Psychohistory & Mob Dynamics

Why most election polls couldn’t predict a sunrise

The term “Psychohistory” represents a fictitious science in which future historical events can be predicted by applying statistics to the behaviors of large groups of people.

Isaac Asimov coined the term in his seminal work “The Foundation” and suggested that predicting the future was similar to predicting the flow of a gas: The movement of an individual gas molecule was impossible to know, however the collective action of the entire cloud is predictable to a high degree of certainty.

Every tool we use to predict major events (like elections) relies on this same basic premise of Psychohistory. Expand the scope to a large enough scale– ask tons of people who they are voting for–average out the results with as many other polls as possible, and Voila! you should have a highly accurate statistical prediction of the future.

And yet, statistics failed us so very badly this election cycle. As my friend Tom said on election night, “Nate Silver can’t predict a sunrise!”

We all feel duped. So the question is what went wrong? Math is unbiased and impartial and not at fault here. The problem is clearly with the inputs. So how do we find the bad inputs?

Well, first we should look to see who got it right. I don’t mean we should study the crackpots who applied their proprietary election prediction methodologies (mostly comprised of gut instincts). This is as beneficial as studying Felix the Polar Bear, who foresaw the Trump victory by rubbing his snout on a Trump Jack-O’-Lantern.

No, we should look at the polls that got it right, and there was a major respectable poll that got it very right: the USC/L.A. Times Poll.

One particular difference between this poll and rest is what interests me. Instead of reducing each person to a very binary decision, e.g. “Who are you voting for?”, the USC/L.A. Times poll asked people how committed they are to their candidate on a scale of 0–100.

The theory was, the poll could better identify if a person was likely/unlikely to vote, because someone with only marginal conviction can be more easily revealed on an expanded scale. Their apathy becomes clear and measurable on a 100-point scale, in other words, and that translates to not showing up.

Human beings, it turns out, are not binary after all, and Psychohistorical-type predictions can’t be made by merely increasing the sampling total. Accurate inputs are required to begin with, and this means individuals need to be sampled in a non-binary way to expose the true flavor and fervor of their subconscious minds.

Another way of looking at this phenomenon is this: Consciousness is mob dynamics. Our decisions are merely the sum of our internal mob — the electrochemical war between all of our beliefs and experiences — fighting for polarization and the chance to steer the organism.

Unfortunately, most people this election cycle had very apathetic internal mobs and they never gathered enough momentum to show up.

So where do we go from here?

This isn’t a post about how to change the future. It’s a post about how to better predict the future. To make our Psychohistorical predictions more accurate, we clearly need to stop sampling people in such restrictive and binary ways.

The flawed polling methodologies put all of their weight behind what people said. The way in which they said it was not factored in. The polls made the mistake of treating the answer, “I’m voting for Hillary”, as a mark for Hillary. The frown on the person’s face when answering that question would be the more pertinent information. After all, a frown represents any number of a billion chores which could get in the way of voting on November 8th.

Going forward, we need to uncover the dynamic of a person’s internal mob with more cleverly devised approaches, much like the USC/L.A. Times poll had the foresight to accomplish. I suspect that facial tracking could play a role in future polling methodologies, but there are pitfalls here that need to be explored still.

As a side note, we all like to think we are in full control of our subconscious mind and the forces which influence it. This is a pleasant fiction indeed, until you realize you have a sudden craving for ribs while watching House of Cards, or pass by the Merlot section of wines in the grocery store and think, “nah, Merlot sucks.”

The reality is our subconscious is a highly malleable mob, and it’s often influenced by the neuron/moron who is yelling the loudest. Unfortunately for America (and the world) this turned out to be too true.