Quote:
Originally Posted by rickroll
started reading A Drunkards Walk recently, nothing new really but it's an interesting read on keeping me humble when i do data analysis
I read it shortly after it came out, when it was one of the first of a group of books to point out the dangers of statistics, and I found it very interesting. While it makes the familiar point indicated in its subtitle ("How Randomness Rules Our Lives "), it does so in an engaging fashion, arguing that humans are hardwired to refuse to recognize randomness and to see instead (and overinterpret) patterns. In its early pages Mlodinow makes the closely related point—that people are not intuitively equipped to understand probability—through a retelling of the Monte Hall paradox, followed by a useful discussion of why M.D.s are very bad at interpreting probabilities—for example, because they misapply statistics about false positives.
Among his many examples of how badly we judge and evaluate, and of how we impose expectations, is Mlodinow's history of the random events that saw Bill Gates rise from minor league software programmer to the place of power he occupied—which reinforces his point that careers that seem to have some shape to them are often just a series of lucky accidents.
One of the benefits of reading this book as a non-mathematician and someone who's never taken a stats course is that it left me with a better grasp of concepts such as standard deviation, regression toward the mean, and confirmation bias.
Last edited by RussellinToronto; 09-09-2019 at 08:52 PM.