Get Free Email Updates!
Join us for FREE to get instant email updates!
Brexit, Now Trump. It’s Official – Political Election Polls Are All Bogus
The world can be successfully navigated as long as you have the data. If you simply follow the data and adjust your actions accordingly, even if you emotionally disagree with what the data is telling you, then you’ll be successful. This maxim has been the backbone of my life.
-By Caleb Jones
As I reported on this blog a few months ago, the UK recently passed Brexit. The odd thing about this was that ALL the polls, ALL the pundits, and ALL the betting pools, which many believe are more accurate than polls, were wrong. Dead wrong.
I talked about how the polls and betting pools probably can’t be trusted anymore. I wasn’t sure about it going forward, but clearly that was the case with Brexit.
Now you know what happened with the Hillary/Trump election. Pretty much everyone in the universe with real data behind them was convinced that Hillary would not only win, but would win somewhat easily. True, there were the alt-right internet devotees who had always been screaming that “Trump will win in a landslide,” but admittedly, these folks just had a lot of hope, passion, and anger, not data.
Though he lost the popular vote (since America is still a left-wing country and will become more so as time goes on), Trump still won, shocking the punditry and elites alike. Again, the polls, punditry, and betting pools were completely wrong.
Thinking this might be the case, four days before the election, I took a snapshot of all the national polls and the swing state polls, which are the only states that matter. I kept these until the day after the election to see how accurate these polls were. My source for all these polls was the average, aggregate polls at realclearpolitics.com.
Here was my snapshot of the aggregate polls for the swing states on November 4th; four days before the election:
As you can see, Hillary not only wins, but wins pretty easily with a reasonable margin for error.
How did the actual election turn out? It looked like this:
Isn’t that amazing? Out of 11 swing states, the polls got five wrong (correction, they got four wrong; there's an error above). That’s an almost 50% inaccuracy rate. Moreover, many of these wrongly called states were high electoral states. The polls had Hillary winning by 40 electoral votes in these states, but Trump won them, and won them by 82!
Even in the states the that polls called correctly, the polls were still way off in terms of how many votes the candidates actually received. Look at this:
What about national polls? Check out this snapshot I took on November 4th showing polls for all four presidential candidates for the entire five months before the election:
As you can see, Trump was clearly trailing in the polls almost the entire time, yet he won the election.
Here’s some more detailed numbers for the national polls during the week before the election:
Remember, these were the polls a week before the election, so the usual excuse of “it’s too soon for the polls to be accurate” doesn’t apply. These were the real numbers a week before the day of the election, and they were still wildly, insanely wrong.
Why did everyone get this so wrong? I’ve read several articles on this topic, and the short answer is that there is no answer, or at least no explanation that has any consensus. Some say it’s because people now lie in polls more often. Others say it’s because it only applies to people with land phone lines, which tends to skew the samples. Still others say it’s because flaws in how they use demographic modeling.
I can’t say definitively why the polls this year on things like Trump and Brexit have been so horribly wrong, but the fact that A) they’re wrong and B) the experts can’t even explain why they’re wrong, shows me that the system is continuing to break down. Either the breakdown is for technical reasons, or it’s because people in the Western world are becoming more and more irrational. It doesn’t matter. This is a sign of the times, folks.
As of now, I will no longer believe or trust any poll that attempts to predict the outcome of a national election anywhere in the Western world. This really sucks, because it’s going to make it that much harder to anticipate future events, and how to profit from those events. Yet, such is the reality of living in a slowly collapsing empire.
I will still pay attention to polls that ask for people's opinions (not elections, but opinions), but only if the polls are clearly weighted to one side or the other. But, as far as “how are you going to vote?” type of polls, I'm ignoring them as of right now; they're useless to me at this point.
Enjoy the decline, folks. It’s going to start getting much more hairy from here.
Want over 35 hours of how-to podcasts on how to improve your woman life and financial life? Want to be able to coach with me twice a month? Want access to hours of technique-based video and audio? The SMIC Program is a monthly podcast and coaching program where you get access to massive amounts of exclusive, members-only Alpha 2.0 content as soon as you sign up, and you can cancel whenever you want. Click here for the details.
Get Free Email Updates!
Join us for FREE to get instant email updates!
D 2016-11-17 14:25:00
The British comedian Jonathan Pie provides a plausible explanation in this part of his "rant"... https://www.youtube.com/watch?v=GLG9g7BcjKs#t=3m40s
Qlue 2016-11-17 15:27:42
Call me a conspiracy theorist but I believe this all has to do with Julian Assange, they literally took him offline and raided the embassy to prevent him from posting a final leak that would probably put Hillary Clinton in jail for life. The latest leaks posted by wikileaks don't match his deadman's switch hash key, which means wikileaks has been compromised. The wikileaks official subreddit censors any thread asking about Julian Assange. He hasn't been seen since mid October, since the shut down happened by John Kerry. We know that the Panamanian company that implements and owns the voting machines is funded by George Soros who donates to Hillary and pays for liberal activists, so it doesn't take much to pad the votes to give Trump the election and then have Hillary concede/forfeit early on like she did. If there is a secret government, then this is how it went down. Some guys even tracked the plane that Assange would have been on to the US, it's called the Guantanamo express, look it up on google images. Hillary was supposed to win that's why the polls were accurate until Assange threw a wrench into her plan in October, which Trump capitalized on. Trump IS part of the establishment, he was supposed to be a pied piper and wasn't supposed to win. That's why he looks clueless when he's with Obama, exactly like South Park portrayed him. I honestly don't think he wanted to be president, why would a billionaire who can have anything he wants take a stressful job that would age him drastically and lead to an early death? Trump is too selfish and smart for that stress drama non sense. Anyway, I don't think there is a true democracy in the US with all the machines, electoral college, first past the post system, gerrymandering, corporate donations, lobbyists, media control and censorship, etc.. it's glaringly obvious.
Tale 2016-11-17 16:52:06
It's also worth mentioning the 2015 UK general election as part of the trend, where the most likely outcome was a hung parliament requiring a second election, with a Labour/SNP coalition the only other outcome considered possible from polling / expert opinion. The actual result was a Tory majority government, which looking back it seems the betting market gave a ~10% chance. Funnily enough this unlikely result was the only way an EU referendum could occur, leading to Brexit. Like you say no-one seems to know for sure why this is happening, and it's definitely motivated me to speed up my back-up plans just in case this unpredictability bleeds into other areas.
Caleb Jones 2016-11-17 17:38:50
Qlue I think your conspiracy theory is a little too far-fetched, but I agree completely that Trump never planned on actually being president, for several reasons, including the ones you stated.
JJI 2016-11-17 18:09:29
I bet money on both Brexit and Trump winning long before the election was held. I bet in March on both outcomes (June 23rd for Brexit and Nov 8th for Trump). My tip for how to predict the next election is to simply look at the polls and how they sample... the compare that with a poll that doesn't change its sample. Let me explain. What many members of the MSM did (most of them backed Clinton) was to oversample for Dems and Women and Blacks in crucial periods (like after the debate) to make it appear that Clinton had "won the debates" but in fact she had not. The most accurate poll of 2012 was The RAND Continuous 2012 Presidential Election Poll ( here: https://alpdata.rand.org/index.php?page=election2012 ) This poll is different from ordinary polls in that it doesn't ask different people each time, but instead asks THE SAME GROUP of people and tracks their opinions EVERY DAY. In this case, 6000 people, then weighted to match the ACTUAL demographics of the USA (no oversampling possible). The USC Dornsife poll adopted this methodology for this election and you can see the results are very close to the actual outcome. They had Trump winning 15+ occasions during the entire election cycle, and Clinton winning only once (after the DNC) during the same cycle. Additionally, there are two models which have very high accuracy 1) The Primary Model (as highlighted in OP) which is based on the Norpoth method of looking at primaries and then making a prediction based on that.. he predicted a Trump win with 87% certainty. He is now 6 for 6 in terms of election forecasting 2) the Lichtman model aka. "The Keys to the White House" which also predicted a Trump win, boasts a retroactive near perfect record (aside from 1960). Lastly, bookie prices are not predictions but simply reflections of where the big money has gone. Bookmakers aim to make money regardless of who wins (and they usually do! vegas didn't build itself!). William Hill is one of the largest bookmakers in the UK and it revealed that it had two HUGE bets placed on Clinton very early on in the campaign (around $1,000,000) that heavily slashed her odds. The pragmatic and cynical side of me thinks that this bet was placed by friends of the Clinton campaign betting this kind of money to slash her odds sharply and therefore make it look like she was more of a betting "favorite". Considering all of this information and with bookmakers offering prices of 5/1 for Trump after hearing all of this, it was an EASY +EV bet. (look up +EV, or Expected Value, in gambling and poker) Next election, use the USC poll, the Lichtman's 13 Keys to the White house and the Norpoth Primary Model to place your bets, and especially when you get a great price like +400 (4/1, 5.0)
Caleb Jones 2016-11-17 19:10:30
Very interesting. Good info. I'm looking into this.
Shura 2016-11-18 05:47:31
My god, sometimes I think you troll, Mr Jones. You talk about following the data and then go on to consider that a 45.6%vs46% poll is predicting a certain winner. You are mature enough to know the basic mistake there. I don't know what you have read, but the answer to why polls don't work is obvious: The wrongly-called shy vote. It's not shy, but fearful. Left-wing has managed to make people afraid to voice opinions contrary to the mainstream (media). To half the country it is shameful to vote conservative to the point that people feel their own safety, property and even jobs are in danger if their opinions were known. Ask yourself this: How easy do you find to imagine someone being fired for saying he voted for Clinton? And for Trump? You know it doesn't feel the same. I have dated 3 American girls in the last few weeks. The 3 reported crying on Election Night (I avoided telling them about my Election night, at a foreign bar wearing a Trump T-shirt, but it was fun) and one of them broke into tears upon recalling it. I know an American father living here 2 of whose 4 children have announced they shall not come home for Christmas because he voted Trump. My British cousin regularly calls Brexit supporters (all of them, half of the country) racist xenophobes on Facebook. She is a total bigot and my whole extended family like her links to whatever nonsense du jour The Guardian is spewing, as well as her rants against leavers for ruining "her country". No one seems to have any problem with her intolerance. No one seems to like that I respectfully point out the validity of every point of view and the need to avoid insults (I don't care in the slightest, just trying to spread bonhomie). Therefore I have no problem believing even the tales of children crying at school because monster Trump were going to eat them alive for breakfast. This is mass hypnotism of a large part of the population. Hence, "shy" vote. However it's not a problem at all. Polls are very useful if you count 3% additional support for the conservative choice. That's the average overperformance in your analysis, as well as what happened in Brexit. Why would you disregard them? Just apply an offset!
Caleb Jones 2016-11-18 11:31:35
My god, sometimes I think you troll, Mr Jones.Incorrect. I always give numerous facts / stats to back up any opinion I state. I don't just toss around emotional insults and invective for no reason.
You talk about following the data and then go on to consider that a 45.6%vs46% poll is predicting a certain winner. You are mature enough to know the basic mistake there.If you look at my data above, it's more than that. It's making that same mistake over and over again in numerous polls, numerous different polling techniques, in numerous regions, and by numerous different margins.
I don’t know what you have read, but the answer to why polls don’t work is obvious: The wrongly-called shy vote. It’s not shy, but fearful. Left-wing has managed to make people afraid to voice opinions contrary to the mainstream (media). To half the country it is shameful to vote conservative to the point that people feel their own safety, property and even jobs are in danger if their opinions were known.Yes, people lying in polls could be one of the reasons, as I clearly stated in the article.
Hence, “shy” vote. However it’s not a problem at all. Polls are very useful if you count 3% additional support for the conservative choice.You're pulling that 3% figure out of your ass, and it's inaccurate, as I'm about to show you. If there was a very clear and specific methodology to determine that shy vote, then we could apply it to all polls to get a high degree of accuracy across the board. This is not the case, if you simply look at the AP data I posted above. Trump's overperformance ranged from negative 3 percent to positive 8 percent. It's all over the map. So no, political election polls are still useless to me at this point. Whenever they come up with a reasonably accurate, scientifically-determined dataset for measuring the amount of right-wing pussies who lie in polls, you let me know.
Makeshift 2016-11-18 13:57:30
It's simple: he who controls the memes controls the world. Kek chose the God Emperor as his prophet, and his will is absolute.
Ron1 2016-11-21 01:54:18
Florida - polls: Trump, winner: Trump. Shouldn't there be a Yes in second table (=correct prediction), 4 not 5 wrong predictions?
Shura 2016-11-21 04:01:02
There are ways to account for shy vote, like Trafalgar Group did (https://twitter.com/trfgrp) but what incentive do you find for the liberal media to pay for them? The kind of indirect questions they use has been known to be effective by sociologists for a long time (asking not about the respondent, but about what he sees around him). The 3% is a rule of thumb, yes Duh, it's not me who started playing fast and loose with the numbers. 😉 However, notice the outliers: Utah, where a third party candidate predictably got quite less that what he was polling, and Nevada and New Mexico, notorious for democrat fraud by way of illegal immigrants.
Caleb Jones 2016-11-21 10:23:53
Florida – polls: Trump, winner: Trump. Shouldn’t there be a Yes in second table (=correct prediction), 4 not 5 wrong predictions?Shit. Yep. Sorry for the error.
There are ways to account for shy vote, like Trafalgar Group didI need to see it laid out in a very specific and predictable way across the board. When I get some time I'll check them out, but I doubt they have what I'm asking for.
but what incentive do you find for the liberal media to pay for them?There is none, which is another reason the western world is fucked and it's time to get out.
Curt 2016-11-21 10:24:48
Your mistake here is not looking at margin of error. You are right, polls are always wrong. However, they are useful for approximating the final result. You state multiple times that polls are "wildly" wrong without defining "wildly". By your own graph, national polls showed Clinton +2.3 on Nov 4th. The final result was about Clinton +1.0. That's an error of 1.3, an extremely good result historically. The state polls did quite a bit worse, but individual state polls tend to be slightly less accurate historically for various reasons. Yes, polls missed the final result because the election ended up being extremely close. However, during years that polling shows more of a clear lead, they can still be useful.
Caleb Jones 2016-11-21 10:58:04
You are right, polls are always wrongThat's not my point. My point was, for most of my lifetime, the polls were almost always right. Things have changed now.
Your mistake here is not looking at margin of error.I realize and acknowledge there is always a margin for error and that the election was very close. I figured it would be. The problem is, again in my lifetime, I've seen *many* elections this close or closer, and they had margins of error too, yet the polls were still usually accurate. Again, look at the graph I posted above that shows the nationwide polling five months prior to the election. Trump was down, often way down, pretty much the entire time. Yet he won. Something has changed. Either societally, in the polling systems that are used, or in the media. Or all three. Perhaps the only thing that has changed is that the margin for error is now wider than it has been historically. My point is *something* has changed. I agree that if you see a poll that's obvious, like 80/20, then it might be useful. The problem is we haven't seen major elections in the US like that in many decades. They're almost always close, and thus, useless. Europe too. (Look at the Brexit vote.)
Way_of_Man 2016-11-21 23:42:20
What's interesting is the market's reaction. The stock market is already insanely over inflated. A win for Hillary was predicted to be good or at least "stabilizing" for the markets because she's part of the old guard and would do her best to keep the status quo on Wall Street. Trump is unpredictable. And the market typically dislikes unknown factors. Or so says the old logic. And initially, the futures in stock indices seemed to reflect this idea in that they were taking a dive the more and more good news there was for Trump. Then the unthinkable happened. Trump was announced as the winner and the markets rallied to new highs. This is literally insanity. The markets have no logical reason to be this optimistic. The Trump Train should have caused a sharp drop and in the end, it didn't. The last time I saw this much irrational exuberance over stock prices was 07. Look out below. I'm just hoping I can slush together enough extra capital to short the shit out of the market when the time comes.
Caleb Jones 2016-11-22 11:44:19
The stock market is already insanely over inflated.I check several key economic figures every morning, including the Dow, and man, that thing just keeps rising and rising with no end in sight. It may break 19,000 soon... all for a tiny fraction of its real value. Absolutely insane. When the market crashes next, I'm going to be buying up as much as I can. It's going to be a bloodbath.