Nate Silver 538
Comments
-
dignin said:So, are you guys telling me it's not 2016?
https://www.reddit.com/r/fivethirtyeight/comments/jhxv9e/fivey_responds_to_criticism_of_the_2016_model/
0 -
Lerxst1992 said:Jearlpam0925 said:mrussel1 said:Lerxst1992 said:mrussel1 said:Lerxst1992 said:mrussel1 said:Lerxst1992 said:I’ll say although I hate to admit it trafalgar was right last time, and hear back about a broken clock, perhaps a tune instead.
https://www.youtube.com/watch?v=BxlYSV2dMY8It’s about the harm handicappers could be doing to our elections.This is a topic about 538 and I recall the debate started with a few trump supporters meandering on here stating how these forecasts are absurd given what actually happened in 2016. Then the talk turned to WI, and how it was “solid.” So I joined in, because I think the long departed trump supporters were making an accurate point. I put up side by side images of 538s site 2016 vs now for WI, which looked nearly identical (factoring in 2 former GOP govs running a legit 3rd party campaign). I wouldn’t call that solid at that time.As a Biden supporter, I’ve been very critical of 538s forecasts and believe I even got Nates attention on social media a week ago about his 87% Biden in PA. That just seems outright absurd given the polling then. And even the polling itself shouldn’t be fully trusted. But, looking at the huge increase in covid cases the last 2 days, it would be catastrophic for the forecasts to be wrong. If they create a few thousand “lazy blue non voters” in key states, does Silver and the others take any accountability if their processes are wrong again? I am hoping for a Biden landslide, but seeing many key states within MoE given how absurdly bad trump is doing...I said 538 did a poor job. They deserve the blame for the job that they did. And it does not matter how often their readers vote. It is literally their job is to figure out who is showing up to vote, and which polls captured that the best. And it’s concerning they are taking the current election where many of the swing states are MoE and forecasting it at 87%. It’s just an opinion. You are welcome to have a different one.Did either of you bother to review their process?
Of course 538 tries to figure out turnout. One party wins when turnout is high, he other wins when it isn’t. Their forecast would have zero chance of being close to accurate without being able to predict the approximate makeup and size of the electorate
Of course they will be wrong a percentage of the time. It’s the 12% next to trumps name that is absurd.
538 has about ten references to turnout in their process, here are a few:
“The national snapshot accounts for projected voter turnout in each state based on population growth since 2016, changes in how easy it is to vote since 2016, and how close the race is in that state currently — closer-polling states tend to have higher turnout. National polls are not used in the national snapshot; it’s simply a summation of the snapshots in the 50 states and Washington, D.C.“There could also be some challenges related to polling during COVID-19. In primary elections conducted during the pandemic, for instance, turnout was hard to predict. In some ways, the pandemic makes voting easier (expanded options to vote by mail in many states), but it also makes it harder in other ways (it’s difficult to socially distance if you must vote in person).0 -
mrussel1 said:Lerxst1992 said:Jearlpam0925 said:mrussel1 said:Lerxst1992 said:mrussel1 said:Lerxst1992 said:mrussel1 said:Lerxst1992 said:I’ll say although I hate to admit it trafalgar was right last time, and hear back about a broken clock, perhaps a tune instead.
https://www.youtube.com/watch?v=BxlYSV2dMY8It’s about the harm handicappers could be doing to our elections.This is a topic about 538 and I recall the debate started with a few trump supporters meandering on here stating how these forecasts are absurd given what actually happened in 2016. Then the talk turned to WI, and how it was “solid.” So I joined in, because I think the long departed trump supporters were making an accurate point. I put up side by side images of 538s site 2016 vs now for WI, which looked nearly identical (factoring in 2 former GOP govs running a legit 3rd party campaign). I wouldn’t call that solid at that time.As a Biden supporter, I’ve been very critical of 538s forecasts and believe I even got Nates attention on social media a week ago about his 87% Biden in PA. That just seems outright absurd given the polling then. And even the polling itself shouldn’t be fully trusted. But, looking at the huge increase in covid cases the last 2 days, it would be catastrophic for the forecasts to be wrong. If they create a few thousand “lazy blue non voters” in key states, does Silver and the others take any accountability if their processes are wrong again? I am hoping for a Biden landslide, but seeing many key states within MoE given how absurdly bad trump is doing...I said 538 did a poor job. They deserve the blame for the job that they did. And it does not matter how often their readers vote. It is literally their job is to figure out who is showing up to vote, and which polls captured that the best. And it’s concerning they are taking the current election where many of the swing states are MoE and forecasting it at 87%. It’s just an opinion. You are welcome to have a different one.Did either of you bother to review their process?
Of course 538 tries to figure out turnout. One party wins when turnout is high, he other wins when it isn’t. Their forecast would have zero chance of being close to accurate without being able to predict the approximate makeup and size of the electorate
Of course they will be wrong a percentage of the time. It’s the 12% next to trumps name that is absurd.
538 has about ten references to turnout in their process, here are a few:
“The national snapshot accounts for projected voter turnout in each state based on population growth since 2016, changes in how easy it is to vote since 2016, and how close the race is in that state currently — closer-polling states tend to have higher turnout. National polls are not used in the national snapshot; it’s simply a summation of the snapshots in the 50 states and Washington, D.C.“There could also be some challenges related to polling during COVID-19. In primary elections conducted during the pandemic, for instance, turnout was hard to predict. In some ways, the pandemic makes voting easier (expanded options to vote by mail in many states), but it also makes it harder in other ways (it’s difficult to socially distance if you must vote in person).Fortunately, passive aggressive is perfectly acceptable on this forum. Respecting differing opinions, of course, is not.0 -
I think people here have graciously received and responded to your criticism like any other's. The difference is to keep belaboring and reiterating the same point makes people throw up their hands.
Frankly, to me, it seems your greatest concern is based on some PTSD of 2016. Which I totally get, but to say if Biden isn't President on January 21st then 538 is terrible, unfair and inaccurate is silly. It's a forecast. A 5, 10, 15, 20% chance for Trump doesn't mean 0% chance of winning.
I think that concern and anxiety would be better used making phone calls, texts, and knocking on doors.Post edited by Jearlpam0925 on0 -
Lerxst1992 said:Jearlpam0925 said:mrussel1 said:Lerxst1992 said:mrussel1 said:Lerxst1992 said:mrussel1 said:Lerxst1992 said:I’ll say although I hate to admit it trafalgar was right last time, and hear back about a broken clock, perhaps a tune instead.
https://www.youtube.com/watch?v=BxlYSV2dMY8It’s about the harm handicappers could be doing to our elections.This is a topic about 538 and I recall the debate started with a few trump supporters meandering on here stating how these forecasts are absurd given what actually happened in 2016. Then the talk turned to WI, and how it was “solid.” So I joined in, because I think the long departed trump supporters were making an accurate point. I put up side by side images of 538s site 2016 vs now for WI, which looked nearly identical (factoring in 2 former GOP govs running a legit 3rd party campaign). I wouldn’t call that solid at that time.As a Biden supporter, I’ve been very critical of 538s forecasts and believe I even got Nates attention on social media a week ago about his 87% Biden in PA. That just seems outright absurd given the polling then. And even the polling itself shouldn’t be fully trusted. But, looking at the huge increase in covid cases the last 2 days, it would be catastrophic for the forecasts to be wrong. If they create a few thousand “lazy blue non voters” in key states, does Silver and the others take any accountability if their processes are wrong again? I am hoping for a Biden landslide, but seeing many key states within MoE given how absurdly bad trump is doing...I said 538 did a poor job. They deserve the blame for the job that they did. And it does not matter how often their readers vote. It is literally their job is to figure out who is showing up to vote, and which polls captured that the best. And it’s concerning they are taking the current election where many of the swing states are MoE and forecasting it at 87%. It’s just an opinion. You are welcome to have a different one.Did either of you bother to review their process?
Of course 538 tries to figure out turnout. One party wins when turnout is high, he other wins when it isn’t. Their forecast would have zero chance of being close to accurate without being able to predict the approximate makeup and size of the electorate
Of course they will be wrong a percentage of the time. It’s the 12% next to trumps name that is absurd.
538 has about ten references to turnout in their process, here are a few:
“The national snapshot accounts for projected voter turnout in each state based on population growth since 2016, changes in how easy it is to vote since 2016, and how close the race is in that state currently — closer-polling states tend to have higher turnout. National polls are not used in the national snapshot; it’s simply a summation of the snapshots in the 50 states and Washington, D.C.“There could also be some challenges related to polling during COVID-19. In primary elections conducted during the pandemic, for instance, turnout was hard to predict. In some ways, the pandemic makes voting easier (expanded options to vote by mail in many states), but it also makes it harder in other ways (it’s difficult to socially distance if you must vote in person).
Next, I'm not sure what your point is about turnout. Given that the EC works off of winner-takes-all in the vast majority of states, it seems wholly inconsequential whether a state wins by ten points or two - it only matters which combinations of states are won and lost. As you wrote, closer-polling states tend to have higher turnout - but that doesn't favour one party over the other, it just heightens the sense of urgency and draws both sides out to the voting booth, meaning it remains competitive. Fear about the other side winning is and will continue to be exploited by both Biden and Trump, which has up and down-side potential for both depending on whose messaging resonates louder.
Finally, in terms of 'grading' 538, it's absurd to grade statisticians based on who wins the election, as that isn't something anyone can estimate because it's a probability (it's like if I told you there's a 1 in 2 million chance of you dying in your shower last night, and you saying "that can't be true, because I didn't die last night"). A far more reasonable way to grade them would be to look at the final vote results, to evaluate whether they were close to the central tendency of the polls' expected corrected results, and if not to assure they were within the proposed margins of error.
'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 10 -
benjs said:Lerxst1992 said:Jearlpam0925 said:mrussel1 said:Lerxst1992 said:mrussel1 said:Lerxst1992 said:mrussel1 said:Lerxst1992 said:I’ll say although I hate to admit it trafalgar was right last time, and hear back about a broken clock, perhaps a tune instead.
https://www.youtube.com/watch?v=BxlYSV2dMY8It’s about the harm handicappers could be doing to our elections.This is a topic about 538 and I recall the debate started with a few trump supporters meandering on here stating how these forecasts are absurd given what actually happened in 2016. Then the talk turned to WI, and how it was “solid.” So I joined in, because I think the long departed trump supporters were making an accurate point. I put up side by side images of 538s site 2016 vs now for WI, which looked nearly identical (factoring in 2 former GOP govs running a legit 3rd party campaign). I wouldn’t call that solid at that time.As a Biden supporter, I’ve been very critical of 538s forecasts and believe I even got Nates attention on social media a week ago about his 87% Biden in PA. That just seems outright absurd given the polling then. And even the polling itself shouldn’t be fully trusted. But, looking at the huge increase in covid cases the last 2 days, it would be catastrophic for the forecasts to be wrong. If they create a few thousand “lazy blue non voters” in key states, does Silver and the others take any accountability if their processes are wrong again? I am hoping for a Biden landslide, but seeing many key states within MoE given how absurdly bad trump is doing...I said 538 did a poor job. They deserve the blame for the job that they did. And it does not matter how often their readers vote. It is literally their job is to figure out who is showing up to vote, and which polls captured that the best. And it’s concerning they are taking the current election where many of the swing states are MoE and forecasting it at 87%. It’s just an opinion. You are welcome to have a different one.Did either of you bother to review their process?
Of course 538 tries to figure out turnout. One party wins when turnout is high, he other wins when it isn’t. Their forecast would have zero chance of being close to accurate without being able to predict the approximate makeup and size of the electorate
Of course they will be wrong a percentage of the time. It’s the 12% next to trumps name that is absurd.
538 has about ten references to turnout in their process, here are a few:
“The national snapshot accounts for projected voter turnout in each state based on population growth since 2016, changes in how easy it is to vote since 2016, and how close the race is in that state currently — closer-polling states tend to have higher turnout. National polls are not used in the national snapshot; it’s simply a summation of the snapshots in the 50 states and Washington, D.C.“There could also be some challenges related to polling during COVID-19. In primary elections conducted during the pandemic, for instance, turnout was hard to predict. In some ways, the pandemic makes voting easier (expanded options to vote by mail in many states), but it also makes it harder in other ways (it’s difficult to socially distance if you must vote in person).
Next, I'm not sure what your point is about turnout. Given that the EC works off of winner-takes-all in the vast majority of states, it seems wholly inconsequential whether a state wins by ten points or two - it only matters which combinations of states are won and lost. As you wrote, closer-polling states tend to have higher turnout - but that doesn't favour one party over the other, it just heightens the sense of urgency and draws both sides out to the voting booth, meaning it remains competitive. Fear about the other side winning is and will continue to be exploited by both Biden and Trump, which has up and down-side potential for both depending on whose messaging resonates louder.
Finally, in terms of 'grading' 538, it's absurd to grade statisticians based on who wins the election, as that isn't something anyone can estimate because it's a probability (it's like if I told you there's a 1 in 2 million chance of you dying in your shower last night, and you saying "that can't be true, because I didn't die last night"). A far more reasonable way to grade them would be to look at the final vote results, to evaluate whether they were close to the central tendency of the polls' expected corrected results, and if not to assure they were within the proposed margins of error.They are more than statisticians. They are making judgements on the quality of specific polls and adjusting the results, something they were dead wrong about in 2016.Specifically they got undecideds and demographics completely wrong in their 2016 tipping point state analysis. They also ignored the most accurate polls. They show their specific 2016 math on that state by state. It’s not supposed to be offensive that commenters here would fail to look at the specifics of the topic and instead pretend someone does not understand the basics of what they do, someone who has posted 538s support on this forum multiple times? Anything to preserve that bubble and safe space here. Just like trump supporters do on conservative forums.I give 538 a lot of credit for being transparent regarding their math.
https://projects.fivethirtyeight.com/2016-election-forecast/(Click on PA)Post edited by Lerxst1992 on0 -
Using 538s tipping point analysis, and their supporters main point that 2020 is different because of the 50% threshold Clinton never reached in most swing states, while ignoring he impact of quality 3rd party candidates . Looking at 538s base polling data today, and using 50.9% as a comfort level, Biden is at 216 electoral votes.216. To many, that spells on 87% likelihood, which is fine, but it’s far from that for my liking.0
-
Lerxst1992 said:Using 538s tipping point analysis, and their supporters main point that 2020 is different because of the 50% threshold Clinton never reached in most swing states, while ignoring he impact of quality 3rd party candidates . Looking at 538s base polling data today, and using 50.9% as a comfort level, Biden is at 216 electoral votes.216. To many, that spells on 87% likelihood, which is fine, but it’s far from that for my liking.Remember the Thomas Nine !! (10/02/2018)
The Golden Age is 2 months away. And guess what….. you’re gonna love it! (teskeinc 11.19.24)
1998: Noblesville; 2003: Noblesville; 2009: EV Nashville, Chicago, Chicago
2010: St Louis, Columbus, Noblesville; 2011: EV Chicago, East Troy, East Troy
2013: London ON, Wrigley; 2014: Cincy, St Louis, Moline (NO CODE)
2016: Lexington, Wrigley #1; 2018: Wrigley, Wrigley, Boston, Boston
2020: Oakland, Oakland: 2021: EV Ohana, Ohana, Ohana, Ohana
2022: Oakland, Oakland, Nashville, Louisville; 2023: Chicago, Chicago, Noblesville
2024: Noblesville, Wrigley, Wrigley, Ohana, Ohana; 2025: Pitt1, Pitt20 -
Confidence levels are high.www.myspace.com0
-
The Juggler said:Confidence levels are high.0
-
.Gern Blansten said:Lerxst1992 said:Using 538s tipping point analysis, and their supporters main point that 2020 is different because of the 50% threshold Clinton never reached in most swing states, while ignoring he impact of quality 3rd party candidates . Looking at 538s base polling data today, and using 50.9% as a comfort level, Biden is at 216 electoral votes.216. To many, that spells on 87% likelihood, which is fine, but it’s far from that for my liking.It’s perfectly accurate. I said 538s BASE polling data, before 538 adjusts for undecideds and demographic. Here is NV 2020, base polling is 50.1. it’s steps 2 thru 4 where 538 made significant errors in 2016. On the tipping point map below, every state above NH has Biden under 50.9% base polling0
-
I don't know how anyone looks at that graphic and feels good. New Hampshire, Michigan, Minnesota, Nevada, Wisconsin, Pennsylvania, Arizona, Florida, North Carolina. All light blue. I don's see Trump losing any outside of maybe Minnesota and Nevada. And the light pink ones? Color them dark red right now.
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
2024 Napa, Wrigley, Wrigley0 -
I sure feel a lot better that they are any shade of blue at this point....Remember the Thomas Nine !! (10/02/2018)
The Golden Age is 2 months away. And guess what….. you’re gonna love it! (teskeinc 11.19.24)
1998: Noblesville; 2003: Noblesville; 2009: EV Nashville, Chicago, Chicago
2010: St Louis, Columbus, Noblesville; 2011: EV Chicago, East Troy, East Troy
2013: London ON, Wrigley; 2014: Cincy, St Louis, Moline (NO CODE)
2016: Lexington, Wrigley #1; 2018: Wrigley, Wrigley, Boston, Boston
2020: Oakland, Oakland: 2021: EV Ohana, Ohana, Ohana, Ohana
2022: Oakland, Oakland, Nashville, Louisville; 2023: Chicago, Chicago, Noblesville
2024: Noblesville, Wrigley, Wrigley, Ohana, Ohana; 2025: Pitt1, Pitt20 -
OnWis97 said:I don't know how anyone looks at that graphic and feels good. New Hampshire, Michigan, Minnesota, Nevada, Wisconsin, Pennsylvania, Arizona, Florida, North Carolina. All light blue. I don's see Trump losing any outside of maybe Minnesota and Nevada. And the light pink ones? Color them dark red right now.
Are you not going to allow yourself any level of confidence until 12:00pm on 1/20/21?Post edited by The Juggler onwww.myspace.com0 -
The Juggler said:OnWis97 said:I don't know how anyone looks at that graphic and feels good. New Hampshire, Michigan, Minnesota, Nevada, Wisconsin, Pennsylvania, Arizona, Florida, North Carolina. All light blue. I don's see Trump losing any outside of maybe Minnesota and Nevada. And the light pink ones? Color them dark red right now.
Are you not going to allow yourself any level of confidence until 12:00pm on 1/20/21?'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 10 -
Put it this way, folks. Here's the deal. Biden has a better chance at winning Texas than Trump has of winning Pennsylvania. There are a million different scenarios for Biden to become president without winning Texas and there are hardly any for Trump without PA. That's number one.
Number two. Get those confidence levels up.www.myspace.com0 -
The PTSD from 2016 is hard to shake. There are some very confident pollsters on twitter that would tell us to chill the fuck out.
If he wins again I will just shut the fuck up and accept the fact that I am outnumbered. Not sure what else to do.Remember the Thomas Nine !! (10/02/2018)
The Golden Age is 2 months away. And guess what….. you’re gonna love it! (teskeinc 11.19.24)
1998: Noblesville; 2003: Noblesville; 2009: EV Nashville, Chicago, Chicago
2010: St Louis, Columbus, Noblesville; 2011: EV Chicago, East Troy, East Troy
2013: London ON, Wrigley; 2014: Cincy, St Louis, Moline (NO CODE)
2016: Lexington, Wrigley #1; 2018: Wrigley, Wrigley, Boston, Boston
2020: Oakland, Oakland: 2021: EV Ohana, Ohana, Ohana, Ohana
2022: Oakland, Oakland, Nashville, Louisville; 2023: Chicago, Chicago, Noblesville
2024: Noblesville, Wrigley, Wrigley, Ohana, Ohana; 2025: Pitt1, Pitt20 -
benjs said:The Juggler said:OnWis97 said:I don't know how anyone looks at that graphic and feels good. New Hampshire, Michigan, Minnesota, Nevada, Wisconsin, Pennsylvania, Arizona, Florida, North Carolina. All light blue. I don's see Trump losing any outside of maybe Minnesota and Nevada. And the light pink ones? Color them dark red right now.
Are you not going to allow yourself any level of confidence until 12:00pm on 1/20/21?___________________________________________
"...I changed by not changing at all..."0 -
Question, though, and now this is me totally doing the touchdown dance early: if you were to go to inauguration the theme should be U-Hauls right? Like just thousands of U-Hauls showing up to escort dipshit?
Just droves of U-Hauls laying on the horn of Pennsylvania Avenue. I mean I wouldn't go to see Joe sworn in, but more to enjoy this motherfucker going bye-bye.
And there ya have it, folks - I have officially jinxed us all.0 -
The Juggler said:Put it this way, folks. Here's the deal. Biden has a better chance at winning Texas than Trump has of winning Pennsylvania. There are a million different scenarios for Biden to become president without winning Texas and there are hardly any for Trump without PA. That's number one.
Number two. Get those confidence levels up.
I'm with you, I'm not nearly as worried this time around. Maybe I should be but I'm not.0
Categories
- All Categories
- 148.8K Pearl Jam's Music and Activism
- 110K The Porch
- 274 Vitalogy
- 35K Given To Fly (live)
- 3.5K Words and Music...Communication
- 39.1K Flea Market
- 39.1K Lost Dogs
- 58.7K Not Pearl Jam's Music
- 10.6K Musicians and Gearheads
- 29.1K Other Music
- 17.8K Poetry, Prose, Music & Art
- 1.1K The Art Wall
- 56.7K Non-Pearl Jam Discussion
- 22.2K A Moving Train
- 31.7K All Encompassing Trip
- 2.9K Technical Stuff and Help