As the title suggests, the aim of the book is to apply simulation in grasping the fundamental and advanced areas of statistical computing; as such, it combines probability theory with computer programming. While the reader is free to choose his or her own programming language, the author has selected the R programming language to solve the exercises (R is a free statistical package with programming facilities). Some of the topics covered in the book include random number generation, simulating statistical models (multivariate normal, Markov chains, Poisson process, and so on), Monte Carlo and Markov chain Monte Carlo (MCMC) methods, Bayesian and bootstrap computations, and continuous time models (Brownian motion, stochastic differential equations, and so on).

These topics are spread evenly across six chapters. Additionally, there are three useful appendices comprising probability reminders, programming in R, and answers to exercises, followed by a helpful bibliography and a subject index.

As mentioned in the preface, there are advantages as well as disadvantages of a computer simulation-based approach as taken by the author. While simulation-based results provide only approximate solutions to problems, it is also true that one can cover a wider domain of problems and more importantly those involving huge and complex systems where purely theoretical analysis is tedious if not impossible.

Some plus points of the book:

(1) It covers both traditional (like MCMC) and advanced topics (such as reversible MCMC) of statistical computing.

(2) It explains the applications with requisite theory.

(3) It explains the topics with exercises and solutions in R, a free package.

(4) It includes a chapter on continuous time models and a useful appendix on programming in R.

However, I do not agree with the author’s remark in chapter 1 (p. 1) that π is not random at all. He is unaware that π has been successfully used as a random number generator. Here is why. First, there will be no cycle in case of π, as, being an irrational number, it involves a non-terminating and non-repeating fraction. Voss’ proposal that every time we run the code the output can only start with the digits 3.14159 ... is simply absurd, because when we are simulating we shall always start from an arbitrary position and not from the beginning. Moreover, if one picks up a string of arbitrary length from an arbitrary position (this position serves as the seed of the random number generator in case of π, while the length is determined by our sample size requirement), it is likely to pass the common randomness tests barring a few exceptions [1]. The author is also not aware that previous attempts to disprove the randomness of π have been already rejected by the noted statistician George Marsaglia [2]. Last but not least, it must be emphasized that numbers generated from an arbitrary seed need not be random, a drawback most random number generators suffer from, as pointed out clearly by Matsumoto et al. [3].

The book would benefit from a chapter on various sampling algorithms, such as simple random sampling without replacement. It is otherwise interesting and self-contained, although some knowledge of probability and programming is desirable. The target readership includes postgraduates, teachers, researchers, and practitioners of statistics and computing.