angrysoba
Philosophile
Yeah, I suppose there are those who think their vote is pointless and therefore don't cast it.
-
Security incident: ISF was recently accessed by intruders. Please change your password, and change it anywhere else you used it. Read more
2020 Presidential Election
- Thread starter Skeptic Ginger
- Start date
- Status
Yeah, I suppose there are those who think their vote is pointless and therefore don't cast it.
Hlafordlaes
Disorder of Kilopi
Catalonia Now Interfering in US Elections!!
You read it right, someone from Catalonia will be voting in a US presidential election for the very first time since 1976. Damn foreigners and other suspicious types think 2020 concerns -- get this -- the entire planet.
You read it right, someone from Catalonia will be voting in a US presidential election for the very first time since 1976. Damn foreigners and other suspicious types think 2020 concerns -- get this -- the entire planet.
The Great Zaganza
Maledictorian
- Joined
- Aug 14, 2016
- Messages
- 33,038
You read it right, someone from Catalonia will be voting in a US presidential election for the very first time since 1976. Damn foreigners and other suspicious types think 2020 concerns -- get this -- the entire planet.
wait -since when is the US part of the planet?
538 has also started crunching number on specific "How it will play out" scenarios, which are always interesting to think about.
Trump wins in a landslide (Wins the popular vote by double digits) -- Less than 1%
Biden wins in a landslide -- 30%
Biden wins the popular vote but loses the EC -- 10%
Trump wins the popular vote but loses the EC -- Less than 1%
No one wins, it goes to the House (Neither candidate gets 270 votes, House picks the President) -- Less than 1%
The Presidency will be determined on a recount -- 5%
We get the exact same EC Map in 2020 that we did in 2016 -- Less than 1%
A 269 to 269 EC Vote Tie -- 1%
Biden wins at least one state Trump won in 2016 -- 92%
Trump wins at least on state that Hillary won in 2016 -- 37%
Trump wins in a landslide (Wins the popular vote by double digits) -- Less than 1%
Biden wins in a landslide -- 30%
Biden wins the popular vote but loses the EC -- 10%
Trump wins the popular vote but loses the EC -- Less than 1%
No one wins, it goes to the House (Neither candidate gets 270 votes, House picks the President) -- Less than 1%
The Presidency will be determined on a recount -- 5%
We get the exact same EC Map in 2020 that we did in 2016 -- Less than 1%
A 269 to 269 EC Vote Tie -- 1%
Biden wins at least one state Trump won in 2016 -- 92%
Trump wins at least on state that Hillary won in 2016 -- 37%
Segnosaur
Penultimate Amazing
That might only work if you somehow make the lack of a pet sound "unamerican".
"Saint Ronald Reagan had a pet. Why doesn't Donald?"
Sent from my LM-X320 using Tapatalk
There are a lot of IFs behind those numbers. It's not quite the Bible Code, but I have a very hard time believing that they aren't wildly oversell their ability to make accurate predictions by even doing these calculations.
xjx388
Penultimate Amazing
- Joined
- Dec 1, 2010
- Messages
- 11,392
In other words, "We have no idea what is going to happen."
I don't think they ever claim to know what is going to happen.
They claim to be able to take polling data, apply their model to normalise it and then provide an approximate probability (with error bars) for various outcomes.
Those probabilities will then change as few information becomes available (for example one candidate moving significantly in the polls), or if they make changes to their model.
IMO it's a reasonable approach.
At the moment, it appears that the most likely outcome is that Biden will win both the Electoral College and the popular vote. If President Trumps poll numbers were to improve significantly, or if he was successful in making postal voting illegal, then I'd expect that to change.
I don't get the bolded parts.
If it's a 269-269 tie, the House picks the President.
So how can that be at 1%, and the House picking the President at less than 1%?
Shouldn't the House picking the President have an equal or higher probability than a tie?
Last edited:
In general people tend to say that kind of thing if what the pundit says is uncomfortable
.Somebody who would very much like President Trump to be reelected either has to accept that he is currently trailing very badly in the polls and currently looks relatively unlikely to be reelected or they can try to find ways to discredit that information.
One way is to say that the polls are simply wrong. 2016 gives some ammunition for that though in most cases the later polls were "right" within the margin of error (margin of error being something which is ignored by pretty much everyone).
Another is to attempt to discredit the pundit. Claiming that the pundit is making vague predictions is one way to do that. A lot of people who don't understand polling and probability want pundits to make early, specific, accurate, predictions. FiveThirtyEight don't do any of those things. They change their forecasts regularly in light of new information, they are clear about the uncertainty in those forecasts and because the forecasts are updated regularly, most of the earlier ones are "wrong".
Dunno.
This is the current list of forecasts:
https://projects.fivethirtyeight.com/2020-election-forecast/
The 269-269 EC tie isn't there as a different line to the house deciding any more.
Ok, thanks. Probably a typo somewhere.
Oh the 269-269 thing came from their actual election predictions in which Biden wins 77 out every 100 predicted scenarios and Trump wins in 22 out of every 100 predicted scenarios. The 269-269 just accounts for the odd percentage out and they might be using different metrics for that then their "Scenario Probabilities."
Should have made that clearer.
So basically it's 77% Chance Biden Wins, 22% Chance Trump Wins, 1% Chance "Other" and "Other" includes a 269-269 split as one of its possibilities.
Should have made that clearer.
So basically it's 77% Chance Biden Wins, 22% Chance Trump Wins, 1% Chance "Other" and "Other" includes a 269-269 split as one of its possibilities.
Last edited:
No problem.
Was looking at the 538 model for 2016.
At this point in the race, Clinton was around 60%, then rose to 87% after the debates, down to 64% after Comey's infamous letter, and up to 71% on election day (all % are probabilities to win).
pgwenthold
Penultimate Amazing
- Joined
- Sep 19, 2001
- Messages
- 21,829
I am wondering, have they modified their algorithm since then?
If they went back and recalculated 2016 using the current algorithm, would they give the same result?
thaiboxerken
Penultimate Amazing
- Joined
- Sep 17, 2001
- Messages
- 35,109
But how do they figure GOP cheating and Russian involvement in their algorithms?
Well you can't. That's a completely different system to simulate.
That's why the question is less "Will Biden win the election" and more "Will Trump care?"
xjx388
Penultimate Amazing
- Joined
- Dec 1, 2010
- Messages
- 11,392
In general people tend to say that kind of thing if what the pundit says is uncomfortable
Somebody who would very much like President Trump to be reelected either has to accept that he is currently trailing very badly in the polls and currently looks relatively unlikely to be reelected or they can try to find ways to discredit that information.
One way is to say that the polls are simply wrong. 2016 gives some ammunition for that though in most cases the later polls were "right" within the margin of error (margin of error being something which is ignored by pretty much everyone).
Another is to attempt to discredit the pundit. Claiming that the pundit is making vague predictions is one way to do that. A lot of people who don't understand polling and probability want pundits to make early, specific, accurate, predictions. FiveThirtyEight don't do any of those things. They change their forecasts regularly in light of new information, they are clear about the uncertainty in those forecasts and because the forecasts are updated regularly, most of the earlier ones are "wrong".
I have no problem with 538's approach. I'm not discrediting them. I'm simply pointing out that the information they are providing is nothing more than an educated guess. It's a summation of the polls and their numbers are only as good as those polls are. It's like trying to tell us what the weather will probably be like 43 days from now.
There's a danger in putting too much stock in these kinds of forecasts: voter apathy. We saw it to some extent in 2016. Clinton had a 71% chance, according to 538 and we all know what happened. In an election such as this one, where it's so important to get Biden and anti-Trump voters motivated to actually vote for Biden, I fear that information like this might lead to voter apathy.
Yeah I also remember people on this very board spending all of October where almost literally every post from them was jumping up and down having absolute hissy fits over the fact that 538 was so "Bullish" on Trump, giving him mid-high twenty percent chances when everyone else was giving him some microscopic chance.
- Status