We have been taught that in a sample of numerous (hundred and thousands) of observations, the largest number of observations tend to converge toward the "mean," with most outcomes lying near the mean (i.e. 95% in the range of -/+ two standard deviations from the mean). The Bell Curve model naturally emphasizes convergence to the mean and with this model, it is easy to dismiss the observations in the either end of the tails, as being uninfluential, highly improbable "outliers."
Taleb presents to the reader however, numerous examples of such "outlier" events that transformed our paradigms. His chapters contain not only examples of real people and past events, but also compelling analyses of how we (as well as scholars and doctors) rationalize outlier events and proves the logical fallacy of such behavior. People behave this way when they are fooled by confirmation error, the fallacy silent evidence, future blindness, retrospective distortion and many more logical fallacies.
His book is filled with inductive logic and a lot of vocabulary requires logical comprehension; for example, it took me ages to differentiate between scalable and non-scalable (If you have a non-scalable job, you are paid by the hour you work). And what about Mediocristan and Extremistan! (In Mediocristan, nothing is "scalable" and one single extreme observation cannot affect the sum of all values in the distribution. In Extremistan on the other hand, extreme occurrences cansignificantly affect statistical properties) It is also complete with colorful narratives and the author's ego, which made it a very entertaining read. I should probably read it at least once again though, in order to fully appreciate the epistemology and logic behind the colorful narratives.
No comments:
Post a Comment