|
On November 05 2020 00:00 Wegandi wrote:Show nested quote +On November 04 2020 23:54 Sbrubbles wrote:
Question for those with better understanding of the US electoral system: I understand the argument for smaller population states to have proportionally more "votes" (even if I don't agree). That said, is there an argument for the electoral votes being winner-take-all and not some proportional split? I don't remember this being in any of the Federalist papers or heavily debated at the ratifying conventions (I'd have to double check). There's nothing saying you can't except the Democrats would never go for it as they'd lose more than they gain. They disproportionately win the bigger electoral states. Last thing they want is to split up Cali, NY, IL, etc.
Wouldn't that heavily depend on which system more closely mirrors the popular vote? Democrats have won the popular vote every time since 2008, so if proportional representation were implemented in a way to better mirror the popular vote, I think Democrats would jump at the opportunity. Of course, it'd have to be done across the board. It makes no sense for a state like Texas change to proportional representation while Cali keeps Winner Takes All.
|
On November 05 2020 00:04 Wegandi wrote:Show nested quote +On November 04 2020 23:59 Stratos_speAr wrote:On November 04 2020 23:57 Wegandi wrote:On November 04 2020 23:53 Stratos_speAr wrote:On November 04 2020 23:51 Wegandi wrote:On November 04 2020 23:45 Stratos_speAr wrote:
Another interesting side story here is that many people (like Wegandi) tried to dunk hard on pollsters and modelers like 538, whereas every model had the current projected outcome within the realm of their "likely" outcomes.
There will be a lot of talk about how polling of individual states was off (like FL), but the actual models did a pretty damn good job. What? The polls had Biden beating Trump by 1% in Florida (that's a huge error), had Biden winning WI by 7%, MI by 4-5%, etc. It is across the board the polls are off by 4%+ in one direction. The polls sucked and weren't close. Look how bad this is lol https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html I'll give you a chance to read my post again. More slowly this time, please. If you think the pollsters did a good job at accurately forecasting this election you're not going to convince me. The fact that 538 had Biden at 80%+ and he barely won is not a good look (coupled with the polls). Models. We're talking models. There was also plenty of education last night on what statistical probabilities mean and I know you were here. If 538 gave Biden an 80% chance to win and he wins, do you know what that means? They were right.Not only this. But do you know what it means if Trump wins? They were still right, because a 10% chance is a significant chance to win. How could they ever be wrong then? The margins do matter. The polling matters. Biden winning WI by 12% or .3% is very different for instance.
Yes, the polling matters.
But we're talking about the models, which take an aggregate of polling data, possible error, and other non-polling political factors to estimate the chance that something happens.
89% chance of winning doesn't mean that they predicted an 89% share of votes.
Models from 538 and several other outlets had this exact scenario in mind and it was within the distribution of "most likely scenarios" for pretty much all of them.
You tried to dunk on 538 and other models last night, and you're looking quite foolish now because their models are looking pretty good. That's the issue here.
|
Good luck with this. I'd say have fun but that doesn't sound likely.
|
On November 05 2020 00:05 Diavlo wrote:Show nested quote +On November 04 2020 23:45 Stratos_speAr wrote:
Another interesting side story here is that many people (like Wegandi) tried to dunk hard on pollsters and modelers like 538, whereas every model had the current projected outcome within the realm of their "likely" outcomes.
There will be a lot of talk about how polling of individual states was off (like FL), but the actual models did a pretty damn good job. I think its difficult for most people (myself included) to grasp mathematical models. Especially when used in everyday life context. The concerning thing to me is that people with little knowledge or understanding on complex topics have such strong opinions on them. And how unwilling they are to just accept that they don't know. It was not clear to me just how much this problem is widespread before the COVID crisis. I'm an MD who worked in a virology lab for over a year and specialize in molecular biology. And yet I have no pertinent opinion on epidemiology, public health and pandemic strategies... So how the fuck are so many people with absolutely no medical background talking with such authority on these subjects?
It's arrogance.
That's about it.
|
Yeah, how did they not see the exact same thing as 4 years ago happening? I'm sure republicans pulled out lots of tricks in the swing states/specific areas, but that's too be expected too...
|
On November 05 2020 00:11 Stratos_speAr wrote:Show nested quote +On November 05 2020 00:05 Diavlo wrote:On November 04 2020 23:45 Stratos_speAr wrote:
Another interesting side story here is that many people (like Wegandi) tried to dunk hard on pollsters and modelers like 538, whereas every model had the current projected outcome within the realm of their "likely" outcomes.
There will be a lot of talk about how polling of individual states was off (like FL), but the actual models did a pretty damn good job. I think its difficult for most people (myself included) to grasp mathematical models. Especially when used in everyday life context. The concerning thing to me is that people with little knowledge or understanding on complex topics have such strong opinions on them. And how unwilling they are to just accept that they don't know. It was not clear to me just how much this problem is widespread before the COVID crisis. I'm an MD who worked in a virology lab for over a year and specialize in molecular biology. And yet I have no pertinent opinion on epidemiology, public health and pandemic strategies... So how the fuck are so many people with absolutely no medical background talking with such authority on these subjects? It's arrogance. That's about it.
And a small mountain of Dunning Kruger
|
On November 05 2020 00:07 JimmiC wrote:Show nested quote +On November 04 2020 23:57 Wegandi wrote:On November 04 2020 23:53 Stratos_speAr wrote:On November 04 2020 23:51 Wegandi wrote:On November 04 2020 23:45 Stratos_speAr wrote:
Another interesting side story here is that many people (like Wegandi) tried to dunk hard on pollsters and modelers like 538, whereas every model had the current projected outcome within the realm of their "likely" outcomes.
There will be a lot of talk about how polling of individual states was off (like FL), but the actual models did a pretty damn good job. What? The polls had Biden beating Trump by 1% in Florida (that's a huge error), had Biden winning WI by 7%, MI by 4-5%, etc. It is across the board the polls are off by 4%+ in one direction. The polls sucked and weren't close. Look how bad this is lol https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html I'll give you a chance to read my post again. More slowly this time, please. If you think the pollsters did a good job at accurately forecasting this election you're not going to convince me. The fact that 538 had Biden at 80%+ and he barely won is not a good look (coupled with the polls). People really don't seem to understand odds and projections. Basically they nailed it. Texas and flordia were outside shots that the Dems were not favored in and they didn't get there, both toss ups that favored Reps. Zona was a toss up that favored Dems it got there. They were right on Col and Nev. They pretty much nailed the rust belt. The order that the votes come in does not matter. They were not projecting that, they were projecting final results. Anyone who at this point thinks polling is bad and shouldn't be trusted does not understand how it works or math.
The RCP average for national vote had Biden +7.2% and that's with one Rasmussen poll being a huge Trump outlier. Biden is likely to win with 2%. You had reportedly AA+ polls calling biden with 8-10 point leads days before the election. This is way outside even the ludicrous 4%+ MOE's. They weren't close.
https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html
|
On November 05 2020 00:04 suprTonttu wrote:Show nested quote +On November 04 2020 23:57 AsariCommando wrote:On November 04 2020 23:54 Sbrubbles wrote:
Question for those with better understanding of the US electoral system: I understand the argument for smaller population states to have proportionally more "votes" (even if I don't agree). That said, is there an argument for the electoral votes being winner-take-all and not some proportional split? There really isn't an argument for the winner-take-all system. Nebraska and Maine have moved away from it and I hope other states do as well. It's really quite ridiculous. That certainly is quite ridiculous. Considering California has 55 "votes". That is kinda big portion of 270 "votes" that is needed to win. Or Texas' 38 etc. Sure they are both big and populous, but it just makes that bigger cities dominate way too much. Makes that smaller towns become irrelevant in my eyes when e.g. one big city has more votes than rest of the state combined. Why shouldn't 1 big city have more votes then the rest of the state combined when it has more people living in it then the rest of the state combined?
Especially when your looking at a national stage with the EC.
I can understand wanting some rural/urban balance at the local level (max state) level but why should the singular leader of the entire USA be anyone but the person with the most votes?
|
On November 05 2020 00:12 Wegandi wrote:Show nested quote +On November 05 2020 00:07 JimmiC wrote:On November 04 2020 23:57 Wegandi wrote:On November 04 2020 23:53 Stratos_speAr wrote:On November 04 2020 23:51 Wegandi wrote:On November 04 2020 23:45 Stratos_speAr wrote:
Another interesting side story here is that many people (like Wegandi) tried to dunk hard on pollsters and modelers like 538, whereas every model had the current projected outcome within the realm of their "likely" outcomes.
There will be a lot of talk about how polling of individual states was off (like FL), but the actual models did a pretty damn good job. What? The polls had Biden beating Trump by 1% in Florida (that's a huge error), had Biden winning WI by 7%, MI by 4-5%, etc. It is across the board the polls are off by 4%+ in one direction. The polls sucked and weren't close. Look how bad this is lol https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html I'll give you a chance to read my post again. More slowly this time, please. If you think the pollsters did a good job at accurately forecasting this election you're not going to convince me. The fact that 538 had Biden at 80%+ and he barely won is not a good look (coupled with the polls). People really don't seem to understand odds and projections. Basically they nailed it. Texas and flordia were outside shots that the Dems were not favored in and they didn't get there, both toss ups that favored Reps. Zona was a toss up that favored Dems it got there. They were right on Col and Nev. They pretty much nailed the rust belt. The order that the votes come in does not matter. They were not projecting that, they were projecting final results. Anyone who at this point thinks polling is bad and shouldn't be trusted does not understand how it works or math. The RCP average for national vote had Biden +7.2% and that's with one Rasmussen poll being a huge Trump outlier. Biden is likely to win with 2%. You had reportedly AA+ polls calling biden with 8-10 point leads days before the election. This is way outside even the ludicrous 4%+ MOE's. They weren't close. https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html
We're not talking about polls.
We're not talking about polls.
Say it within me now, slowly.
We're not talking about polls.
We're not talking about polls.
We're not talking about polls.
We're not talking about polls.
We're not talking about polls.
We're not talking about polls.
User was warned for this post.
|
I should apologize for my post about pollsters and poll aggregators made yesterday. I got caught up in the moment and my ignorance was very apparent. Upon reflection, a narrow Biden victory (which hasn't happened yet!) still does fit into the 89% or 95% or whatever was forecasted in the models.
|
On November 05 2020 00:08 ghrur wrote:Show nested quote +On November 05 2020 00:00 Wegandi wrote:On November 04 2020 23:54 Sbrubbles wrote:
Question for those with better understanding of the US electoral system: I understand the argument for smaller population states to have proportionally more "votes" (even if I don't agree). That said, is there an argument for the electoral votes being winner-take-all and not some proportional split? I don't remember this being in any of the Federalist papers or heavily debated at the ratifying conventions (I'd have to double check). There's nothing saying you can't except the Democrats would never go for it as they'd lose more than they gain. They disproportionately win the bigger electoral states. Last thing they want is to split up Cali, NY, IL, etc. Wouldn't that heavily depend on which system more closely mirrors the popular vote? Democrats have won the popular vote every time since 2008, so if proportional representation were implemented in a way to better mirror the popular vote, I think Democrats would jump at the opportunity. Of course, it'd have to be done across the board. It makes no sense for a state like Texas change to proportional representation while Cali keeps Winner Takes All.
It wouldn't be a 1:1 with popular vote especially depending how they decide to do it and I suspect voter turnout is artificially depressed in 1 party non-competitive states so who knows, but yeah. GOP tends to run up the vote % higher in flyover country so would potentially lose less than they'd gain from their 40%+ vote shares of CA, IL, NY, etc.
|
From the Biden camp. All rests on the Judge in PA hearing the GOP case today.
|
I also think it's really important to note that most models explicitly said that they don't take into account voter suppression, i.e. they assume that the SC won't throw out a bunch of votes in places like MI or PA.
|
Note: my latest maps show only ~135 million votes counted so far. The rest are likely to favor Biden, due to how mail in ballots are usually counted late. If they do, then we may actually see that 7% popular vote margin.
Doesn't really matter. Biden has it even without PA.
|
On November 05 2020 00:14 Stratos_speAr wrote:Show nested quote +On November 05 2020 00:12 Wegandi wrote:On November 05 2020 00:07 JimmiC wrote:On November 04 2020 23:57 Wegandi wrote:On November 04 2020 23:53 Stratos_speAr wrote:On November 04 2020 23:51 Wegandi wrote:On November 04 2020 23:45 Stratos_speAr wrote:
Another interesting side story here is that many people (like Wegandi) tried to dunk hard on pollsters and modelers like 538, whereas every model had the current projected outcome within the realm of their "likely" outcomes.
There will be a lot of talk about how polling of individual states was off (like FL), but the actual models did a pretty damn good job. What? The polls had Biden beating Trump by 1% in Florida (that's a huge error), had Biden winning WI by 7%, MI by 4-5%, etc. It is across the board the polls are off by 4%+ in one direction. The polls sucked and weren't close. Look how bad this is lol https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html I'll give you a chance to read my post again. More slowly this time, please. If you think the pollsters did a good job at accurately forecasting this election you're not going to convince me. The fact that 538 had Biden at 80%+ and he barely won is not a good look (coupled with the polls). People really don't seem to understand odds and projections. Basically they nailed it. Texas and flordia were outside shots that the Dems were not favored in and they didn't get there, both toss ups that favored Reps. Zona was a toss up that favored Dems it got there. They were right on Col and Nev. They pretty much nailed the rust belt. The order that the votes come in does not matter. They were not projecting that, they were projecting final results. Anyone who at this point thinks polling is bad and shouldn't be trusted does not understand how it works or math. The RCP average for national vote had Biden +7.2% and that's with one Rasmussen poll being a huge Trump outlier. Biden is likely to win with 2%. You had reportedly AA+ polls calling biden with 8-10 point leads days before the election. This is way outside even the ludicrous 4%+ MOE's. They weren't close. https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html We're not talking about polls.
We're not talking about polls.Say it within me now, slowly. We're not talking about polls.
We're not talking about polls.We're not talking about polls.
We're not talking about polls.We're not talking about polls.
We're not talking about polls.
I'm not interested in tautology about models wherein you posit that if Biden wins they're right and if Trump wins they're also right. Why are you ignoring how awful the polls were?
|
How many times does it need to be repeated, those are just mathematical that make statistical predictions. They are not Nostradamus, and not meant to be.
I read a million twitter post saying 538 was "wrong again". Those people do not understand statistics; if Silver had given Trump 2% chances of winning and Trump had won, he still wouldn't have been "wrong".
They have repeated the whole cycle that Trump had a chance to win.
|
|
|
On November 05 2020 00:20 Nevuk wrote:
Doesn't really matter. Biden has it even without PA.
The margins in several states are small enough that having PA would be quite nice for him
|
On November 05 2020 00:21 Wegandi wrote:Show nested quote +On November 05 2020 00:14 Stratos_speAr wrote:On November 05 2020 00:12 Wegandi wrote:On November 05 2020 00:07 JimmiC wrote:On November 04 2020 23:57 Wegandi wrote:On November 04 2020 23:53 Stratos_speAr wrote:On November 04 2020 23:51 Wegandi wrote:On November 04 2020 23:45 Stratos_speAr wrote:
Another interesting side story here is that many people (like Wegandi) tried to dunk hard on pollsters and modelers like 538, whereas every model had the current projected outcome within the realm of their "likely" outcomes.
There will be a lot of talk about how polling of individual states was off (like FL), but the actual models did a pretty damn good job. What? The polls had Biden beating Trump by 1% in Florida (that's a huge error), had Biden winning WI by 7%, MI by 4-5%, etc. It is across the board the polls are off by 4%+ in one direction. The polls sucked and weren't close. Look how bad this is lol https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html I'll give you a chance to read my post again. More slowly this time, please. If you think the pollsters did a good job at accurately forecasting this election you're not going to convince me. The fact that 538 had Biden at 80%+ and he barely won is not a good look (coupled with the polls). People really don't seem to understand odds and projections. Basically they nailed it. Texas and flordia were outside shots that the Dems were not favored in and they didn't get there, both toss ups that favored Reps. Zona was a toss up that favored Dems it got there. They were right on Col and Nev. They pretty much nailed the rust belt. The order that the votes come in does not matter. They were not projecting that, they were projecting final results. Anyone who at this point thinks polling is bad and shouldn't be trusted does not understand how it works or math. The RCP average for national vote had Biden +7.2% and that's with one Rasmussen poll being a huge Trump outlier. Biden is likely to win with 2%. You had reportedly AA+ polls calling biden with 8-10 point leads days before the election. This is way outside even the ludicrous 4%+ MOE's. They weren't close. https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html We're not talking about polls.
We're not talking about polls.Say it within me now, slowly. We're not talking about polls.
We're not talking about polls.We're not talking about polls.
We're not talking about polls.We're not talking about polls.
We're not talking about polls. I'm not interested in tautology about models wherein you posit that if Biden wins they're right and if Trump wins they're also right. Why are you ignoring how awful the polls were?
Because
We're not talking about polls.
Why are you having such a hard time comprehending that polls and models are two different things? I already mentioned that polling errors will be analyzed by the entire world later. I'm calling you out for trying to bash statistical models when they pretty much all did a good job.
How many times does it need to be repeated, those are just mathematical that make statistical predictions. They are not Nostradamus, and not meant to be.
I read a million twitter post saying 538 was "wrong again". Those people do not understand statistics; if Silver had given Trump 2% chances of winning and Trump had won, he still wouldn't have been "wrong".
They have repeated the whole cycle that Trump had a chance to win.
They literally ran like four different pieces over the weekend about how their model gives Trump a significant chance to win.
Ignorance like Wegandi's is purposeful at this point.
|
On November 05 2020 00:20 Nevuk wrote:https://twitter.com/ElectProject/status/1323897829044195329Note: my latest maps show only ~135 million votes counted so far. The rest are likely to favor Biden, due to how mail in ballots are usually counted late. If they do, then we may actually see that 7% popular vote margin. Doesn't really matter. Biden has it even without PA.
Yeah, Arizona appears locked in at long last with 3.4 points to Biden at 99% reporting, and both MI and WI appear to have swung his way with outstanding early vote totals in pro-Biden areas. Both of the latter two might even be able to be final-ish later today (recounts are very likely given their margins, but even if margins stayed exactly the same the numbers involved seem too large to change in terms of historical recounts).
|
|
|
|
|
|
EPT
