Read on Twitter
As I am watching the election coverage, I can't help but think about it in a broader context -- that of Asimov's Psychohistory, to be precise. We are talking about a collective outcome integrating hundreds of millions of individual human wills.
Surely modern-day Hari Seldons (Nate Silver?) should be able to predict the outcome to the sixth decimal point. Yet it's not happening. Why?
To explain what I mean about a collective outcome from 100,000,000s of individual wills, I can predict the probability of being killed in motor vehicle accident next year in USA to the fifth decimal point. It will be 0.00011.
https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in_U.S._by_year
https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in_U.S._by_year
But I wouldn't dare predict the way this election goes even now, 11:20 pm on Nov. 3. Something is different.
Thinking more on why US presidential elections are so hard to predict. Yes, voting decisions are not independent, and such feedback influences easily generate mathematical chaos. Yes, people lie to pollsters. Yes, exogenous events can have a big effect.
But it's not enough.
But it's not enough.
It seems to me that the major reason for unpredictability of US presidential elections is the Electoral College.
It's almost like the Founders decided to build a "Chaos Generator" (and succeeded very well).
It's almost like the Founders decided to build a "Chaos Generator" (and succeeded very well).
