Good article. I agree with “ Biden doesn’t have it in the bag” but I’d push back on the polling isn’t broken and this is not the same as 2016.
RCP swing state aggregates are inching even closer, and FL NC AZ PA are within margin of error. Trump wins if he takes those states, and all 4 have a large NCW population which pollsters could be missing
“Trump is just a polling error of winning” just like last time, except it’s relevant for swing states this time as opposed to the national average. Yet 538 still has trump at 12%, which seems absurd to me.
But it's not that absurd. If Biden's national polling is 10+ then that would of course drive his chances higher. And of course Biden doesn't have "it in the bag", but to think he's not a heavy favorite overall would be to ignore the obvious.
4 years ago he had like a 3 in 10 chance of winning and got lucky. This time around he has literally been trailing Biden for 3 years in all polling. So 1 in 10 definitely sounds about right to me.
Correct. Which, of course, is not nothing.
I tend to have to restrain myself in these convos because I really just don't want to come off as an asshole when it comes to statistics and forecasts.
Except in stats you know the population. If someone is calling 2016 luck they missed the lead story about that election.
If you are predicting a roll of the dice, you know exactly how many combinations and probabilities.
If you are predicting the super bowl winner, you know exactly how many teams are in the competition.
But with trump there is a massive unknown that most polls, hundreds of them, got completely wrong - How many people in a specific demo will show up. NOT how they will vote, but should they even be included in the stat. And trumps base is the biggest demo by far, making a blanket polling error a distinct possibility, especially in swing states where his demo is largest.
Thats like calling dice and not knowing if they have six sides or 8. Or if there are 32 NFL teams or 42.
Last time 57 million non college whites voted. If 57.3 million show up this time, polls could be completely wrong. If 2% of trump supporters hate getting called for polls because the media s*cks, and lie, polls are wrong. It’s not like trump tells them to hate the media and establishment? In a football game, the results are verifiable. What a trumpster tells a pollster? No dice.
I am not saying that’s my prediction, just that 12% is absurd, considering HOW the 2016 victory occurred for him.
Umm....Jim Comey doing something no FBI director in the history of the country has ever done 10 days before the election, which DIRECTLY benefitted the guy who won, is the definition of Trump getting lucky. It is not going out on a limb to say that HRC would be president without that event happening. HIs campaign did not even expect to win.
Short of something like that happening again, and seeing as how the Hunter story is a flop, all available polling we have seems to be pretty damn accurate again (with the obvious caveat being Trump stealing the election and you can't really set up a model for that).
Yeah, turning to the Senate model @ The Economist. They have 74% chance of a Dem 51+ majority. 10% split evenly, 16% Repub majority. I think I might take the Dems' bet on that now.
Good article. I agree with “ Biden doesn’t have it in the bag” but I’d push back on the polling isn’t broken and this is not the same as 2016.
RCP swing state aggregates are inching even closer, and FL NC AZ PA are within margin of error. Trump wins if he takes those states, and all 4 have a large NCW population which pollsters could be missing
“Trump is just a polling error of winning” just like last time, except it’s relevant for swing states this time as opposed to the national average. Yet 538 still has trump at 12%, which seems absurd to me.
But it's not that absurd. If Biden's national polling is 10+ then that would of course drive his chances higher. And of course Biden doesn't have "it in the bag", but to think he's not a heavy favorite overall would be to ignore the obvious.
4 years ago he had like a 3 in 10 chance of winning and got lucky. This time around he has literally been trailing Biden for 3 years in all polling. So 1 in 10 definitely sounds about right to me.
Correct. Which, of course, is not nothing.
I tend to have to restrain myself in these convos because I really just don't want to come off as an asshole when it comes to statistics and forecasts.
Except in stats you know the population. If someone is calling 2016 luck they missed the lead story about that election.
If you are predicting a roll of the dice, you know exactly how many combinations and probabilities.
If you are predicting the super bowl winner, you know exactly how many teams are in the competition.
But with trump there is a massive unknown that most polls, hundreds of them, got completely wrong - How many people in a specific demo will show up. NOT how they will vote, but should they even be included in the stat. And trumps base is the biggest demo by far, making a blanket polling error a distinct possibility, especially in swing states where his demo is largest.
Thats like calling dice and not knowing if they have six sides or 8. Or if there are 32 NFL teams or 42.
Last time 57 million non college whites voted. If 57.3 million show up this time, polls could be completely wrong. If 2% of trump supporters hate getting called for polls because the media s*cks, and lie, polls are wrong. It’s not like trump tells them to hate the media and establishment? In a football game, the results are verifiable. What a trumpster tells a pollster? No dice.
I am not saying that’s my prediction, just that 12% is absurd, considering HOW the 2016 victory occurred for him.
Umm....Jim Comey doing something no FBI director in the history of the country has ever done 10 days before the election, which DIRECTLY benefitted the guy who won, is the definition of Trump getting lucky. It is not going out on a limb to say that HRC would be president without that event happening. HIs campaign did not even expect to win.
Short of something like that happening again, and seeing as how the Hunter story is a flop, all available polling we have seems to be pretty damn accurate again (with the obvious caveat being Trump stealing the election and you can't really set up a model for that).
It'll be interesting to see what other October surprises they have. The Hunter thing does pretty much seem to be a battle between staunch Trump fans and staunch Trump haters (though I am a bit nervous that it got some mainstream CBS exposure). But there's no way it's their last surprise.
1995 Milwaukee 1998 Alpine, Alpine 2003 Albany, Boston, Boston, Boston 2004 Boston, Boston 2006 Hartford, St. Paul (Petty), St. Paul (Petty) 2011 Alpine, Alpine 2013 Wrigley 2014 St. Paul 2016 Fenway, Fenway, Wrigley, Wrigley 2018 Missoula, Wrigley, Wrigley 2021 Asbury Park 2022 St Louis 2023 Austin, Austin
Good article. I agree with “ Biden doesn’t have it in the bag” but I’d push back on the polling isn’t broken and this is not the same as 2016.
RCP swing state aggregates are inching even closer, and FL NC AZ PA are within margin of error. Trump wins if he takes those states, and all 4 have a large NCW population which pollsters could be missing
“Trump is just a polling error of winning” just like last time, except it’s relevant for swing states this time as opposed to the national average. Yet 538 still has trump at 12%, which seems absurd to me.
But it's not that absurd. If Biden's national polling is 10+ then that would of course drive his chances higher. And of course Biden doesn't have "it in the bag", but to think he's not a heavy favorite overall would be to ignore the obvious.
4 years ago he had like a 3 in 10 chance of winning and got lucky. This time around he has literally been trailing Biden for 3 years in all polling. So 1 in 10 definitely sounds about right to me.
Correct. Which, of course, is not nothing.
I tend to have to restrain myself in these convos because I really just don't want to come off as an asshole when it comes to statistics and forecasts.
Except in stats you know the population. If someone is calling 2016 luck they missed the lead story about that election.
If you are predicting a roll of the dice, you know exactly how many combinations and probabilities.
If you are predicting the super bowl winner, you know exactly how many teams are in the competition.
But with trump there is a massive unknown that most polls, hundreds of them, got completely wrong - How many people in a specific demo will show up. NOT how they will vote, but should they even be included in the stat. And trumps base is the biggest demo by far, making a blanket polling error a distinct possibility, especially in swing states where his demo is largest.
Thats like calling dice and not knowing if they have six sides or 8. Or if there are 32 NFL teams or 42.
Last time 57 million non college whites voted. If 57.3 million show up this time, polls could be completely wrong. If 2% of trump supporters hate getting called for polls because the media s*cks, and lie, polls are wrong. It’s not like trump tells them to hate the media and establishment? In a football game, the results are verifiable. What a trumpster tells a pollster? No dice.
I am not saying that’s my prediction, just that 12% is absurd, considering HOW the 2016 victory occurred for him.
Umm....Jim Comey doing something no FBI director in the history of the country has ever done 10 days before the election, which DIRECTLY benefitted the guy who won, is the definition of Trump getting lucky. It is not going out on a limb to say that HRC would be president without that event happening. HIs campaign did not even expect to win.
Short of something like that happening again, and seeing as how the Hunter story is a flop, all available polling we have seems to be pretty damn accurate again (with the obvious caveat being Trump stealing the election and you can't really set up a model for that).
It'll be interesting to see what other October surprises they have. The Hunter thing does pretty much seem to be a battle between staunch Trump fans and staunch Trump haters (though I am a bit nervous that it got some mainstream CBS exposure). But there's no way it's their last surprise.
I find it shocking that Trump feels this would be the topic that resonates with voters given current events. Each day that he appears to self-sabotage, I get more concerned about him challenging the results of the election under a loss. At this point, I feel the odds are high for both a loss, and a challenge, and wonder if this is becoming his default plan.
'05 - TO, '06 - TO 1, '08 - NYC 1 & 2, '09 - TO, Chi 1 & 2, '10 - Buffalo, NYC 1 & 2, '11 - TO 1 & 2, Hamilton, '13 - Buffalo, Brooklyn 1 & 2, '15 - Global Citizen, '16 - TO 1 & 2, Chi 2
EV
Toronto Film Festival 9/11/2007, '08 - Toronto 1 & 2, '09 - Albany 1, '11 - Chicago 1
I never thought I'd be the proactive type on 'Merica stuff, but if this loon is getting trounced and he's clearly getting fucked up in the votes, yet he directs his acolyte loons to storm capitols and shit - I might actually have to show up to some of that and tell them to go fuck off.
I will say - this is the kind of stuff we're talking about:
Make our Republican presence known’
This is the first presidential election in nearly 40 years that the Republican National Committee has been free to sponsor such “ballot security” operations without permission from a federal court. A 1982 consent decree restricted these activities after the party sent teams of gun-toting men to minority neighbourhoods during a New Jersey election wearing uniforms saying “Ballot Security Task Force”.
That consent decree expired in 2018 and a federal judge declined Democratic attempts to renew it."
538 isn't a poll, it's a forecast. It's not a number based on a betting market. I don't know what else to say - all of these things have been accounted for: you cannot rely on bad polling, or this mysterious demo of Trump voters, or something like unaffiliated voters. All of these have been accounted for, and which are absurd on their head.
As been caveated many times - the only way Trump wins is through the EC, and it's with some kind of significant mishap - like votes go missing, there's a data breach, a huge sample size of ballots are rejected, etc etc. i.e. voter turnout is low.
Currently, 538 has the national outcome at a 8.4% margin of victory for Biden and PA at 6.3%. This seems like a pretty conservative take.
538 is a forecast based on the aggregate of polls-
538 adjusts polling results based on their assessment of the polls quality and other factors such as incumbency, the economy and demographics.
For PA, 538s aggregate is +6.1% Biden and RCPs aggregate this afternoon dropped to 3.8%. I am not sure why 538 is higher, and looking at recent polling it looks to be too high.
Looking at the last 6 PA polls listed on RCP (All reputable polls) the average Biden lead is 3.25%. Arizona’s aggregate is 3.1% Biden. FL and NC are smaller leads for Biden.
The bottom line is all of 4 of these states are within the margin of error (per RCP) meaning the risk the pollster makes a mistake projecting their sample to what they expect to be the actual vote is very possible. Yes the only way trump wins is thru the EC but since these states are within MoE, that means all it takes is a sampling error for the polls to be wrong...again...and trump to win.
Since trumps base is NCWs and NCWs are the largest voting block and these 4 states have large NCW voting share, the sampling risk is actually high when projecting the NCW vote to the actual estimated vote. This is what happened last time and looks possible this time. 12% Trump when the tipping point state looks to be 3%? That does not seem reasonable.
538 isn't a poll, it's a forecast. It's not a number based on a betting market. I don't know what else to say - all of these things have been accounted for: you cannot rely on bad polling, or this mysterious demo of Trump voters, or something like unaffiliated voters. All of these have been accounted for, and which are absurd on their head.
As been caveated many times - the only way Trump wins is through the EC, and it's with some kind of significant mishap - like votes go missing, there's a data breach, a huge sample size of ballots are rejected, etc etc. i.e. voter turnout is low.
Currently, 538 has the national outcome at a 8.4% margin of victory for Biden and PA at 6.3%. This seems like a pretty conservative take.
538 is a forecast based on the aggregate of polls-
538 adjusts polling results based on their assessment of the polls quality and other factors such as incumbency, the economy and demographics.
For PA, 538s aggregate is +6.1% Biden and RCPs aggregate this afternoon dropped to 3.8%. I am not sure why 538 is higher, and looking at recent polling it looks to be too high.
Looking at the last 6 PA polls listed on RCP (All reputable polls) the average Biden lead is 3.25%. Arizona’s aggregate is 3.1% Biden. FL and NC are smaller leads for Biden.
The bottom line is all of 4 of these states are within the margin of error (per RCP) meaning the risk the pollster makes a mistake projecting their sample to what they expect to be the actual vote is very possible. Yes the only way trump wins is thru the EC but since these states are within MoE, that means all it takes is a sampling error for the polls to be wrong...again...and trump to win.
Since trumps base is NCWs and NCWs are the largest voting block and these 4 states have large NCW voting share, the sampling risk is actually high when projecting the NCW vote to the actual estimated vote. This is what happened last time and looks possible this time. 12% Trump chance to win when the tipping point state looks to be 3% Biden lead? That does not seem reasonable.
I love how people think that if it ends up within the MOE, that it must move closer to Trump. lol Because...2016 of course!
Biden is comfortably ahead. He has been comfortably ahead for almost three years now. 10.7% at this point. That is unheard of for a challenger to an incumbent in recent times. So with that in mind, logic dictates that a poll like the latest one in WI tonight that has Biden up 6%, could easily be closer to Biden +10 than Biden +2. Either way...it's Biden +.something...
Also, I think because you are looking at an average of a bunch of polls, the MOE becomes smaller than looking at just one poll. Right?
Nothing is locked in, but there’s a reason that Trump was attacking Biden last year and keeping his hands off Sanders. I guarantee they had internal polling showing his strength and that’s only improved.
Nothing is locked in, but there’s a reason that Trump was attacking Biden last year and keeping his hands off Sanders. I guarantee they had internal polling showing his strength and that’s only improved.
Even so, all this idiot had to do was tell people to wear masks and head his health experts advice to have a MUCH better chance in November.
Nothing is locked in, but there’s a reason that Trump was attacking Biden last year and keeping his hands off Sanders. I guarantee they had internal polling showing his strength and that’s only improved.
Even so, all this idiot had to do was tell people to wear masks and head his health experts advice to have a MUCH better chance in November.
Fucking moron.
Absolutely this. Once I realized if this shithead did just like 5% more regarding the pandemic he'd probably still win. People just can't get it twisted that there's this vast swell of "good" in people coming out to "fight back" or something. Americans are garbage and as a whole we wouldn't have needed much to re-elect this waste of dogshit. Turns out he's done less than 0% and seems to be actively sabotaging himself.
I love how people think that if it ends up within the MOE, that it must move closer to Trump. lol Because...2016 of course!
Biden is comfortably ahead. He has been comfortably ahead for almost three years now. 10.7% at this point. That is unheard of for a challenger to an incumbent in recent times. So with that in mind, logic dictates that a poll like the latest one in WI tonight that has Biden up 6%, could easily be closer to Biden +10 than Biden +2. Either way...it's Biden +.something...
Also, I think because you are looking at an average of a bunch of polls, the MOE becomes smaller than looking at just one poll. Right?
Thats an interesting point. It looks like RCP is leaving off SurveyMonkey's daily tracking poll which has Biden +8
538 rates SM a D- poll yet this seems to be the major reason 538 is Biden +6 and RCP is +3.
PA is critical now as its 538s tipping point state. It looks to be +3 for pollsters rated D and above
The next tipping-point state is FL. I dont think Biden wants to pin his chances on FL.
Comparing apples to apples (RCP) 2020 PA looks reasonably comparable to 2016 at this point in time, and it is the tipping point state. Hillary was actually doing better. Someone mentioned Comey letter hurt Clinton here and I agree, but I think there is a good chance trump and/or the senate comes up with some BS in the next 2 weeks that could do something to the polls. Early voting could save the Dems. Hopefully.
If Biden misses in PA, the next tipping point is FL.
No way FL is anything other than 50/50 and no way this is an 87% chance Biden election. I can’t explain how trump gets covid and says he wants everyone to get it and the polls tighten, but he did and they did.
Prediction- by Election Day 538 will have no choice but to have their odds fairly close to Clinton’s 2016 chance of 70%.
Edit, AZ just jumped ahead of FL on the snake tipping point chart. Not sure if that’s good news or bad, because a few months ago Biden had a sizable FL lead.
I completely agree. This is scary news. 4 conservative justices who claim to believe in states rights and limited govt decided to intervene and override a critical state court ruling.
I have no doubt Barrett would join the conservatives in key election decisions once she is on there. Roberts is voting with Dems only to protect the reputation of his court for the history books, like he has done often since Kennedy retired.
This election is getting close and it may take Pelosi fighting SCOTUS with ever trick in her book if there is a key party line SCOTUS vote
Comparing apples to apples (RCP) 2020 PA looks reasonably comparable to 2016 at this point in time, and it is the tipping point state. Hillary was actually doing better. Someone mentioned Comey letter hurt Clinton here and I agree, but I think there is a good chance trump and/or the senate comes up with some BS in the next 2 weeks that could do something to the polls. Early voting could save the Dems. Hopefully.
If Biden misses in PA, the next tipping point is FL.
No way FL is anything other than 50/50 and no way this is an 87% chance Biden election. I can’t explain how trump gets covid and says he wants everyone to get it and the polls tighten, but he did and they did.
Prediction- by Election Day 538 will have no choice but to have their odds fairly close to Clinton’s 2016 chance of 70%.
Edit, AZ just jumped ahead of FL on the snake tipping point chart. Not sure if that’s good news or bad, because a few months ago Biden had a sizable FL lead.
There's no doubt that PA is tightening. Recent polls have shown that. It's going to come down to turnout.
It is tightening a little. But didn't we expect that after Trump's horrible debate performance and covid diagnosis. It was eventually gonna tight a little as that stuff faded. But RCP is laughable at 3.8%. Biden's lead is about 2% higher than it was a few months ago and waaay more steadier than HRC's lead was (at this point in '16 she was still basking in the glow of the access hollywood tape bump a few days prior to the Comey bombshell--that stuff is not happening this time around).
And 538 includes Quinnipiac which is a very pro democratic poll, so it could go both ways.
RCP had Clinton +1.7 last time in PA and 538 was +3.7, so if I were Nate I wouldn’t be so dismissive about RCP. I’d say RCP a better job In PA last time. I hope Quinnipiac is right, but my expectation in an 87% chance race is that the tipping point state would have a more comfortable lead, one that’s nearly as solid as the national. I don’t think Nate can make that claim .
I’ve been chirping at Nate on social media about PA the last couple days, glad he’s paying attention
And 538 includes Quinnipiac which is a very pro democratic poll, so it could go both ways.
RCP had Clinton +1.7 last time in PA and 538 was +3.7, so if I were Nate I wouldn’t be so dismissive about RCP. I’d say RCP a better job In PA last time. I hope Quinnipiac is right, but my expectation in an 87% chance race is that the tipping point state would have a more comfortable lead, one that’s nearly as solid as the national. I don’t think Nate can make that claim .
I’ve been chirping at Nate on social media about PA the last couple days, glad he’s paying attention
While 538 includes Quinnipiac along with other polls which tend to lean a specific direction, they do adjust for biases to counteract. Add to that the wisdom of the crowd benefit by aggregating across different polls.
'05 - TO, '06 - TO 1, '08 - NYC 1 & 2, '09 - TO, Chi 1 & 2, '10 - Buffalo, NYC 1 & 2, '11 - TO 1 & 2, Hamilton, '13 - Buffalo, Brooklyn 1 & 2, '15 - Global Citizen, '16 - TO 1 & 2, Chi 2
EV
Toronto Film Festival 9/11/2007, '08 - Toronto 1 & 2, '09 - Albany 1, '11 - Chicago 1
Comments
Short of something like that happening again, and seeing as how the Hunter story is a flop, all available polling we have seems to be pretty damn accurate again (with the obvious caveat being Trump stealing the election and you can't really set up a model for that).
It'll be interesting to see what other October surprises they have. The Hunter thing does pretty much seem to be a battle between staunch Trump fans and staunch Trump haters (though I am a bit nervous that it got some mainstream CBS exposure). But there's no way it's their last surprise.
2013 Wrigley 2014 St. Paul 2016 Fenway, Fenway, Wrigley, Wrigley 2018 Missoula, Wrigley, Wrigley 2021 Asbury Park 2022 St Louis 2023 Austin, Austin
EV
Toronto Film Festival 9/11/2007, '08 - Toronto 1 & 2, '09 - Albany 1, '11 - Chicago 1
Make our Republican presence known’
This is the first presidential election in nearly 40 years that the Republican National Committee has been free to sponsor such “ballot security” operations without permission from a federal court. A 1982 consent decree restricted these activities after the party sent teams of gun-toting men to minority neighbourhoods during a New Jersey election wearing uniforms saying “Ballot Security Task Force”.
That consent decree expired in 2018 and a federal judge declined Democratic attempts to renew it."
https://www.aljazeera.com/news/2020/10/15/army-for-trump-preps-poll-watching-operation-raising-concerns
That's kind of a big deal.
Since trumps base is NCWs and NCWs are the largest voting block and these 4 states have large NCW voting share, the sampling risk is actually high when projecting the NCW vote to the actual estimated vote. This is what happened last time and looks possible this time. 12% Trump when the tipping point state looks to be 3%? That does not seem reasonable.
Since trumps base is NCWs and NCWs are the largest voting block and these 4 states have large NCW voting share, the sampling risk is actually high when projecting the NCW vote to the actual estimated vote. This is what happened last time and looks possible this time. 12% Trump chance to win when the tipping point state looks to be 3% Biden lead? That does not seem reasonable.
.
Biden is comfortably ahead. He has been comfortably ahead for almost three years now. 10.7% at this point. That is unheard of for a challenger to an incumbent in recent times. So with that in mind, logic dictates that a poll like the latest one in WI tonight that has Biden up 6%, could easily be closer to Biden +10 than Biden +2. Either way...it's Biden +.something...
Also, I think because you are looking at an average of a bunch of polls, the MOE becomes smaller than looking at just one poll. Right?
Fucking moron.
Thats an interesting point. It looks like RCP is leaving off SurveyMonkey's daily tracking poll which has Biden +8
538 rates SM a D- poll yet this seems to be the major reason 538 is Biden +6 and RCP is +3.
PA is critical now as its 538s tipping point state. It looks to be +3 for pollsters rated D and above
The next tipping-point state is FL. I dont think Biden wants to pin his chances on FL.
If Biden misses in PA, the next tipping point is FL.
Prediction- by Election Day 538 will have no choice but to have their odds fairly close to Clinton’s 2016 chance of 70%.
I have no doubt Barrett would join the conservatives in key election decisions once she is on there. Roberts is voting with Dems only to protect the reputation of his court for the history books, like he has done often since Kennedy retired.
This election is getting close and it may take Pelosi fighting SCOTUS with ever trick in her book if there is a key party line SCOTUS vote
RCP had Clinton +1.7 last time in PA and 538 was +3.7, so if I were Nate I wouldn’t be so dismissive about RCP. I’d say RCP a better job In PA last time. I hope Quinnipiac is right, but my expectation in an 87% chance race is that the tipping point state would have a more comfortable lead, one that’s nearly as solid as the national. I don’t think Nate can make that claim .
EV
Toronto Film Festival 9/11/2007, '08 - Toronto 1 & 2, '09 - Albany 1, '11 - Chicago 1