No mo' Bad Luck? Book on drop in Std Dev lowering cost of uncertainty
February 19, 2023 11:26 AM Subscribe
Help me remember a book from 10 or 20 years ago that discussed how a general decrease in modern processes' and activities' standard deviation has altered our outlook through lowering costs of uncertainty: in cases where people used to blame bad luck, now we tend not to because of much greater expected precision about time and consistency of results. So, in shipping, you can't say 'Oh the ship sank or it's really late but it's nature, it's not my fault.' Or, apparently, duelling disappeared because of the decrease in randomness.
Response by poster: It turned out that the book was Peter L Bernstein's "Against the Gods" (1996)--my friend remembered it somehow.
posted by diodotos at 2:12 PM on April 8, 2023 [1 favorite]
posted by diodotos at 2:12 PM on April 8, 2023 [1 favorite]
So, I ended up reading the book and unfortunately, your description doesn't quite match the details. Yes, the central thesis is about quantifying variance, but it doesn't describe any decline in variation over centuries (that would be quite the feat of historical data mining!). Rather, the book describes the evolution of insurance, why you want it and how to correctly charge for it. And I definitely don't recall reading anything about dueling.
Where this particular book ends up going is options, and specifically quantifying risk using Black-Sholes, and was written two years before LTCM blew up and took a billion dollar bailout. Or the MBS blowup. So that's an interesting perspective.
Anyways the key thesis is that variance -- up or down -- is the trick to pricing options. It's not suggesting that variance dropped over time, but one can infer that if it did, insurance (and options specifically) would get cheaper. The book is rather quiet on how to reduce variance of outcomes, presumably because the author is in the business of trading stocks for a living; his Wikipedia bio suggests he managed a billion dollar trading firm in the 50's.
posted by pwnguin at 9:37 AM on September 26, 2023
Where this particular book ends up going is options, and specifically quantifying risk using Black-Sholes, and was written two years before LTCM blew up and took a billion dollar bailout. Or the MBS blowup. So that's an interesting perspective.
Anyways the key thesis is that variance -- up or down -- is the trick to pricing options. It's not suggesting that variance dropped over time, but one can infer that if it did, insurance (and options specifically) would get cheaper. The book is rather quiet on how to reduce variance of outcomes, presumably because the author is in the business of trading stocks for a living; his Wikipedia bio suggests he managed a billion dollar trading firm in the 50's.
posted by pwnguin at 9:37 AM on September 26, 2023
This thread is closed to new comments.
Or the one freakonomics book, since I'm pretty sure Levitt has told a story about calibrating when to leave for an airport using the metric of "number of missed flights." Googling finds How Not to Be Wrong covers this in book form in a chapter.
posted by pwnguin at 6:00 PM on March 5, 2023