The Flaw of Averages: Why We Underestimate Risk in the Face of Uncertainty, Sam L Savage
A businessmen's book. That is, carefully written for people who aren't the sharpest knives in the drawer. Nevertheless, interesting.
The biggest drawback of the book is its commercial nature, both in the writer and in the intended readers. Savage is trying not to tax his targeted readership's limited patience, and at the same time, he's got a consulting business to support. He pimps the book's website, flawofaverages.com, a total of 42 times over the course of the book. That is not an exaggeration. I did a search on that phrase in the ebook. 42 times. He refers readers to the website nearly once a chapter. (Short book. Short chapters, all 47 of them. Did I mention the limited attention spans?) This is annoying.
But hey, it's a breezy read, and not boring. It's also got helpful reminders of mental traps to avoid, not all of which I knew. (I am a collector of them. Nice to catch a few more.) Also, it's got a great visualization of the Central Limit Theorem. It's so easy to do that I typed this myself.
First, this is the uniform distribution of throwing a single die. Each x represents a 1/6th chance.
x x x x x x
1 2 3 4 5 6
Next, this is the distribution of throwing two dice. Each x represents a 1/36th chance.
x
xxx
xxxxx
xxxxxxx
xxxxxxxxx
xxxxxxxxxxx
23456789X12
Last, throwing three bones. Each x represents a 1/108th chance, rounded up. (At full resolution, with each x representing a 1/216th chance, it would've been twice as tall.)
xx
xxxx
xxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxxxx
xxxxxxxx
xxxxxxxx
xxxxxxxxxx
xxxxxxxxxx
xxxxxxxxxxxx
xxxxxxxxxxxxxx
xxxxxxxxxxxxxxxx
3456789X12345678
Amazing how quick the bell forms, including clearly visible inflection points. Three dice. Uniform distribution to a proto-bell in three steps. Even if you know the proof, it can be amazing to see.