Rating: 8.3/10.

Tells the story of how statistics emerged as a scientific discipline in the 20th century. The title comes from an apocryphal story by Fisher describing an experiment to see if a lady can taste the difference between two ways of making tea. The book describes the lives and circumstances of the people involved, explains the math pretty well using words, without getting too technical with equations. Some of the earliest personnel:

- Karl Pearson (1857-1936) was the founder of mathematical statistics, devised methods of estimating statistical parameters from data, founded the journal
*Biometrika*, applied these methods to confirm Darwin’s theory of natural selection. He had a dominating personality, and his son Egon Pearson also became a famous statistician. - William Gosset (1876-1937) discovered the Student t-distribution while working for Guinness, improving methods to brew beer. He had to publish under the pseudonym “Student” because Guinness wouldn’t let their employees publish.
- Ronald Fisher (1890-1962) was a genius that invented a lot of modern statistics including MLE for estimating parameters, ANOVA, experimental design. Originally used these methods to study the effects of fertilizers on crop variation, eventually became a distinguished professor. Fisher did not get along with Pearson, also dismissed evidence that smoking caused cancer long after it was accepted by the scientific community.
- Jerzy Neyman (1894-1981) invents the standard textbook formulation of hypothesis testing against a null hypothesis, introduces the concept of confidence interval. Fisher and many others are skeptical since it’s unclear what is the interpretation of p-value and 95% probability of the 95% confidence interval.

The first half was more coherent as it described the earliest days of statistics as an emerging field. The second half of the book is a series of quick summary biographies of a number of statisticians, such as Kolmogorov (axiomizing probability), Wilcoxon (nonparametric stats), Tukey (robust estimation). I felt this was not very coherent since it jumped around rapidly between too many minor characters without any central theme. One annoying pattern was whenever there was a woman, the author had to emphasize her gender and spend a few paragraphs talking about how remarkable it is for a woman to do statistics at that time, leaving less space to how she contributed to the field.

Undoubtedly, statistics changed the way we do science, and this book tells the story of how it happened. The book was published in 2001 so it doesn’t include the recent developments in machine learning and the difference between the two cultures. Also, instead of so many short biographies in the latter half, I would’ve preferred a discussion of the current research directions and open problems of mathematical statistics.