|
Replies: 7
| visibility 566
|
All-In [31449]
TigerPulse: 100%
Posts: 10176
Joined: 1/28/15
|
Nate Silver's FiveThirtyEight model
Nov 1, 2020, 9:49 AM
|
|
has shifted even further towards a likely Biden win, now giving him 90% probability. This is NUTS. For comparison, they had Trump at 30% in 2016. History has borne out that their model is well calibrated. Silver and his editorial staff have put out several podcasts in the last week almost apologizing for their model showing what it does. Keep in mind this says nothing about the margin of victory or even which states will be won, but the overall probability of a win. They've had a few points that I found interesting.
1. Trump still has a 10% probability in their model of an EC victory, but for him to win the polls will have to be even further off and outside the MOE than they were in 2016, and they explain the changes in the model and the high quality polls they include that would seem to suggest that the results won't be quite as off as 2016.
2. Trump's path to victory is quite small, and if he were to lose FL, OH, GA, NC, AZ, OR PA it's basically over, provided Biden's lead in MN, WI, and MI proves to be accurate. Trump would have to sweep all of those, or pull a surprise flip elsewhere (NV? MI?), to get to 270.
3. The "shy Trump voter" theory for polls being what they are has no evidence whatsoever
4. Pennsylvania delivers the EC victory in 37% of their simulations. I hope for all of our sakes that it doesn't come down to that just that one state. With all of the chicanery regarding mailed-in ballots and when/whether they will be counted, it would be 2000 all over again, but worse.
IMO a best case scenario is a Biden win and the R's somehow keep 51 in the senate... problem being if turnout is there to give Biden the WH... the senate likely goes Blue too.
|
|
|
|
 |
Rock Defender [53]
TigerPulse: 90%
Posts: 35
Joined: 11/30/98
|
Re: Nate Silver's FiveThirtyEight model
Nov 1, 2020, 10:27 AM
|
|
|
|
|
|
|
|
All-In [34105]
TigerPulse: 100%
Posts: 33608
Joined: 9/13/99
|
This page is very interesting
Nov 1, 2020, 10:34 AM
|
|
You can see how the odds change if certain states swung one way or the other. It shows that a Biden win in Florida, which might be known sooner than most battleground states, would all but end the contest.
https://projects.fivethirtyeight.com/trump-biden-election-map/ Explore The Ways Trump Or Biden Could Win The Election Choose whether President Trump or Joe Biden will win each state in the 2020 presidential election and we’ll recalculate whether they have a path to 270 electoral votes and what their chance is of winning the Electoral College.
|
|
|
|
|
|
Rock Defender [53]
TigerPulse: 90%
Posts: 35
Joined: 11/30/98
|
Re: This page is very interesting
Nov 1, 2020, 10:41 AM
|
|
|
|
|
|
|
|
Orange Blooded [4365]
TigerPulse: 80%
Posts: 8370
Joined: 1/4/17
|
Re: This page is very interesting
Nov 1, 2020, 12:03 PM
[ in reply to This page is very interesting ] |
|
spooneye,
Why would you think Florida might be known early - given their history ??
|
|
|
|
|
|
All-In [34105]
TigerPulse: 100%
Posts: 33608
Joined: 9/13/99
|
Re: This page is very interesting
Nov 1, 2020, 6:56 PM
|
|
Florida has been doing voting-by-mail for years (probably because of snowbirds). They have it set up so that they are already processing ballots. Their results will probably be known on election night.
https://projects.fivethirtyeight.com/election-results-timing/
|
|
|
|
|
|
All-TigerNet [12943]
TigerPulse: 100%
Posts: 6228
Joined: 1/26/16
|
Trump wins FL by 3+ points***
Nov 1, 2020, 3:04 PM
|
|
|
|
|
|
|
|
Hall of Famer [22940]
TigerPulse: 100%
Posts: 16894
Joined: 12/2/00
|
90% prediction is stupidly outrageous
Nov 1, 2020, 6:33 PM
|
|
I’d still give Trump a 55/45 chance.
Trump will win FL easily. May sneak out a victory in PA. Believe Biden wins AZ.
|
|
|
|
|
|
Replies: 7
| visibility 566
|
|
|
50
