jj’s Reviews > Everything Is Predictable: How Bayesian Statistics Explain Our World > Status Update

Separate but tangentially-related thoughts:
- this is now dubber the "Summer I Discovered Bayesianism"
and
- I am enjoying myself immensely

Oh dear, Chivers is quite convincing in giving evidence of his main theses (ie I am becoming convinced): "It meant that Turing and his team had to use priors—to assume that some combinations of letters were more likely than others. So the three-letter sequence E-I-N, ein, German for “a” or “one,” would be more likely—he reasoned—than the sequence J-X-Q. Words like “WIND” or “CONVOY” would be more likely than “ITCH”..

When I first heard about the "stats wars" I laughed it off as silly. Yo there was blood drawn: "Fisher, who called Bayes’ theorem a “staggering falsity,”93 “perhaps the only mistake to which the mathematical world has so deeply committed itself,”94 which “must be wholly rejected,”95 had won."

What have I gotten myself into...
"motivations—he admits there’s a war between frequentists and Bayesians, and says in his book, “Consider this [book] a piece of wartime propaganda, ... My goal with this book is not to broker a peace treaty; my goal is to win the war.” It would certainly help win the war if it turned out that frequentists were racist."

What a name drop: "John Maynard Keynes, the great economist and liberal (and my great-great-uncle, so I ought to declare an interest here when I try to downplay the awfulness of it all), was another."

Lol: "Quetelet made various mistakes—he hadn’t realized that there are other ways for a quantity to be normally distributed, for instance, and also went around gleefully applying the normal curve to everything he saw. A later statistician would diagnose the condition of “Quetelismus,” of seeing the normal distribution everywhere you look."

Realized why this book is 'reading weird to me'. "Remember: Bayesian inference is distinguished by a broad view of probability, not by the use of Bayes' theorem." (McElreath)

Chivers is making his claims based on Bayes theorem (fair, he gets to decide objective of his book). But Bayes theorem is, at simplest and broadly, a mathematical notation of probability theory; Bayes was just the first to write it down as is.

Chivers is drinking the Bayesian kool-aid (the title is a big clue) but not a bad entry-level book on probability theory and Bayes. Amusing math history and anecdotes but missing papa Gauss??

I will give credit though that this math history chapter is doing well to contextualize my earlier textbook reading about math being (paraphrasing) explanations of phenomena that can (and still sometimes do, obvs not the foundational concepts that have been argued and tested and used through centuries) themselves be debated and discussed

You don't "reduce" the chance, they're independent events: "In the case of the four rolls of a single die, your chance of not seeing a six on any one throw is 5/6, or p ≈ 0.83. If you roll it again, your chance of not seeing a six on either throw is 0.83 times 0.83, or just shy of 0.7. Each time you roll the die, you reduce the chance of not seeing a six by 17 percent."