mean there are plenty of cases where the opposite happened.

[Fmr. Gov. NickG]

The model now shows decimals when you toggle the popular vote.

Current popular vote prediction is Biden 52.4 - Trump 46.3

Has anyone questioned Silver why his model takes for granted that the race will tighten?
I mean there are plenty of cases where the opposite happened.

It's not necessarily that he thinks the race will tighten.  He thinks the race will converge on the "fundamentals", and right now the "fundamentals" (especially I believe the stock market) predict a tossup race.

Right now, the predictions for each state are only 40-50% based on the polls, and 50-60% based on everything else Silver has decided might be relevant.

[GeorgiaModerate]

The model now shows decimals when you toggle the popular vote.

Current popular vote prediction is Biden 52.4 - Trump 46.3

Has anyone questioned Silver why his model takes for granted that the race will tighten?
I mean there are plenty of cases where the opposite happened.

It's not necessarily that he thinks the race will tighten.  He thinks the race will converge on the "fundamentals", and right now the "fundamentals" (especially I believe the stock market) predict a tossup race.

Right now, the predictions for each state are only 40-50% based on the polls, and 50-60% based on everything else Silver has decided might be relevant.

This is one of the fundamental (no pun intended) differences between the 538 and Economist models.  538 assumes the fundamentals indicate a close race, while the Economist assumes that the fundamentals suggest a significant Biden lead.

[Skye]

The model now shows decimals when you toggle the popular vote.

Current popular vote prediction is Biden 52.4 - Trump 46.3

Has anyone questioned Silver why his model takes for granted that the race will tighten?
I mean there are plenty of cases where the opposite happened.

FWIW, he also tweeted this:

https://twitter.com/NateSilver538/status/1294372448335532034

[Landslide Lyndon]

The model now shows decimals when you toggle the popular vote.

Current popular vote prediction is Biden 52.4 - Trump 46.3

Has anyone questioned Silver why his model takes for granted that the race will tighten?
I mean there are plenty of cases where the opposite happened.

It's not necessarily that he thinks the race will tighten.  He thinks the race will converge on the "fundamentals", and right now the "fundamentals" (especially I believe the stock market) predict a tossup race.

The fundamentals are hundreds of thousands of dead, record unemployment, the worst civil unrest in more than 50 years, and a chronically unpopular president whose approval is hovering around 40%. I'll believe that the race will tighten when I see it.  

[Crumpets]

https://twitter.com/NateSilver538/status/1299449209402396675

[Interlocutor is just not there yet]

It's crazy how 4% makes the difference between Trump winning a squeaker & Biden having the best election performance since Bush '88

[Obama-Biden Democrat]

It's crazy how 4% makes the difference between Trump winning a squeaker & Biden having the best election performance since Bush '88

Democrats are on the wrong side of a delayed realignment. 2016 showed huge collapses in the Midwest, there was a giant counter trend in the Sun Belt, but TX,GA,AZ were too solidly Republican to flip in one election cycle. WI, PA and MI were flippable as they were only tilt/lean D, not safe like the Republican Sunbelt

For those who prefer the SUNBELT STACK, it is possible in a +8 NPV Biden landslide.

[wbrocks67]

If his model is really changing that much because of Trafalgar polls, then it’s junk

[Crumpets]

Just got this beautiful realignment map. Note that even ME-1 goes for Trump.

Biden 271
Trump 267

[Ferguson97]

Just got this beautiful realignment map. Note that even ME-1 goes for Trump.

Biden 271
Trump 267

Oh god I can imagine the pundits salivating already.

But seriously, imagine the panic if Minnesota, New Hampshire, all of Maine, Wisconsin, and Pennsylvania vote for Trump... only for Biden to turn it around and win Arizona, Florida, and Michigan. Talk about mood wiplash.

[Interlocutor is just not there yet]

Hell, imagine if that's Biden's map - Michigan + Georgia

So long Rust Belt, hello Sun Belt!

[Crumpets]

Hell, imagine if that's Biden's map - Michigan + Georgia

So long Rust Belt, hello Sun Belt!

It would even work without GA or MI if he wins NC, although he'd have zero room for error with faithless electors.

[Kuumo]

Hell, imagine if that's Biden's map - Michigan + Georgia

So long Rust Belt, hello Sun Belt!

It would even work without GA or MI if he wins NC, although he'd have zero room for error with faithless electors.

That would be horrible for the senate unless trends aren't real for Tina Smith, Gary Peters, and Jeanne Shaheen. I would think that the electors would stick with Biden because if one of them voted for someone else, they would receive numerous death threats. Imagine if this happened:

The final election results look like this.

President
Biden/Harris: 270 EV
Trump/Pence: 268 EV

Senate
Democrats: 50 (+3)
Republicans: 50 (-3)
(Republicans win GA-special)


And then when the electoral college votes, we get this:
Biden/Harris: 269 EV
Trump/Pence: 268 EV
Other/Other: 1 EV

The House delegations elect Trump president and the Senate deadlocks 50-50. Pence tries to cast a tie-breaking vote, which is challenged by Democrats in the courts. When it reaches the Supreme Court, the ruling is in favor of Pence, who is then re-elected vice president by the Senate and his own tie-breaking vote. Mitch McConnell is then re-elected Majority Leader by the Senate.

One crazy Biden-state elector changes the outcome of the election from a Democratic trifecta to four more years of Trump/Pence and at least two more years of Majority Leader McConnell.

[Ferguson97]

Hell, imagine if that's Biden's map - Michigan + Georgia

So long Rust Belt, hello Sun Belt!

It would even work without GA or MI if he wins NC, although he'd have zero room for error with faithless electors.

That would be horrible for the senate unless trends aren't real for Tina Smith, Gary Peters, and Jeanne Shaheen. I would think that the electors would stick with Biden because if one of them voted for someone else, they would receive numerous death threats. Imagine if this happened:

The final election results look like this.

President
Biden/Harris: 270 EV
Trump/Pence: 268 EV

Senate
Democrats: 50 (+3)
Republicans: 50 (-3)
(Republicans win GA-special)


And then when the electoral college votes, we get this:
Biden/Harris: 269 EV
Trump/Pence: 268 EV
Other/Other: 1 EV

The House delegations elect Trump president and the Senate deadlocks 50-50. Pence tries to cast a tie-breaking vote, which is challenged by Democrats in the courts. When it reaches the Supreme Court, the ruling is in favor of Pence, who is then re-elected vice president by the Senate and his own tie-breaking vote. Mitch McConnell is then re-elected Majority Leader by the Senate.

One crazy Biden-state elector changes the outcome of the election from a Democratic trifecta to four more years of Trump/Pence and at least two more years of Majority Leader McConnell.

The riots that would follow this would make those first few nights of Floyd look like nothing...

[wbrocks67]

The national polling average really dropped an *ENTIRE POINT* because of the Zogby poll. Really.