Status Updates From Everything Is Predictable: ...

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."

I'm sure even an "outsider" can see that's no small difference: "He wasn’t an Anglican. Nor was he a Catholic. The two doctrines are different, but not all that different—they differ on what seem to the outsider relatively small points. The Catholics believe salvation comes only through the Church, whereas the Anglicans believe that having faith in Jesus Christ and following his teachings get you to Heaven,..."

Gauss was key to the Bayesian framework we know today; ND also overlaps with Bayes approach but again they are different equations that are intertwined but fundamentally and mechanistically differ

OK I said what I just said but cmon you can't just say "It is, without exaggeration, perhaps the most important single equation in history." and without any reference & when there is exists Pythagoras' theory (architecture), logarithms!! (even just it's role in allowing complex computation, CALCULUS, law of gravity (anything space or flight), wave equation (from acoustics to optics), even Normal/Gaussian distribution

Ok, now that I figured out what was irking me I can level-set and proceed onward.

That’s extending a theory of probability further than, well, logical. Bayes sticks close to what people think “probability” and “uncertainty” is but it is not in-and-of-itself logic. We make inferences with it yes, just like any probability theories.

Also, Chivers throws the term “priors” (a very specific Bayes term) to mean whatever thing he wants to apply to whatever example he is using.

That said, what DOES bother me about this book is that it seems to conflate Bayes framework with logical argumentation as a whole. “Logic itself… is just a special case of Bayesian reasoning…”.

me to read this book scrutinizingly as I do textbooks (by “scrutinizing” here I mean deeply reading for logically sound argumentation as a process for me to understand the logic of the underlying concept(s)). This book is a different media for a different audience written by someone with whom this isn’t their academic field of study. Which is fine. I just shouldn’t base on same metrics.

I figured out what was irking me. First, my biases: I started reading this as a student in epi & biostats + someone new to Bayes concepts but looking for a non-textbook reading to supplement (as leisure reading) my current learning activities. Ie) I’m reading as a new student to this topic who is currently active learning this concept through academic curricula. This is not academic curricula and it is unfair of

I just learnt a new word to differentiate Chivers’ Bayesian fervour with more measured (and sane) use of Bayesian theory — Bayesianism, which pedestals Bayesian theory as a “narrative description of rational belief” (McElreath) rather than a probabilistic theory that is more/less useful depending on the situation.

(After brief readings of alt bayes refs) is this just the bayesian way?? This overconfidence that Bayes is best.. shoot. What have I gotten into...

Not that the Bayesian logic is inaccurate (I am a baby in this Baysian world, neigh an embryo) but it’s that his claims are so very shiny with a spin to it that’s on the side of overpromise.

“It explains why scientific results can be “statistically significant” and yet still be very wrong”. Even I know you’re purposefully saying so to overexcite (at best) or mixing statistical paradigms (at worst).

I was really excited about this book since it was the first Bayes-related book I ever bought and had hoped it would ease me into to “new” (for me) paradigm. But not even a quarter into the Intro there are some claim that give me pause.

This book would not pass peer-review which, fine, that’s not Chivers purpose & he is a writer not statistician. But now im going with caution.