C++ Design Patterns and Derivatives Pricing (Mathematics, Finance and Risk) (Mathematics, Finance and Risk)

Joshi not only explains quickly how to implement the basic pricing techniques (trees, monte carlo), but also offers some good guidance on modern OO C++ techniques (patterns) along the way.

My only criticism is that the book is too short. I would have gladly read another 50% on top.


A Superb, Practical Introduction to Quant Programming   December 23, 2004
 4 out of 7 found this review helpful

This book stands out because it has clearly been written by a quant practitioner rather than a pure theorist. It regularly focuses on the real life challenge of producing a performant, maintainable system which can be enhanced over time with minimal impact.

Though it does not claim to be an introduction to C++ or Mathematical Finance, its onion-skin approach is effective in guiding those with a very basic knowledge in these two areas into some reasonably advanced terrain.


A fast-paced introduction to sophisticated C++   December 8, 2004
 7 out of 7 found this review helpful

C++ is a big language. It has a lot of features and even more in the way of clever ideas for combining these features to create effective code. Most books on the subject are dauntingly huge, at least to the novice. Perhaps the greatest virtue of this book is that it covers a lot of ground while being slim enough (under 200 widely-spaced pages) to be easily read from cover to cover.

The book is clearly aimed at those in quantitative finance and the examples are designed to be of genuinely useful code that a quant might write. This is a refreshing change from more general books where the examples are either silly ("a Sauternes is inherited from a Bordeaux and has a pour() member function") or do the kind of low-level manipulation (such as container classes) that few programmers write. The examples are given in full, even if this involves repeating a good deal of code, but I suppose this does give the reader the satisfaction of quickly skimming two or three pages from time to time.

Aside from a basic acquaintance with mathematical finance, the reader is required to have a rudimentary knowledge of C++. However the more advanced language features, such as virtual functions and templates, are explained concisely as they are introduced. Writing good object-oriented code depends very much on knowing not so much the syntatical rules but why the language features are there and when to use them. This is the emphasis of this book; it gives very clear well-reasoned guidance for effective use of the language. One very important aspect of this is the use of so-called "design patterns". I'm not sure this term is ever defined but it becomes clear through many examples that it refers to clever ways to combine language features to achieve particular generic goals.

There are places where the author introduces several ideas at once and almost trips over himself in an effort to explain them all. But once the reader becomes accustomed to this and is therefore prepared to wait a little for an explanation of something, this is an effective way of getting through the material rapidly. Consequently this is a very readable book, in which the alert reader can learn a lot in a short time. All in all, this book is ideal for someone starting to write C++ for finance or even for a more experienced quant who feels his code is insufficiently sophisticated.


The boss and his/her requests   October 20, 2004
 2 out of 3 found this review helpful

Having implemented mathematical finance solutions in a variety of programming languages, I found in Joshi's book the ideal guide to move to the C++ way of doing things.
The book certainly requires some knowledge of C++ (at least an introductory book) but leads the reader through a learning process focused on the reasons why C++ is the most used language in financial engineering.
It starts for the simplest non-OO implementation, highlighting (pros and) cons, and then moves on to the appropriate C++ solution. In the way, fundamental design patterns are introduced and used (strategy, decorator, bridge...). By the end of the book the reader will have seen how to build a Monte Carlo engine for exotic options and how to simplify its interface through the more complicated factory pattern.
Numerically, the book is mainly focused on Monte Carlo methods but it also spends a chapter on trees (binomial) and one on templatised root finding (for the implied vol). PDEs are not dealt with.
The never-ending requests of the "evil boss" are a constant driver in a quest for reusability and generality and the book clearly shows how to achieve these. Knowing how evil bosses can be I would say that this book is certainly a BUY!!!