"The Drunkard's Walk"

(How Randomness Rules Our Lives)
-by leonard mlodinow
2008
My aunts and mother believe that every time they find a penny, it's a message from their dead mother.
My brother insists that whenever we play cards, the trump card must always be "predicted". He gets mystically orgasmic when someone is right (or even vaguely close).
Am i the ONLY person in my family not a superstitious simpleton, ignorant of any understanding of statistics and probabilities?
Mind you, i have my own statistic bugaboo - the fact that i don't buy it. The whole concept, i mean. For the most part i do, but there's one element i can't reconcile, the notion that if you deal with large enough numbers, say dice rolls in the thousands, the distribution will become more and more predictable the higher you go...you'll eventually get the same percentage of sixes as ones...and I DON'T QUITE BUY IT. It seems just another form of mystical-wistical mumbo jumbo...for all that to work out, it seems that dice would have to possess memory. But DICE CAN'T REMEMBER, i say! Which mlodinow would agree with, to a point. He would say that you shouldn't be amazed by six sixes in a row (or even sixteen or twenty-six), but he would also say that when you get to high enough numbers, the probabilities even out.
It feels like statisticians want to have their cake, and eat it too.
But one thing i do accept, as this book shows, is that human probabilistic intuition is grievously flawed, and the sooner we understand that, the better our lives will get. And this book offers a wonderful, accessible understanding of how these things work. Mlodinow (who co-wrote "A Briefer History of Time" with stephen hawking) talks about regression toward the mean, which shows that great or awful results are aberrations, and subsequent events always trend back toward the average. This trips us up, because humans are easily swayed by extreme events (for example, flight instructors fallaciously believe that yelling at a student after a bad flight is a successful teaching tactic). And perhaps the greatest human statistical misconception is the belief that success and failure are based on merit. Statistics show that both results are largely random. For example, john grisham had his first book rejected by publishers twenty-six times. Dr. seuss, twenty-seven. Or take pulitzer-prize winning novelist john kennedy toole, whose first book was published eleven years after he committed suicide after repeated rejections by publishers. Success and failure (at least by the myopic, zero-sum standards of this society) are at least as much about persistence as ability.
I can imagine a more sharply-focused book, strictly dealing with exposing human misconceptions. The title refers to the randomness (like bouncing molecules) within order. Leonard perhaps neglects the topic of coincidence, and how "miraculous" events are actually commonplace...something we fail to appreciate, because we're not always looking where we would have to, to see them all. So we get boggle-eyed by seemingly-astounding coincidence. Mlodinow does, however, take the time to delve deeply into the history of probabilities. And he offers many rich examples...like how we know that a significant percentage of college sports games are fixed, or how we can be certain that wine-tasting guides are bullshit. Or how financial advisers and movie studio executives are rewarded or punished on the basis of nearly 100% bullshit.
As for my brother, if ESP existed, every casino in the world would go out of business.
As for my mother and aunts, if they started finding spanish doubloons everywhere...there STILL would be a rational explanation.
As for me, i may have to go roll a die one thousand times, just to accept the magical powers of statistical probabilities once and for all.
A wonderful book.