"... since polling — to a greater extent than the conventional wisdom acknowledged — had shown a fairly competitive race with critical weaknesses for Clinton in the Electoral College. It’s possible, perhaps even likely, that Clinton will eventually win the popular vote as more votes come in from California. But in a broader sense? It’s the most shocking political development of my lifetime."
I'm reading that this morning because I was wondering how Nate Silver — after getting all those clicks — would protect his lucrative brand.
Subscribe to:
Post Comments (Atom)
Posts
49 comments:
In an extremely narrow sense, Nate's projections were valid.
But in a broader sense, he probably overweighted everything he believed was correct.
Probabilistic projection models mean never having to admit you were wrong.
Hey, maybe Nate can predict the following:
With Hilary's defeat, will all those huge donors to the Clinton Foundation to do "good works":
a.) increase,
b.) stay the same,
c.) drop like a stone?
I mean, it was never about the buying and selling influence, right?
Nate will survive but Sam Wang has to eat a bug!
There's still hope for Hillary. Isn't there some sort of National Human Resources Grievance Committee she can file a complaint with?
In America, money trumps sex. Who'd have thunk it?
Certainly not the woman who spent a decade lusting after her newfound status as a multi-hundred millionaire.
Establishment/Clinton Narrative: We are all women and Muslims.
Trump narrative: We are all working class whites.
The Democratic party should just stop existing.
He's just an ordinary analytical guy who got trapped by many of the things that cause problems in statistical analysis all the time
- relying too much on historical regression (past accuracy of polls),
- poor sample construction (the polls he used were his samples and they didn't represent the population of possible polls well)
-classic meta-analysis defects (studies of studies are historically problematic and in a true analytical effort, they are to be discarded)
Hey, did that left-wing dipshit who promised to eat a bug if Trump exceeded 240 EV done so? I know Madonna backed out of her blowjob promise.
Nate was, honestly, saying that her support was wide but really thin and that her firewall existed, but if one state went down, she had a problem. I thought he was pretty decent.
CA will put crooked Hillary over the top with the pop-vote. If CA were allowed to pick our president, we'd all be screwed forever.
I think Trump should make Hillary Ambassador to Libya. but - she must reside there.
Can Nate predict how long it will be before Megyn Kelly is fired at Fox.
The Murdoch boys from London will want to audition and hire their own Bimbos.
Hey, did that left-wing dipshit who promised to eat a bug if Trump exceeded 240 EV done so?
Supposedly he's working on it
Popular votes in an electoral college system is always going to be misleading. In a place like California where the Presidential results are a foregone conclusion, it depresses turnout both because (a) it feels like voting is pointless and (b) there is little effort put into getting people to the polls. If we switched to a popular vote system, candidates would spend a lot more time and money in California and Texas. To Trump, it didn't matter if he lost California by 5% or 15% or 50%; he's not going to win and he gains basically nothing for even large gains.
His reputation as a seer has taken a hit, but he was nowhere near as obnoxious as most of his peers.
I thought Nate Silver really came through. He appeared to be a scrupulously straight shooter. He went into yesterday with odds in HC's favor between 2:1 and 3:1 and continued to point out that she was within a normal polling error of losing badly. All this when other pundits were making wildly wrong predictions of 95% plus for HC and vociferously chastising him for not falling in line.
When the weatherman says "30% chance of rain tomorrow and the prediction models are volatile", well, if it rains tomorrow, are you surprised. Not really.
"In America, money trumps sex. Who'd have thunk it? "
-- I thought Clinton out spent Trump too?
I'm reading that this morning because I was wondering how Nate Silver — after getting all those clicks — would protect his lucrative brand.
Nate’s silver is tarnished – to anyone with ties to reality, that is. And Real Clear Fuck-up’s bogus poll average should be seen as the nonsense it really is. But they’ll be forgiven because they serve a useful propaganda function. They contribute the false trappings of expertise to the bleating of the pitiful pundits and the witless talking heads. I expect they’ll each be cited as proof of something negative about Trump before a month is up.
But all is good. A false description of reality will only benefit Trump in the long run – just as it has now. Unless I miss my guess they will continue to underestimate Trump; “they” being the eGOP, the Democrats, the MSM and other assorted political miscreants.
Ficta at 9:16, +1. I was on fivethirtyeight all night because they started out NOT predicting a Clinton landslide - just saying that she appeared to have more paths to EC victory than Trump.
He went into yesterday with odds in HC's favor between 2:1 and 3:1 and continued to point out that she was within a normal polling error of losing badly.
Yes indeed - Nate Silver predicted Clinton’s victory and also predicted Clinton’s defeat. Which is to say: Nate Silver didn’t predict shit, really.
In Nate's defence, he' is not a pollster. He takes others' polls and makes predictions on them. At least that is my understanding of his work.
Nate silver is another election loser.
Sam Wang, try one of these:
https://sciencebob.com/wp-content/uploads/2015/04/rhinoceros.jpg
Feel free to insert at either end.
Ficta said...
I thought Nate Silver really came through. He appeared to be a scrupulously straight shooter. He went into yesterday with odds in HC's favor between 2:1 and 3:1 and continued to point out that she was within a normal polling error of losing badly. All this when other pundits were making wildly wrong predictions of 95% plus for HC and vociferously chastising him for not falling in line.
Of all the poll accumulators Nate Silver has come out looking the best, by a large margin. Silver did a good job teasing out a reasonably accurate signal from what turns out to have been quite biased data. You have to applaud him for this. This being said, his success was based more on his personal intuition than on sound statistical principles. Personal intuition inevitably fails at some point. But not this time, so Silver lives to see another day.
The losers are the pollsters, who did a remarkably poor job in aggregate.
First Brexit and now this. I think from now on people will be far more skeptical of poll predictions, especially on controversial candidates and issues.
Nate Silver reputation is only bolstered by this election. It's Sam Wang, Huff Po and Upshot at the NYT who whiffed badly. 538 has given Trump a 30% to 35% chance for the past two weeks and was getting slammed by Wang, Huff Po and NYT for overestimating Trump's chances and accused of doing so intentionally to boost clicks.
Silver also repeatedly noted recently that Hillary had to win the popular vote by more than 1.5% because her votes are poorly distributed. This proved exactly right.
He also noted that Trump only needed a moderate polling error, in the range of 2-3% in national polls to win. That too proved correct.
"The losers are the pollsters, who did a remarkably poor job in aggregate."
But there-in lies the rub ie if pollsters can no longer be trusted, a Nate Silver type is useless.
Rhythm and Balls said... [hush][hide comment]
There's still hope for Hillary. Isn't there some sort of National Human Resources Grievance Committee she can file a complaint with?
In America, money trumps sex. Who'd have thunk it?
Certainly not the woman who spent a decade lusting after her newfound status as a multi-hundred millionaire.
Establishment/Clinton Narrative: We are all women and Muslims.
Trump narrative: We are all working class whites.
The Democratic party should just stop existing.
I want to know how working class blacks, asian, hispanic, Indian and every other "working class" group voted.
Hugh Hewitt is a seer who wrote the book on Trump. It's called "If It's Not Close, They Can't Cheat." Of course, Hewitt was a NeverTrumper until right before the end.
Election 2016 IV: A New Hope
Nate Silver: We've analyzed their attack, sir, and there is a danger. Should I have your ship standing by?
HRC: Evacuate? In our moment of triumph? I think you overestimate their chances.
Lieutenant: Uh... she DID abandon ship though.
Michael K said...
Hugh Hewitt is a seer who wrote the book on Trump. It's called "If It's Not Close, They Can't Cheat." Of course, Hewitt was a NeverTrumper until right before the end.
11/9/16, 9:38 AM
It seems to have been quite a clean election. I wonder if Trump's refusing to prevouch for the integrity of the results, reinforced by the Tuesday Nevada lawsuit, facilitated this. Gosh, that was stupid and crazy of him.
OK,
MSM: Evacuate? In our moment of triumph? I think you overestimate their chances.
Matthew Sablan said...
Lieutenant: Uh... she DID abandon ship though.
11/9/16, 9:43 AM
YES! Ye gods and little fishes, what a coward!
When in 2012 Obama, much to my surprise, won his second term, I made a $100 bet with my wife that Hillary would be the next president. I did so in the fully cynical assumption that the US was absolutely beyond redemption, in the hands of corrupt government and corrupt business and corrupt culture.
It is with great pleasure that today I have to go to the bank and get a nice crisp $100 bill to pay my wife her winnings.
Maybe, if I am humble enough about it, she will use it to take us both out to eat somewhere nice.
To be fair to Silver, he got in a Twitter war with other pollsters that he wasn't as confident on Hillary's chances as others. So, in the end, I think, as pollsters go, he comes out looking better than others.
Thanks to WikiLeaks, we can now point to solid evidence that the DC press corp whored itself out as the unofficial and secret agents of the Democratic Party.
It also exposed the mendacity of the pollsters who slammed their thumbs on the scales, arguing that more Dems go to the polls than Repubs, and that Repub registration was declining and turning independent.
According to the Hill's story on pollsters, they were giving Hillary a 7-8 percentage point advantage when 4 percent should have been the case.
Of course, there was also that polling company that solicited Podesta's business by describing how they rig their polls.
And the Cubs also won.
Nate Silver gave Trump a better chance than most prognosticators That is his defense.
http://elections.huffingtonpost.com/2016/forecast/president
1200-times more accurate than AGW science, though.
Bad Lieutenant said...
Trump's refusing to prevouch for the integrity of the results,
I think even this is going to work to his favor. When the left starts grousing about her winning the popular vote we get to remind them that delegitimzing election results is un-American. According to them at least.
I was bouncing back and forth between here and 538 blog. Never had the TV on once - that media is dead to me.
Anyway, I was watching some of the comments coming in when 538 made it clear that Trump was going to win. Literally, heads were exploding and I think in the right scenario, Nate would have been a dead man.
I thought Silver did pretty well this cycle.
There are acute limitations with our current polling methodologies - the samples are small and are dependent on understanding the demographic political alignments for accuracy. When you have a realignment election, this methodology just does not work.
Went several times to 538 last night for election coverage- it was disappointingly lame and lagging. And I saw that 538 is as proudly progressive as Jezebel or Daily Kos- article after article on wonderful heroic Hillary and mean nasty Trump and his evil white male supporters. Garbage... And 538, like Jezebel, Kos, Vice and other democrat party fluffer blogs went dark before the end came. Poor democrats just couldn't bear to face the music when the Fat Lady started warming up her tonsils...
Silver always allowed for the probability that the polls could be off. He is willing to adjust. Too many in the media are not. These news organizations need to clean out a lot of their political reporters and pundits. Retain or hire those who are open minded
@Michael Fitzgerald Odd. I saw regular posts right up until H conceded. There was some strange behavior in what posts loaded sometimes through the evening. Sounds like their web host choked on traffic maybe.
It's worth remembering that almost every single toss-up state went to Trump. This may be a testament to the great passion of his supporters, but it is also a testament to luck. If the election had been held a week ago or a week from now the results may have been different. We'll never know.
It is the nature of the winner-take-all contest that the closeness of the result is immaterial to the narrative.
Nate Silver, one hit wonder. And from what we know now from wikileaks re media/dem collusion, its not unreasonable to ask if he was getting access to internal poll data.
Real Clear Politics: "There are NO Americans in Baghdad"
Firefly: "You know, in certain older civilized cultures, when men failed as entirely as you have, they would throw themselves on their swords."
Why do these people still have jobs?
Well, Silver did hedge his bets the last weekend, but I think he had help in doing so that some of the other prediction teams didn't get (looking at you Nate Cohn, you chump!).
If I were Silver's shoes last week, I would have been asking direct questions of the Clinton Campaign as to why she was spending money and time in Michigan and Pennsylvania (and Virginia!). It isn't enough to just wave your hands and say she was just hedging her bets because Trump was doing it- you should allow your opponent to waste valuable time and resources contesting a state that is out of reach. Silver is smart guy, and he almost certainly did ask that question, and being that high a profile prognosticator, it is likely that someone in the Clinton Campaign shared the internal numbers with him. This explains, by the way, why the fireworks show was cancelled- the Clinton team was no longer confident of being declared the winner any earlier than late Wednesday morning.
In Silver's defense, his input almost all of the time is the shitty and/or biased polls conducted by the media's polling whores. This is especially true of the state-level polls- the national polls aren't quite as bad, though the ABC/WaPo polling was shit-stained toilet paper most of the cycle. Late last week he started to "adjust" some of the polls, and I read what description he gave for doing so, but it really didn't make a lot of sense to me- that is why I think what he was actually doing was incorporating Clinton's internal polls into his model, but was probably explicitly told not to say that he had such data.
"Hey, did that left-wing dipshit who promised to eat a bug if Trump exceeded 240 EV done so?"
There appears to be a 35% chance he will, plus/minus 35%.
Would you be happy with Nate Silver if you'd bet your life savings on his odds. No. Nuf said.
Post a Comment