Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

CajunBlazer

(5,648 posts)
Mon Sep 26, 2016, 06:14 PM Sep 2016

Hillary's chances of winning on 538 site are improving as new state polls are added.

Nate Silver's 538 site now : Chances of winning: Clinton 55.0% - Trump 45.0%

Improved from this morning: Chances of winning: Clinton 51.5% - Trump 48.5%

Separation increased from 3% to 10% - That's significant in one day.

17 replies = new reply since forum marked as read
Highlight: NoneDon't highlight anything 5 newestHighlight 5 most recent replies
Hillary's chances of winning on 538 site are improving as new state polls are added. (Original Post) CajunBlazer Sep 2016 OP
Nate ochem Sep 2016 #1
Just because a site is slow to add new poll results does not mean there're "accurate" CajunBlazer Sep 2016 #6
He also predicted ochem Sep 2016 #8
You are citing Silver's predictions based only on early poll data CajunBlazer Sep 2016 #10
you prove my point ochem Sep 2016 #17
Oh goodie Doctor Jack Sep 2016 #2
Maybe you are chosing to read the wrong OP's CajunBlazer Sep 2016 #7
No, mostly annoyed at 538 for their seemingly random bullshit Doctor Jack Sep 2016 #11
Well, your sarcasm is still misplaced CajunBlazer Sep 2016 #14
Yes it is, but the more I am seeing the different polls the more I am wondering still_one Sep 2016 #3
I think what you are seeing is more volitility in the polled populations... CajunBlazer Sep 2016 #9
I don't know, but this was what happend after President Obama's first debate with Romney with still_one Sep 2016 #15
Well, remember that Obama wasn't at his best in that first debate CajunBlazer Sep 2016 #16
Please stop BlueInPhilly Sep 2016 #4
Your not right or wrong, but.... CajunBlazer Sep 2016 #5
The sheer number of polls out there... Wounded Bear Sep 2016 #12
I wouldn't be believing conspiracy theories if I were you CajunBlazer Sep 2016 #13

CajunBlazer

(5,648 posts)
6. Just because a site is slow to add new poll results does not mean there're "accurate"
Mon Sep 26, 2016, 06:59 PM
Sep 2016

Quite the opposite, it means that statistical predictions like that of the Princeton group are often out of date.

The reason why Silvers projections change daily, or even more often, is because new polls are entered into the statistical software that he developed as they are reported. The software is rerun as soon as the data is entered and displayed automatically on the website. So that his site always has latest view which includes the latest polls.

If other sites change less often it is because they enter new poll data less frequently and rerun their software less often so they don't produce their latest view of the data as often.

So, all other things being equal, would you rather predictions that are up to date, or old out of date predictions?

Hard to argue with Silver's results. He picked the winner in 50 out of 50 states in 2012 and 49 out of 50 in 2008 using the same software and methodology.

CajunBlazer

(5,648 posts)
10. You are citing Silver's predictions based only on early poll data
Mon Sep 26, 2016, 07:20 PM
Sep 2016

Go back and look at his predictions as the Republican nomination process progressed. That why his predictions change with each new set of polls. Poll can only provide info as to who people will vote for on the day they are polled. As people change their mind (for instance when their guy drops out of the race) the polls change and therefore the predictions change. But again they can be accurate for only for one point in time

Doctor Jack

(3,072 posts)
11. No, mostly annoyed at 538 for their seemingly random bullshit
Mon Sep 26, 2016, 07:21 PM
Sep 2016

Nothing against the OP. All of the sarcasm is directed at 538

CajunBlazer

(5,648 posts)
14. Well, your sarcasm is still misplaced
Mon Sep 26, 2016, 07:40 PM
Sep 2016

All you are seeing on 538 is new data flowing into a computer program so the computer spits out new results. Sarcasm is wasted when applied to uncaring data. The electorate is in flux more than usual, but then it is a crazy election cycle.

However, historically the candidate leading in the polls going into the first debate has lost only twice - in 1980 and 2000 and the candidate leading in the polls after the first debate has always won. I f those trends are to continue, the polling data should stabilize soon.

still_one

(91,965 posts)
3. Yes it is, but the more I am seeing the different polls the more I am wondering
Mon Sep 26, 2016, 06:22 PM
Sep 2016

if some of the new methodologies being used by some of the pollsters are distorting the results, and this is represented by the crazy volatilities that seem to be occurring


CajunBlazer

(5,648 posts)
9. I think what you are seeing is more volitility in the polled populations...
Mon Sep 26, 2016, 07:15 PM
Sep 2016

...rather than volatility in the polls themselves.

What I mean by that is that is voters that are changing their minds on a constant basis on who they will support. For instance, if the decisions of the people of one state remained the same over a period of time, poll results of those people might vary from poll to poll, but the combined average of all of the state polls (which is more accurate stylistically) would stay more or less the same over time.

What we have been seeing instead is Hillary's lead in the state poll averages have been dropping more or less constantly since the convention. We are reaching a critical point in this election cycle. Never has the leader in the polls after the first debate lost the election.

Hopefully this turnaround is the first sign of a complete turnaround in the poll trends to date.

CajunBlazer

(5,648 posts)
16. Well, remember that Obama wasn't at his best in that first debate
Mon Sep 26, 2016, 08:10 PM
Sep 2016

The consensus was he lost that first debate and then hewaxed Romney's clock in the remaining debates.

BlueInPhilly

(870 posts)
4. Please stop
Mon Sep 26, 2016, 06:22 PM
Sep 2016

Constantly posting 538 odds. The ever changing odds only proves how volatile Silver's model is. Not really helping.

I do expect some retort from you telling me how wrong I am. So retort away. But for now, PLEASE stop giving so much credence to Silver's model.

CajunBlazer

(5,648 posts)
5. Your not right or wrong, but....
Mon Sep 26, 2016, 06:48 PM
Sep 2016

.... why are you reading a post you don't want to read? Just stop reading my posts on 538 which are always indicated as such in the title.

By the way the reason why Silvers projections change daily, or even more often, is new polls are entered into the statistical software that he developed so that his site always has latest view based on the latest polls. He picked the winner in 50 out of 50 states in 2012 and 49 out of 50 in 2008 using the same software and methodology.

Wounded Bear

(58,440 posts)
12. The sheer number of polls out there...
Mon Sep 26, 2016, 07:21 PM
Sep 2016

and being quoted leads me to believe that there is some serious data manipulation going on to provide pre-selected results.

The media is loving it because it drives the horse race and the ratings, but I fear for all the disinformation that is bound to be in these myriad polls.

538 isn't helping. Too much noise getting into the model tells me there must be some serious flaws in the underlying methodology and/or the algorithms he's using.

CajunBlazer

(5,648 posts)
13. I wouldn't be believing conspiracy theories if I were you
Mon Sep 26, 2016, 07:32 PM
Sep 2016

The very same statistical methodology using the very same polls allowed Silver to pick the Presidential winner in 50 out 50 states in 2012 and 49 out of 50 states in 2008.

I don't think that there is "data manipulation" in the polls nor do I believe there is anything wrong with Silvers methodology. It looks like people are having a hard time sticking with their earlier decisions in this crazy election year.

It doesn't seem as if Hillary has been losing support so much as Trump has been gaining support - probably because more and more Republicans have decided to hold their nose and vote for him while the Stein supporters are sticking to their guns. That too might change as the race gets closer and Stein voters come face to face with the real possibility of a Trump victory.

Latest Discussions»Retired Forums»2016 Postmortem»Hillary's chances of winn...