Pedro Pinto's Reviews > The Black Swan: The Impact of the Highly Improbable
After 10 years of my first reading I did something that I usually do not do and re-read this book.
I did it because I'm currently reading the Incerto collection from NNT and this is the second book. I'm really glad I did as I think I liked it more in 2024 than I did in 2014, as most of the book I had already forgotten with the exception of some key ideas, thus this book was almost in the most valuable part of your library, the one that was not yet read (your library should be divided in 3 parts instead of 2 by increasing order of importance: 1.- the books you have read; 2.- the books you have read but almost forgot about it and 3.- the unread books).
NNT with a compelling, witty writing style takes us into this marvelous world, that is the one we live in (although we tend not to see it).
The key ideas that I will keep from the book are the following:
1. The key traits of a Black Swan: a) it is an outlier, b) it produces extreme impacts (positive or negative); c) with the benefit of hindsight everyone can explain it and see it coming (specially the pundits).
2. The dangers of the bell curve (normal distribution) specially in Quadrant IV of events and all that it entails.
3. The danger of opacity on the human mind, driven by: a) illusion of understanding, b) retrospective distortion and, c) overvaluation of factual information and handicap of authoritative and scholar individuals.
4. Clearly identify and treat different events that are scalable (where power-laws prevail, Quadrant III and IV - Extremistan ) and the ones that are non-scalable (Quadrant I and II - Mediocristan) where the platonic world and normal distribution can be used.
5.- There is a big difference of Absence of Evidence with Evidence of Absence.
6.- Probability should be considered as a liberal art (not a quantitative one).
7.- Additional data/information (d2d) usually is useless and sometimes toxic.
8.- Simpler statistical methods most of the times are better than the highly sophisticated ones.
9.- Fractal world - repetition of geometry at different scales, but within those scales, numerical or statistical measures are proportional.
10.- Avoid Locke's madman syndrome - to reason correctly from erroneous data
11.- Watch-out with the saying that all models are wrong, but some are useful as some are also very harmful.
12.- Avoid optimization, learn to love redundancy (inspire yourself in nature)
13.- If you try to understand complexity use simplicity.
Finally, by re-reading this book, my curiosity on the books, amongst others, of Poincaré and Mandelbrot has increase significantly and have included several in my wish list! The books I have added were:
1. Why most things fail - Paul Ormerod
2. Critical mass - Philip Ball
3. Ubiquity - Buchanan
4. The fractal geometry of nature - Mandelbrot
5. What I learned losing a million dollars
6. The (mis)behavior of markets - Mandelbrot
Hopefully you will enjoy it as much as i did!