Disclaimer: I am not a Trump supporter

I don't like either of the candidates and because I think the odds of a war between the US and Russia go up dramatically with a Clinton presidency I refuse to sacrifice even an illusion of complicity by engaging in a vote for a neoliberal war-monger even if Noam Chomsky tells me she's a lesser of two evils. Chomsky, despite his apparent beliefs to the contrary can not predict the future and has presented zero evidence that he is remotely proficient at forecasting and so his belief that "any democrat" would be safer than Trump clearly falls well outside the range of his expertise.

Taking Chomsky's own assumptions at face value, if I believe everything that Trump says to be true there is a chance (because the decision is not solely his) that he will restrict Muslim immigration in an attempt to fight terrorism, and a chance he will restrict Mexican illegal immigration. If I believe everything Clinton says then I can expect tensions to rise between us and the country possessing the most nuclear warheads in the world (aside from us) which has the chance of instigating a nuclear war. Given Chomsky's past stated fear of the potential for nuclear war I find it perfectly reasonable to give this primacy in my analysis over non-violent immigration policies. That being said, the potential for those policies to turn violent, among many other things, keeps me from voting for Trump (not that my vote matters).

Intelligence is no defense against bias

Despite in the past holding up potential nuclear warfare as one of the biggest existential threats to mankind for this election Chomsky has now replaced the potentially imminent annihilation of mankind due to nuclear conflict with the less imminent annihilation of mankind from global warming as the "key issue" saying that since all Republicans dismiss global warming (though Mitt Romney didn't, I suppose Chomsky would have voted for him?) any Democrat would be better (and that's not overly simplistic?). While Chomsky and his supporters like to hold him up as some neutral and unbiased arbiter of truth, the evidence indicates that he is only human and subject to the same confirmation biases as the rest of us. Let's remember that Chomsky lives in the North, in Massachusetts (a Blue State). That he is interpreting the data to support the same candidate most of the people around him are supporting should not come as a surprise. Being the Einstein of linguistics does not make you immune to cognitive bias.

But What About The Polls?

In Nate Silver's "The Signal and the Noise" he tells the story of professional gambler named Haralabos "Bob" Voulgaris, a sports bettor who, according to Silver, makes over a million dollars in a bad year. Armed only with a degree in philosophy and the knowledge of his likely financial prospects given said degree, Voulgaris decided that in order to make enough money to free himself from his current occupation as an airport skycap he had just enough money to make one big bet. When deciding what bet to make did he look at what the mainstream had to say? In fact, just the opposite.

According to Silver, "Voulgaris was never a big believer in conventional wisdom—it’s in large part its shortcomings that make his lifestyle possible..." (Silver, Nate. The Signal and the Noise: Why So Many Predictions Fail-but Some Don't (p. 234). Penguin Group US. Kindle Edition.)

The bet he settled on was the L.A. Lakers to win the 1999 NBA Championships. Though the team seemed objectively good, an apparent feud between its two best players, Kobe Bryant and Shaquel O'Neal, convinced most people (including the book-makers) that their odds of winning should be reduced, especially given incidents like O'Neal getting ejected from a game after losing his temper. If winning depended both on the two best players on the team working together, as well as remaining in the game, how could any sane person think such a dysfunctional team stood a chance?

According to Silver, as Voulgaris saw it the Lakers were still playing well on paper and he liked what the coach was doing with the team, so he placed a bet of $80,000 (practically all of his money) on the Lakers winning. If that happened he would make a million dollars.

At first his prediction looked good, the Lakers performed well for the remainder of the Season. This would be like Trump performing well during the primaries. But in the playoffs they faced numerous near defeats and odds defying comebacks. In the deciding game 7 of the Western Conference semi-finals the Lakers were down 16 points in the last two minutes of the third quarter puts the odds of them winning THAT GAME 15-to-1 AGAINST. Imagine what their odds of winning the entire championship were at this point? But their opponents got unlucky, missing an abnormal number of shots, while the Lakers were able to increase their output. They won the game and eventually went on to win the championship.

Every winner almost lost

My point is only that when you look back at the history of any competition, you will find no shortage of events that had the potential to alter the outcome and seemingly signaled the imminent demise of one of the teams. Does that make Voulgaris' prediction any less right? Between the prediction and the event countless events will occur some of which may be related to the outcome, some of which may be unrelated, some of which may appear unrelated but be related, and some of which may appear related and be totally unrelated. In the case of almost all, if not all, of these events we will never know whether they had any impact on the outcome whatsoever. We tell ourselves in the moment that we see causation, we pretend that means something, and then we move on with our lives.

Real turning points don't exist in elections, just imagined ones

In sports competitions those events are relatively concrete. If the Lakers lost the Western Conference semi-finals, their odds of winning the championship would have been ZERO. So at least in those cases we can say, "Because they lost that game or series of games, that caused them to lose the championship because the loss eliminated their ability to compete for the title." This election, I believe, has far fewer defining moments such as that. Even if Trump had lost the Republican primaries he very well might have had a chance running as a 3rd party. It is difficult, if not impossible to estimate how much good running as a Republican has done him and anyone who suggests otherwise likely also would have failed to predict his success as a Republican candidate as well.

There is an expression in trading circles: "this time it's different." It's usually used to refer mockingly to people who, when faced with an obvious financial bubble dismiss the concerns of "bears" by claiming "this time it's different." But the reason we fall victim to this logic is because things are usually different. The most you can say when things are different is that our normal heuristics likely won't hold up. This is why "this time it's different" is such a ludicrous justification for ignoring worrying economic trends, because it should make a person MORE concerned, not less. The fact that Trump has defied the odds so consistently proves that we don't understand the dynamics of this election. Period.

Of course Trump's odds of winning are low. But Trump's odds of winning have always been low. And yet despite this, and despite spending far less than any other campaign (hence why discussions of campaign finance reform have been mysteriously absent from the Democratic platform ever since Barack Obama out-fundraised Romney by 9% in order secure a lead of 4%) Trump has continued to win.

People can come up with all kinds of reasons for why his poll numbers move the way they do. They can pretend that they know that the recent decline in the polls is "caused" by the revelation of his lewd private comments. But is this a valid assumption? Are people very good at perceiving causation? Of course not. We're notoriously BAD at it. Has there ever been an election, for example, where revelations were NOT made close to the election which the opposition attempted to frame as completely devastating to their opponent? In fact it is a universal trait of elections.

Whatever is "revealed" will be treated by the opposition as if it is the single most Earth-shattering revelation ever revealed in the history of mankind. The fact that his numbers declined at the same time is great promotional material for his opponents, but of course if his numbers had not declined this merely would have been used to "prove" that his supporters are sexist and misogynistic and we live in a patriarchy, though of course his decline in the polls should not be interpreted as proof that we do not live in an inherently misogynistic and patriarchal society because of the first commandment of politics: thou shalt interpret all data so as to promote one's narrative. Whatever event occurs will always be spun and the movement of the polls will always be used to support this narrative. Think of it this way: what if we knew for a fact that the short-term movement of polls was random, do you think that would change anything?

Just look at the wikileaks revelations regarding the Clinton campaign. If instead of Trump's numbers declining, Clinton's had declined what would everyone be saying? "These leaks are devastating to the Clinton campaign, the public is so outraged they're willing to support a sexist, racist, misogynist like Trump." Like the parent of an autistic child who just received a vaccine we look at the confluence of two seemingly related events and assume a causal connection because we are incapable of discerning cultural reality from "real reality." Our culture has formed a connection between vaccines and autism that our biases latch on to and amplify in our minds. So too has our culture formed a connection between movements in polls which are just as likely to be noise as signal which our biases amplify into meaningful data.

Where does that leave us?

Well there are a few critical components that I think give us reason to adjust the odds more in Trump's favor. As I said in a previous post, this time really is different in that it is the first time that I know of where the Republican establishment hates their own candidate as much as their opponents do. We all know that political polls are biased by the people conducting the polls with some polls leaning left and some leaning right, however, if there are more people on the right who are leaning toward Clinton, then in this case we should expect more polls to lean Clinton's way than under normal circumstances. From this alone we can say that support for Trump is likely higher than the polls currently indicate, though how much is unknown. Since some polls have Trump slightly ahead or tied with Clinton, we might reason that there would likely be more such polls if he had more sympathizers in the mainstream.

This is also why Nate Silver's methodology at 538 can no longer be trusted. 538 adjusts the weight of polls based on their historical performance. But it also assumes loyalty to party affiliation. It assumes that a "right leaning" poll will continue to be right leaning and that Hillary Clinton is not right leaning. In fact, Hillary Clinton is in many ways a more conservative politician (her support of Free Trade and Wall Street, her past opposition to gay marriage, etc.), is certainly more "establishment," and is at the very least the stated favorite of many popular conservatives.

Brexit

The next critical data point is the one being highlighted by the Trump campaign: the Brexit vote. Because it is being promoted by the Trump campaign it is obviously going to be dismissed out-of-hand by the opposition and its media surrogates (the majority of the mainstream media) and ironically, this means that its effects will likely be out-sized compared to their predictions as they are likely to ignore or dismiss it because it does not conform to their desired narrative. However, here we do have an incident where the mainstream media and polling apparatus firmly predicted an anti-populist result ("remain") and the result was a populist Brexit vote by a margin nearly identical to Obama's win over Romney in 2012. We should also not forget that Obama was a populist, anti-establishment, pro-change candidate, and ask ourselves whether the forces which elevated him to victory have diminished or increased since that election?

In the case of Brexit we have a situation where the mainstream politicians and press believed that the polls they were conducting were accurate and simply happened to support their narrative and then were completely surprised when reality came crashing down on their heads. And now we are faced with a situation where a historically unpopular candidate, who couldn't sell 3,000 copies of her book (according to the New York Times), who can't get people to come to her rallies and who has to pay for people to disrupt her opponent's rallies, is being presented to us as the inevitable winner because right now the polls appear to be in her favor.

A similar incident occurred in Greece when the Syriza (the socialist party) government held a referendum on whether to leave the EU, the results of which were ignored when the people of Greece overwhelmingly voted to leave as well. I guess the Greek government was expecting a different result.

Breaking News: Polls Change

But does it make sense to assume that the polling numbers will remain where they are? And if not, how is it likely they will move? From April to May Trump's numbers surged 4%, only to correct -5% for the first half of June, before once again turning around and surging again for the next month and a half from a low of 38.3% to 45.7% (7.4%). In just 10 days that number once again dropped 5.8% before turning around YET AGAIN and resuming a steady two month streak of upward momentum that resulted in another 5.1% increase at the same time as Clinton's numbers were demonstrating a steady downward trend that wound up being ~-4%.

One has to wonder what those who take polling numbers as the undisputed word-of-God must have been thinking as Clinton's numbers marched inexorably downward, at every moment confirming what everyone (even her defenders) know about her? That she is not popular. That she is not charismatic. That she is not nice. That she is no one's favorite candidate and at best a lesser of two evil's. What a relief it must have been when the polls turned around and their cognitive dissonance could be dismissed as a momentary lapse of pessimism. Finally, the polls can be trusted once again, now that they support the narrative.

However, if we simply look at the trend what we see in Trump's numbers are prolonged bouts of growth interrupted by abrupt but temporary corrections. In the first half of October Trump saw a decline 3%. If this time really is NOT different, then we can expect them to rally another 4-7% just as they have in the past. Of course, in order for him to win Clinton's numbers would have to decline in a commensurate fashion as they did in December, April, June, and August.

Expect The Unexpected

Of course, this time is different and anyone who thinks they know what is about to happen is deluding themselves. That being said I wouldn't be surprised if going into the election Trump's numbers rise and Clinton's numbers fall. If on the day of the election Clinton's lead has narrowed to the margin of error, a Trump surprise victory would not surprise me at all. I don't say this because I wish it to be true, in fact, the very fact that I do not like either candidate should mean that I am less likely to be biased in either's favor doesn't it?

Does Any of This Really Matter?

For reasons that will have to be addressed in a separate post, I believe that the problems our society will be faced with in the coming decade are so significant, and so utterly ignored by the mainstream (despite being felt by the majority), that whomever is elected will wind up more like a kite in a hurricane than the more usual position of, as Elon Musk says, "Captain of a very large ship with a very small rudder." For example, the USD is likely stuck in a deflationary spiral which will continue to wreak such havoc on the global economy that a major reform not unlike Bretton Woods will be required in the next 5-10 years.

At the same time we have a Higher Education Bubble and Pension Crisis (which are not unrelated to the deflationary forces) which are also likely to come to a head. It is far more likely that whomever is elected will merely provide the same faulty-pattern-recognizers I am talking about with a scapegoat for these events as opposed to someone they can see as a motivating force for positive change, though in both cases they would still be wrong.

The Crowd moves and we along with it. We focus our attention on a few members to cope with the insanity that we exist in a society built on an infinite number of unconscious and unseen forces.