This book is a quick read that sites some very interesting studies/stories. I appreciate that this author has a different way of looking at problems bThis book is a quick read that sites some very interesting studies/stories. I appreciate that this author has a different way of looking at problems but still reminds us that correlation does equal causation (many people forget that from Stat 101). ...more
**spoiler alert** I liked some of the characters, even if they were a little unbelievable. But, I felt like about 3/4 in the book the story changed. A**spoiler alert** I liked some of the characters, even if they were a little unbelievable. But, I felt like about 3/4 in the book the story changed. And toward the end, the format changes and in the last 3 pages the narrator changes from one character to another within the same section of he book. Very inconsistent and disappointing. I wanted to know if she wrote the book, how it came out, did she adopt the girl. Instead the book abruptly ends with the main character asking another to marry her. THE END.........more
It was really hard to keep reading this book. It is very dry and negative. The entire book should have been a brief essay. The message of this book: pIt was really hard to keep reading this book. It is very dry and negative. The entire book should have been a brief essay. The message of this book: people suck at predictions. In the book, he gives an example of a statement from Bertrand Russell that he disagrees with and says that he disagrees with it because he does not believe it is "helping us deal with the problem". I feel like this entire book is setting up a problem that the author does not help us to deal with.
He defines a black swan as an event we cannot predict, but then he tells us to try to predict black swans, just not precise black swans. He wants us to stop predicting, but tell us to “Put yourself in situations where favorable consequences are much larger than unfavorable ones.” Is that not predicting??? It is just making a prediction without using the any statistical tools to help us with probabilities. The author also says that he never makes predations, but makes the statement: "We saw that companies can go bust, while governments remain." Not only is that a prediction, it is inaccurate. Lately, we have seen many governments overthrown or going bankrupt.
He suggests that anyone that makes predictions for a living that should "get another job". He also seems to think that asking a random person on the street how many dentists there are in Manhattan while result in just as good of a forecast as those made by professional economists and statisticians. He also seems to think that reading less newspapers and blogs, watching less TV and narrating less may help us to better be prepared for black swans.
There are a lot of problems with some of his arguments:
He thinks companies should not make 5 year plans because they can not predict everything that will happen in the next five years. Companies should make plans, but be open to adjusting as needed. He suggests that almost all inventions are a "product of serendipity". Many inventions are the result of years of work (like the airplane - it was not stumbled upon)."
Have to talk about his bad statistics:
The law of iterated expectations is not from statistics ( I am a statistician and never heard of it). I looked it up; it is an economic term.
I took until chapter 15 to realize that he is using the term Gaussian to refer to all statistics, not just to the Gaussian bell curve (Normal Distribution). Through out the book he is bashing the bell curve because it does not fit every type of data he can think of. And it is not supposed to! He seems to think it is the only distribution used by statisticians and he couldn’t not be more wrong…we have many distributions: uniform, Bernoulli, lognormal, t, gamma, beta.
His example from page 245-250 is not Gaussian…it is binomial.
He is correct that the normal distribution fails if the observations are not independent, but he must have been asleep in class when they taught that you should always CHECK the assumptions when using any statistical model.
His uses an example of graphs on pages 186-187 to show how making predictions using mathematics and statistics is wrong. But what he fails to point out is that in statistics, it is true the prediction would be the line you see in the graph, but it would also include the confidence intervals around the prediction that get wider and wider as you go further out. ...more