|
Along the way, we meet a Renaissance doctor who made more money at games of chance than treating the sick, a Swiss dynasty and full-scale soap opera of mathematicians, and a mathematician who experienced a religious conversion and made a probabilistic argument for the existence of God. The book begins with a discussion about how much we may underestimate and underrecognize the role of randomness in our lives, and at the end returns to the theme of how much we misattribute success or failure to our own efforts while neglecting the role of chance.
And he discusses some surprising psychological experiments that expose our innate tendency to find patterns in random "streaks" and to attribute intentional control over things we don't actually control (even when we know better). He touches on a wide variety of applications, from baseball (home run records and world series outcomes) to movie industry executive performance and mutual fund management success, illuminating how much of such outcomes are random.
CalTech physicist Leonard Mlodinow, in his book "The Drunkard's Walk: How Randomness Rules Our Lives", offers a fascinating lesson on the development of our understanding of probability and randomness, and how randomness is widely misunderstood and underestimated even today. Many of the examples are surprisingly counter-intuitive, such as the "Monty Hall problem", supporting the point that our brains tend to be wired counter to correct probabilistic reasoning.
Not only does Mlodinow succeed in making mathematical theory and history quite fascinating, but he demonstrates the applicability of randomness in our lives in ways that will make you ponder. The core chapters of the book present a sequence of concepts in probability theory, but rather than just present dry theory, Mlodinow takes the much more interesting approach of presenting the concepts by way of the history of their development, making it not only the story of the development of ideas, but of the colorful characters who contributed to them.
He does a good job of carefully explaining the concepts with good examples.
This was an easy read for me. Probability is a subject with which I struggle.
I've read a lot of introductory statistics material over the years (which of course says a good bit about my ability to understand statistics -- or lack thereof). But if you really are a specialist, then the popular books aren't generally for you, either. I recommend Black Swan as well, but if you have to choose, Drunkard's Walk is better.If you are a specialist in the field, then this book isn't for you. If that was so easy to do, then someone would have done it already. I agree that he does not as effective a job as others do in surveying all of the heuristics and biases. I have NEVER read a book that explains the concepts so well.
Read this book if you want to get a good intuitive understanding of what is going on. Lots of other people have said lots of other things about this book, and for the most part, I agree. If you know a good bit about statistics, then this book is not for you. Of these, by far the more interesting heuristically is the former, and skillfully uses examples (such as random number series) to show how it happens.
Using the Bayesian test, Mlodinow shows that the true question is: what percentage of women who were abused by their husband and were murdered were actually murdered by someone else.Mlodinow also effectively sets forth the issues of how human beings see order in randomness and randomness where there is order. For example, Alan Dershowitz argued that admitting evidence of OJ Simpson's abuse of his wife was irrelevant because only a minuscule number of women who are abused are also murdered by their husband. But all of these books are short and well-written: quite literally, you can read them all (or listen to them unabridged, as I did), and it will help the concepts stick in your head.But one book that this is clearly superior to is The Black Swan, by Nassim Taleb. You can't do better.
Taleb does discuss the problem of not knowing when you have a Gaussian distribution, but his account of the alternative "Mandelbrotian" way of thinking is just opaque (perhaps an occupational hazard, but then he shouldn't do it). They haven't. Taleb sticks with the "people see order when it's random" problem, but more than anything else, The Black Swan focuses on TALEB, not the problem. Moreover, a number of excellent books have appeared over the last couple of years that popularize and explain the Twersky/Kahneman "heuristics and biases" approach to life, so on that side, this book is not truly necessary.But what an explanation of statistics it is.
I think that Predictable Irrational (Dan Ariely), Nudge (Sunstein and Thaler), and Sway (Ori Branfman) are somewhat better than that. Lots of books (mostly textbooks) will explain the tests; what they won't do is give you a good intuitive sense of what these tests are doing, and how they work.Mlodinow also communicates with exceptional clarity about the nature of statistical fallacies. He explains the "normal curve," and then uses it to explain the underlying intuition behind Bayesian reasoning, the chi-squared test, and significance testing, just to name three. Note that what I am talking about is the intuitive notion behind the tests.
Humorous historical and practical explanation of the constant existence of chance in all physical behavior, and probably in the abstract world as well.
I haven't found a better book that does so without being so dry that I fall asleep. My hardcover copy had many instances of missing words in the text. While annoying, this is a simple matter, but hopefully corrected in the paperback edition.Mainly, as others have noted here, it really had only a few really good examples of how to better apply rules of statistics to real life.
|
