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

paulthompson

(2,398 posts)
Wed Mar 30, 2016, 05:39 PM Mar 2016

Nate Silver is about to be wrong in Wisconsin

Lately, there have been news reports like this one, suggesting that Clinton is pulling out of Wisconsin. She and Bill basically have no more events scheduled there for the next week:

https://www.americarisingpac.org/team-clinton-desperately-trying-to-lower-expectations-in-wisconsin/

In my opinion, that's confirmation of the late Sanders surge trend that I and others have put forward. See here:

http://www.democraticunderground.com/12511576753

The Clinton campaign knows that Sanders' numbers always go up as the election date gets closer, never down (outside the South, that is). So if Sanders is already tied or thereabouts one week prior to the election, that means he's going to easily win, and it's just a matter of by how much.

Note the above video came out BEFORE the new poll results from a few hours ago showing Sanders moving into the lead in Wisconsin for the first time. So even Sanders being down slightly a week in advance means he's very likely to win.

However, supposed election numbers expert Nate Silver and the folks at fivethirtyeight.com have yet to figure this out. He still has Wisconsin as a toss-up. He's always a couple of steps behind this election season, because he still hasn't noticed the Sanders late surge trend. And he always has a "fudge factor," which he calls "polls plus," in favor of Clinton. Wisconsin is a typical example. According to an average of pure poll data, Silver has Sanders slightly favored to win there, 56% to 44%. But his "polls plus" has the odds of Clinton winning at 64% to 36%.

So according to Silver, Clinton has a 2 to 1 chance to win the state, when news reports are pointing out that the Clinton campaign is conceding the state in order to concentrate their efforts on New York. If Silver is right, someone needs to tell the Clinton campaign not to pull out of the state!

What I predict will happen is that more polls will come out leading up to the election date showing Sanders in the lead, so the poll average and even the "polls plus" numbers at 538 will eventually come to reflect that. But, as always, Silver will be behind what everyone else already knows, which is that Sanders will win Wisconsin, and the only question is the margin.

And by the way, up until yesterday, 538 had the odds of Clinton winning Wisconsin over 85% for both the poll average and the "polls plus," when every other analyst knew it would be much closer race. Did even most pro-Clinton supporters here think Clinton has an 85% chance of winning Wisconsin? I doubt it.

As always this election season, Silver is behind and has a pro-Clinton tilt. I don't know if it's because he's rooting for Clinton, or just that he repeatedly fails to see the late Sanders surge trend, or both.

36 replies = new reply since forum marked as read
Highlight: NoneDon't highlight anything 5 newestHighlight 5 most recent replies
Nate Silver is about to be wrong in Wisconsin (Original Post) paulthompson Mar 2016 OP
If Sanders wins WI it would not matter much if Hillary blows him out upaloopa Mar 2016 #1
I think it was put away on March 15th. Garrett78 Mar 2016 #7
May April fly by... FarPoint Mar 2016 #28
I think I heard it "will be over by March 1st" ! John Poet Apr 2016 #35
Silver is worthless this year. A static model in a dynamic situation can't work. nt thereismore Mar 2016 #2
Predictwise paulthompson Mar 2016 #3
or they're about to Maricopa the state MisterP Mar 2016 #4
He seems to usually give her a large chance of winning in states that vote on computers. stillwaiting Mar 2016 #5
More misleading gossip about her abandoning the state. Lucinda Mar 2016 #6
Yes, she did paulthompson Mar 2016 #8
Hillary has been to Madison, Milwaukee, La Crosse and Green Bay already Lucinda Mar 2016 #12
Past vs. future paulthompson Mar 2016 #16
Shades of the Cruz Camp suggesting that Carson was dropping out. Garrett78 Mar 2016 #15
Yep. They did to Hillary before with one of the earlier states, but I don't remember which one. Lucinda Mar 2016 #17
So you're upset with the Clinton campaign? paulthompson Mar 2016 #19
They were right about Carson...they were just right too early... Ken Burch Mar 2016 #24
Sorry, but I'm not reading through all that. MerryBlooms Mar 2016 #9
Yr pst = 2 mny lttrs demwing Mar 2016 #20
Noted MerryBlooms Mar 2016 #22
YEP! demwing Mar 2016 #31
Funniest photo ever! MineralMan Mar 2016 #32
Just call him 'Nate the not-so-Great' 4139 Mar 2016 #10
Rather than Nate Silver, he's Nate Tin Ear jfern Mar 2016 #18
feels so right reddread Mar 2016 #11
What does that mean? paulthompson Mar 2016 #14
i never believed their scrapings. reddread Mar 2016 #21
she campaigned in Milwaukee today, Wednesday 3/30 amborin Mar 2016 #13
Dear DU, OilemFirchen Mar 2016 #23
What about CBS? paulthompson Mar 2016 #25
It's called Data Science for a reason RAFisher Mar 2016 #26
adapt! paulthompson Mar 2016 #29
Thanks Dem2 Mar 2016 #30
WI is an Open Primary so they have no real polling day manipulative powers. NorthCarolina Mar 2016 #27
Well, if Nate's going to be wrong, then Hillary is going to win WI? Dem2 Mar 2016 #33
america rising pac sharp_stick Apr 2016 #34
I believe Silver's recipe for combining an average of the polls from Democrats Samantha Apr 2016 #36

upaloopa

(11,417 posts)
1. If Sanders wins WI it would not matter much if Hillary blows him out
Wed Mar 30, 2016, 05:44 PM
Mar 2016

in NY and PA.

I think I heard Hillary can put this thing away by the end of April without WI.

Garrett78

(10,721 posts)
7. I think it was put away on March 15th.
Wed Mar 30, 2016, 05:57 PM
Mar 2016

Chances are her pledged delegate lead will be in the ballpark of 350 by the end of April. At that point, I imagine more people will accept the inevitable. Not everyone, mind you, but some.

FarPoint

(12,206 posts)
28. May April fly by...
Wed Mar 30, 2016, 07:24 PM
Mar 2016

As swiftly as March...I'm ready for closure.. Hillary will finish this by the end of April.

paulthompson

(2,398 posts)
3. Predictwise
Wed Mar 30, 2016, 05:49 PM
Mar 2016

I just checked the predictwise website. According to their data, Sanders has a 79% chance of winning Wisconsin, while Clinton has a 21% chance.

Compare that to Silver's "polls plus" number that has Clinton with a 66% chance of winning. Very different! What does he know that everyone else is missing out on?!

I would say "nothing," because his "polls plus" predictions have been off in a Clinton direction time and time again. Yet apparently he doesn't want to update his model.

stillwaiting

(3,795 posts)
5. He seems to usually give her a large chance of winning in states that vote on computers.
Wed Mar 30, 2016, 05:53 PM
Mar 2016

And, he gives Sanders chances to win in states that don't vote on computers.

He was wrong in Michigan. If votes are being stolen, and I do believe they are since I can't think of one good reason why they wouldn't do so, I guess they either didn't steal enough in Michigan or they didn't steal votes in every state that votes on computers.

Until we organize to demand a change to how we conduct our elections, election results have no credibility for me.

It has been interesting this primary season for sure.

paulthompson

(2,398 posts)
8. Yes, she did
Wed Mar 30, 2016, 05:59 PM
Mar 2016

But watch the video. There's still a week to go until the Wisconsin election day, and that's the ONLY state voting from now until then, so there's no big competition for time and money. There are surprisingly few Hillary or Bill Clinton events scheduled in Wisconsin for that time. Does that mean nothing to you?

Lucinda

(31,170 posts)
12. Hillary has been to Madison, Milwaukee, La Crosse and Green Bay already
Wed Mar 30, 2016, 06:09 PM
Mar 2016

Chelsea has been to Madison, Milwaukee and Waukesha.

All together there have been 9 separate WI campaign events since the 24th. There was a mention about another Hillary visit to be scheduled in the next few days in an article I read yesterday or the day before. They did not ignore or abandon the state.

paulthompson

(2,398 posts)
16. Past vs. future
Wed Mar 30, 2016, 06:15 PM
Mar 2016

I'm not talking about the past, I'm talking about the future. If there are new reports that the Clinton campaign has just decided to abandon Wisconsin, they can't go back in time and cancel events in the past.

I would be surprised if Clinton doesn't show up there at all in the next week. But the normal thing would be to be there most every day, having more than one event most days. Clinton seems more focused on New York already.

Lucinda

(31,170 posts)
17. Yep. They did to Hillary before with one of the earlier states, but I don't remember which one.
Wed Mar 30, 2016, 06:17 PM
Mar 2016

Pretty shady way to try and influence voters IMO.

paulthompson

(2,398 posts)
19. So you're upset with the Clinton campaign?
Wed Mar 30, 2016, 06:29 PM
Mar 2016

Because it's the Clinton campaign trying to lower expectations in Wisconsin. Watch this video, from today:

MerryBlooms

(11,728 posts)
9. Sorry, but I'm not reading through all that.
Wed Mar 30, 2016, 06:00 PM
Mar 2016

Last report on my FB news feed, Sanders was up by 4%, and that's good enough for me.

paulthompson

(2,398 posts)
25. What about CBS?
Wed Mar 30, 2016, 07:06 PM
Mar 2016

See this, from today:



I'm sure we'll hear more of the Clinton campaign lowering expectations in Wisconsin in the coming days.

RAFisher

(466 posts)
26. It's called Data Science for a reason
Wed Mar 30, 2016, 07:08 PM
Mar 2016

I'm a Bernie support but jeez. It's a model. They can be wrong. Statistical models have to have consistency. I can't even tell if the methodology of the 'Poll Plus' is modified. Is it adaptive or is it the same for all elections? From what I'm reading it looks like the methodology has remained the same since January. Even by his own admission back in January he said this model might be completly wrong. If you don't like the Poll Plus model then just ignore it.

It's a science, you are allowed to make mistakes. Silver is trying make a better model. Sometimes it does not work. It's called science.

paulthompson

(2,398 posts)
29. adapt!
Wed Mar 30, 2016, 07:30 PM
Mar 2016

Talking about science, if you find you're wrong, then you change your methods until you get it right. There's no rule that says Silver has to stick with his same model for the entire election season.

If his "polls plus" modifications are wrong, I'm free to not only ignore them, but call them out for being wrong. A lot of people trust those numbers because Nate Silver built up a great reputation from past elections. I thought he did really well in past elections too. But I would argue that this election is different, and he hasn't adapted.

To give one example, Silver subscribes to the theory that high-profile endorsements are very predictive as to who will win. This has been a cardinal rule for poll analysts for many years, and it's a big part of his "polls plus" modifications for each state. But this year, with voters in a rebellious mood, endorsements seem to have no effect, and on the Republican side at least, sometimes a negative effect.

Look for instance at Sanders' landslide wins in Alaska, Hawaii, and Washington. Clinton had nearly all the big endorsements in those states, but that seemed to matter not at all.

Any intelligent scientist would adjust one's model after having it be wrong so consistently. Nate Silver has not, so he deserves our scorn.

This matters because in politics, momentum matters. Voters have a "rally around the winner" tendency. That's why the campaigns for both Sanders and Clinton try to manage the expectations game. The above link to the CBS is just one example of that.

 

NorthCarolina

(11,197 posts)
27. WI is an Open Primary so they have no real polling day manipulative powers.
Wed Mar 30, 2016, 07:14 PM
Mar 2016

Best to just leave the state and cut your losses.

New York OTOH, being a closed primary, I fully suspect to see a re-do of the AZ Primary fiasco complete with voter disenfranchisement, voter registration info "snafus" and a race called solely on the basis of early voting results while thousands still wait in line to vote.

Dem2

(8,166 posts)
33. Well, if Nate's going to be wrong, then Hillary is going to win WI?
Thu Mar 31, 2016, 04:51 PM
Mar 2016

Chance of Winning:

Polls Plus forecast: Bernie 57, Hillary 43

Polls Only forecast: Bernie 74, Clinton 26

sharp_stick

(14,400 posts)
34. america rising pac
Fri Apr 1, 2016, 03:15 PM
Apr 2016

I'm sure they thank you for doing their job for them. Well done.

https://en.wikipedia.org/wiki/America_Rising

Absolutely vile shit being posted on this site recently, way to go with the flow.

Samantha

(9,314 posts)
36. I believe Silver's recipe for combining an average of the polls from Democrats
Fri Apr 1, 2016, 03:30 PM
Apr 2016

is one that does not work. From the very beginning of this primary, because I had studied Sanders' campaigns in Vermont, I believed that it would not be a majority of Dems taking Sanders into the lead, it would be a majority of Independents, a peel-off of some Republicans and a respectable amount of Dems. At the time I first predicted this, he was getting an average of 51 percent of Independents, and that percentage has slowly increased.

Some of the pollsters in Silver's mix are pretty accurate, but others are average to poor. It is that average to poor being thrown into the pot that throws his percentages off. But the bigger problem is that one cannot predict the human factor, those that are undecided and make up their minds at the last minute. And I believe many of those are Independents and some Republicans.

This is just not a run-of-the mill Presidential election. There are a lot of anomalies in this one that are difficult to foresee.

I have a lot of respect for Nate Silver, btw. He is usually dead on.

Sam

Latest Discussions»Retired Forums»2016 Postmortem»Nate Silver is about to b...