Internal Controls Design graphic Internal Controls Design
consulting and research for internal control with risk management

book main page

"A pocket guide to risk mathematics: key concepts every auditor should know": Content for different readers

When I started telling people who were not auditors about this book for auditors their reaction was "That sounds like just what I need." Their message is clear: this book also has value for people who are not auditors.

Content for risk people

Many people working in the area of risk today know less about quantification of risk than they would like to, and perhaps have to bluff their way through conversations with quantification experts. This book will help you understand what the experts are saying.

If your role involves going beyond a conceptual understanding and learning how to do the calculations then this book could be a great way to get started because it will help you to work out what you want to know. That will save you from, perhaps, spending time mastering the symbolic mathematics of techniques that are irrelevant to you or just not very good.

Beyond that, the depth of the conceptual understanding this book provides goes far beyond most textbooks on mathematics, even the thorough ones. You may find you understand some important points that even your expert colleagues do not.

Content for politicians and regulators

Many of the favourite methods of statistics result in systematic understatements of risk if applied in the usual mindless way. Is that something regulators should overlook? It's not hard to understand what is going on and think of ways to discourage it.

To take just one example, some methods for fitting models to historic data involve assuming the data are distributed according to the normal distribution. If the data are not distributed this way then the fitting doesn't work as well as it is assumed to. The usual way to check that the data are normally distributed is to apply a normality test. Here's the catch. Normality tests typically tell you the data are normally distributed unless they can confidently confirm that they are not. Assuming 'innocent unless proven guilty' makes sense in the courtroom but in model fitting it should be the other way around.

Content for theorists

For students and professional academics the most interesting part of this book is probably the explanation of the interpretations of probability. Instead of running through the history of theories about this it focuses on the interpretations actually in use most often today, revealing how people flip between interpretations, often without realising it.

It also starts with the idea of a probabilistic forecaster whose skill is to be assessed, rather than starting with intepretations. This way the starting point is the practical value of probabilities, which is one thing over which there is less controversy.



In the UK try Amazon.

In the USA try Amazon.

Anywhere else, check Wiley's page on where to buy.

  © 2010 Matthew Leitch