When it comes to discussion of probabilities, Nassim Taleb’s opinion carries a lot of weight.
Taleb is author of the New York Times bestseller The Black Swan: Impact of the Highly Improbable.
Using nearly the same logic as I did earlier, Taleb blasted Nate Silver in a series of tweets on election odds.
Tweet #1 – Stochastics
Tweet #2 – Business
Tweet #3 – Variance
Tweet #4 – Mathematical Explanation
Peter Atwater Chimed In On Clinton Bounce
Attitudes the Missing Factor
It is clear that Silver cannot distinguish between the odds of a team winning a ball game the next day and the odds of someone winning a primary a year away or winning an election five months away.
Social mood is in play, and Silver missed it for the entire primary season.
Silver now posts wildly swinging election odds day to day as if he can measure attitudes and sentiment four to six months away.
I commented social attitudes, and on the absurdity of this type of forecasting many times. Here are a number of choices.
- December 10, 2015: Attitudes, Attitudes: Dear Nate Silver, Regarding Donald Trump, You Are Missing Something Big!
- January 31, 2016: Nate Silver’s Continual Underestimation of Donald Trump’s Chances
- June 29, 2016: Here We Go Again: Nate Silver Says “79% Chance Clinton Wins”; Battleground Bloodbath
- August 4, 2016: Peak Hillary
Peak Silliness
Here are a few snips on “Peak Silliness” from my “Peak Hillary” post.
Pure Idiocy
- Supposedly, Hillary has a 79.9% chance of winning in November as of today.
- Supposedly, Trump had a 50.1% chance of winning in November on July 31, just four days ago!
- Supposedly, Hillary had a 77.4% chance of winning in November on July 12.
This is pure idiocy.
Silver is clearly taking the news of the day and projecting it out to November when voters clearly have a time span of about three days.
Social Mood is clearly in control here.
Silver is totally clueless about what social mood will be in November, just as he was totally clueless about social mood the entire Republican nomination process.
Who Better Than Silver?
In response to “Peak Hillary” reader “BH” responded “If not Nate Silver, then who? What better prognosticator is there? Or does Mish just reject that elections can ever be forecast with any significant accuracy?”
Actually, when it comes to this nomination process and election I have done far better than Silver.
If you are looking for someone who understands fat tails better than nearly all of us, then pay attention to Taleb.
I do not profess to be better than Silver at predicting ball games, and Silver might be spot on when the election gets down to the final week.
The closer to the election, the better Silver will get.
Social Attitudes
For now, Silver does not understand social attitudes.
He does not know how to put attitudes into his model, and his wildly changing numbers prove that statement.
While difficult to predict with any degree of certainty, it is a huge mistake to pretend they do not exist.
Attitudes change constantly, but the rate of change in attitudes will slow within 2-3 weeks of the election, and possibly a bit before that, as minds are made up.
Meanwhile, what Taleb labels as “stochastics” is simply huge day-to-day attitude swings based on post-convention bounces and silly things Trump says.
There is no solid evidence that minds have not been made up yet. Until attitudes firm, we are likely to see more huge swings, all leading to ridiculous assessments of the “odds”.
I do not know what the odds are, but I do know they are not (at the moment) 83.1% for Hillary as Silver projects today.
There is simply too much time between now and the election to be that confident, this year, given wild moves in attitudes.
One must also take into consideration voter’s history of eventually ignoring things that Trump says. Curiously, for such a big fan of history, the one piece of history that’s most relevant, Silver ignores.
Mike “Mish” Shedlock
Plus, no one on the left puts in the I am lying to you plug into their models. Everyone knows who they are supposed to be for, and simply will not share that they are not.
LikeLike
The post DNC bounce for Clinton was faked up anyway. Much to be suspected when you see obviously made up pronouncements being trotted out in a synchronized swimming exercise on ALL the major news outlets.
It was transparent as hell.
LikeLike
Trump himself is too “stochastic” to be our leader.
Nate Silver suffers from the same issue that has affected many pundits this cycle – so pointing that out is rather like shooting fish in a barrel.
How about making the case for trump, for clinton, or for someone else, rather than looking to claim superior analysis. THAT is a MUCH more important question at this juncture.
Seems like neither trump nor clinton are fit for office – in character and policy.
So, ha, ha, ha, yes Nate Silver is probably wrong again. Meanwhile… kaboom!!!
LikeLike
Exactly, Maq. To accurately predict some months out, one needs to be able to predict how Trump will behave, and what sorts of things he will say. It’s not completely impossible that he will behave himself, in which case he might win, but at this point it seems far more likely that he will self-destruct.
As for polls showing the race closing, the Ipsos/Reuters poll is the only one that I know of that shows Clinton’s lead growth halting. That poll showed Clinton +5 from July 25-29, and then +4 between July 30-August 3. All other polls I have seen show Clinton’s lead continuing to grow, and in almost all cases show her lead as the widest yet. For example the McClatchy/Marist poll has her at +15, up from +3 a month ago, while the Fox poll shows her at +10, up from +6 a month ago, and from -3 two months ago.
Getting back to Silver, he obviously isn’t basing his forecasts solely on the current situation, but is also accounting for trends. It seems clear that if the election were held today there would be a 100% chance of a Clinton victory, so giving Trump a 15-20% chance seems to account for the possibility that he might get his campaign back on track.
I don’t think there is much chance of that happening, however. I think the mainstream media was complicit in Trump’s primary win, treating his gaffes as humorous in the primary because they viewed him as a candidate that couldn’t win. Now they no longer treat them as humorous, but instead are attacking him relentlessly, and will continue to do so. Will it work? Probably. I think was the only way HIllary could have been elected. I think anyone else would have beaten her.
LikeLike
Gender in this election means more than trust, money, policy, media influence or careless words. When its time to push the button, fill the bubble or punch the card with nobody looking, Ann Coulter could easily vote Clinton and Dan Rather could easily vote Trump.
The difference between male and female voting this election will be absurd. Neither pre-election polls not exit polls will show it.
LikeLike
Damon Runyon–“All life is 6/5 against..” As numerous Presidential elections have shown, predictions several months in advance based on polling are worse than random.
LikeLike
But, you see, the HOPE is that they will act as push polling against the majority gullible/ignorant:
https://en.wikipedia.org/wiki/Push_poll
“A push poll is an interactive marketing technique, most commonly employed during political campaigning, in which an individual or organization attempts to manipulate or alter prospective voters’ views/beliefs under the guise of conducting an opinion poll.”
LikeLike
Election Betting Odds
https://electionbettingodds.com/
LikeLike
The problem with this site is that that it uses BetFair where there is no U.S. betting, so accuracy of U.S. public sentiment that is “willing to put it’s money where its mouth is” is nil. All it reflects is the mainstream Euro media freaking out over Trump and the state of this wacky election cycle.
LikeLike
I recall before the Brexit vote some deep pocket entity made enormous, percentage skewing bets in England, attempting in my opinion to hijack what was popularly seen as an accurate predictive indicator in favor of the elite “stay” position. Now that this Rubicon has been crossed it degrades the usefulness of such betting stats to gauge sentiment.
LikeLike
I don’t think that polling publications actually portray anything except support of the narrative. In the end, the corporate oligarchs get their candidate that shoes in more and more regulation for themselves. Since Reagan, you can see ever bolder moves in controlling MSM to create the “illusion”. We still have the Patriot Act, they use the Constitution as an election ploy, then after the votes are cast, it becomes their toilet paper. TPP, ect is another step in their plans and they are going to get it if they have to knock off a few more people in the process… so what. it’s their club and you have been denied entry. George Carlin was 100% accurate… and he was just a comedian.
LikeLiked by 1 person
“It’s a Big Club and you ain’t in it. You and I are not in the Big Club.”
LikeLike
Yep, that’s the quote. t’s too bad Dick Cheney is not quite in the big club either… maybe they’d all go hunting together.
LikeLike
“Using nearly the same logic as I did earlier, Taleb blasted Nate Silver in a series of tweets on election odds.”
…with the only real difference being Taleb actually knows what he’s talking about, aside from that, the logic is identical.
LOL
LikeLike
That said, (((Silver))) and the rest of his ilk make a living contributing nothing of any tangible value to mankind….been that way for hundreds of years. (((They))) merely serve as middle-men and rabble-rousers….parasites skimming their vig from the public at large.
LikeLike
And lastly, anyone who spends for than a millisecond thinking any of these polls have the slightest meaning re: the outcome of the election is a dolt. Polls are, and have always been, just a fraudulent “rigged” mechanism that those in control use to sway public opinion.
LikeLike
Perhaps Nate is on the Take? As an employee of the Clinton foundation? Good money for spin I bet.
For a few 100k I’ll state cankels will win by 100%
LikeLike
To echo what others have said: the guy works for the NYT and the NYT wants Hellary. It is his job to help make that happen by creating the myth of inevitability and to give the impression that “Everyone ELSE is for her, why are you for Trump, stupid?” It is a form of “gaslighting” although that basically is what the entire MSM is these days.
LikeLike
This is like Giuseppi and Luciano arguing over who can spray the farthest.
LikeLike
http://freedomoutpost.com/dnc-insider-clinton-being-pounded-in-polls-dont-believe-mainstream-media/
Occam’s razor slits a throat.
Taleb doesn’t consider that Silver might be a manipulator.
Frontrunning is not a black swan event.
Hacking or releasing a rumor in background media is not a black swan event.
LikeLike
And then there is Elliott Wave International that has a completely attitude based technical analysis forecast system. They don’t claim odds, but explain one candidate stands a better chance of winning than the other candidate based on the fractal nature of social mood patterns.
Hillary’s popularity rating (based on chart patterns using data since the early 1990’s) peaked a few years ago. Any gains in popularity leading up to the election will only be a bear market rally in nature. Regardless of the election, her popularity is expected to plummet. She could be the worst president since Jimmy Carter or Barack Obama. Carter doesn’t deserve “worst president ever” since he is a nice person who was in office during strong negative social mood.
MISH is definitely looking at elections through the correct lens; however, he did not invent those glasses.
LikeLike
In retrospect, Carter appears to have been a much better President than he appeared to be at the time. I was glad to have him gone, but I have to admit that, unlike the Presidents who followed, he was a fiscally responsible President.
LikeLike
I am a Trump supporter and not a poll supporter. “I want to poll you” is the quickest way to get a “click” or internet e-mail deleted. Most of my friends a Trump supporters and feel the same way about polls. Hillary can get 75% of the people that like polls and easily lose the election because I and my friends will vote.
LikeLike
I have yet to see a Hillary yard sign. Trump signs have been up for months.
LikeLike
Mish’s criticism of Silver is broader than Teleb’s. Mish has two points, Teleb one. Mish says Silver’s fails to recognize that voters’ attitudes have changed, or that Silver has failed to sufficiently weigh these attitude changes in his model. Teleb does not raise the content of Silver’s model, and specifically Teleb does not mention voter attitudes. Mish also asserts that Silver’s forecasts are defective because they too much, too often. It is this point that Teleb elucidates in his Tweets blogged above.
The question still remains. Can elections be forecasted months in advance with any degree of accuracy? If so, how? It is not enough to say that Silver’s forecasts are flawed. How flawed are they? Worthless? Does Mish offer an election forecast, or just an infrequent macro, conclusory overview? How often should a forecast be updated? Maybe discussions about the future outcomes of elections should avoid quantifications besides poll data.
LikeLike
Revised:
Mish’s criticism of Silver is broader than Teleb’s. Mish has two points, Teleb one. Mish says Silver’s fails to recognize that voters’ attitudes have changed, or that Silver has failed to sufficiently weigh these attitude changes in his model. Teleb does not raise the content of Silver’s model, and specifically Teleb does not mention voter attitudes. Mish also asserts that Silver’s forecasts are defective because they change too much, too often. It is this point that Teleb elucidates in his Tweets blogged above.
The question still remains. Can elections be forecast months in advance with any degree of accuracy? If so, how? It is not enough to say that Silver’s forecasts are flawed. How flawed are they? Worthless? Does Mish offer an election forecast, or just an infrequent macro, conclusory overview? How often should a forecast be updated? Maybe discussions about the future outcomes of elections should avoid quantifications besides poll data.
LikeLike
“I do not know what the odds are, but I do know they are not (at the moment) 83.1% for Hillary as Silver projects today.
There is simply too much time between now and the election to be that confident, this year, given wild moves in attitudes.”
Given this in the original post, I suppose my question in my post above was not necessary, and I should have read the post more carefully.
Still, accepting that the odds are not 83.1% for HRC, can a range be specified now with any reasonable degree of confidence. Say 60-70% HRC 55-80%? Can it even be said now with any reasonable degree confidence that HRC is more likely to win than Trump?
LikeLike
I suspect most would conclude Hillary has the better shot. Even I conclude that. Trump has made a serious mess of things. Is it 60-40? 65-35? I don’t know. Taleb did not assign any odds, he just went after Silver.
Somewhere in the low 60’s seems reasonable enough, perhaps 65%. A quick check shows Silver now assigns over 86%.
LikeLike
There are lots of sites where you can bet on the outcome. I did a check of several of them, and they all are closer to Silver’s forecast than to yours. That doesn’t prove anything of course, other than that Silver has a lot of company.
LikeLike
Who Better Than Silver?
In response to “Peak Hillary” reader “BH” responded “If not Nate Silver, then who? What better prognosticator is there? Or does Mish just reject that elections can ever be forecast with any significant accuracy?”
Of course elections cannot CONSISTENTLY be predicted with any significant accuracy. Any more than anything else in the social/economic domain can. Accurately predicting what sentient, optimizing actors will do in the future, will just lead said actors to reevaluate based on your predictions. Like playing Rock Paper Scissors with some guy who publicly makes predictions and believes in them.
LikeLike
Dear Mish,
Silver, Washington Post, New York Times, CNN, MSNBC other major media (except FOX) are trying to portray Clinton as doing fine, possibly as a precursor to election fraud. See https://www.washingtonpost.com/politics/poll-after-conventions-clinton-leads-trump-by-8-points/2016/08/06/517999c0-5b33-11e6-9aee-8075993d73a2_story.html . Notice the obvious lies in the article, such as:
“The Post-ABC poll shows Clinton winning 92 percent support among self-identified Democrats. That compares with 86 percent support just before the Republican convention and is an indication that the Democratic convention helped consolidate supporters of Sen. Bernie Sanders (Vt.) behind her candidacy. Those Sanders Democrats favor Clinton by 86 percent to 5 percent over Trump, larger than the 79 percent to 10 percent in July.”
See how HUGE Trump’s popularity is on Facebook and Social Media, compared to Hillary’s for yourself at http://thetruthdivision.com/2016/08/spread-this-media-rigging-the-polls-hiding-this-evidence-proving-trump-is-winning/ “
We all know that a sizable portion Grassroots Democrats are against the establishment and that was their main reason for supporting Sanders (like they supported Ron Paul in 2008 and 2012). The DNC election fraud/collusion with Clinton Campaign to rig the primary in favor of Hillary, further angered Sanders supporters, who even booed Sanders for his endorsing Hillary (http://www.cnn.com/videos/politics/2016/07/25/bernie-sanders-boo-hillary-clinton-endorsement.cnn )
Sanders supporters do NOT trust Wall Street Agent, Hillary, habitual liar, warmonger. https://www.youtube.com/watch?v=YECWYBe7CtU For Washington Post to say: “Sanders Democrats favor Clinton by 86 percent to 5 percent over Trump” ignores the fact that many of them will NOT vote for Hillary and others will vote for Johnson and Stein. In short, probably close to half of Sanders voters will NOT in some shape or form vote for Hillary.
Combine this, with the fact that Trump has gotten the highest Republican primary turnout in decades, will great number so independents, as well as Trump’ anti TPP position, makes it clear that these “pools” must be rigged, as they did when Ron Paul was running in 2012. See https://www.youtube.com/watch?v=drqw6AX-6wQ and https://www.youtube.com/watch?v=MkvgtOEWq2I
Silver relies on these biased, rigged polls and statements to make the same kind of erroneous conclusions that he made about Trump in the primary.
Your’s in freedom and justice,
Peymon Mottahedeh, President
Freedom Law School
http://www.LiveFreeNow.org (813)444-4800
LikeLike
Pingback: Says Who? – of Photons, Protons, and Proteins
Pingback: Global Trumpism | Elect Steve Balich
Pingback: Predicting the Wave of Nationalism | Munk – Global Conversations
Taleb is misspelled in the title
LikeLike