I have read cover-to-cover a previous edition of this book (when it was published by Wiley in 1998) and recently had an opportunity to carefully peruse this current edition (5th ed. by Bloomberg Press???). What I found is that this is a strange case of how a great book (the 1998 edition) turned into merely a good book (this 5th edition). Because of this regression toward the average, I deducted one star from my review (but still feel that it is good enough for 4 stars).
As you may have noticed, I really loved the older edition of The Economist Numbers Guide that I thankfully own. It is a great overview and introduction of mathematics as it relates to business. There are a lot of great things about that edition of this book. One of the things I admired about it was the range of topics covered, from interest rates and basic probability/statistics all the way up to Markov Chains, linear programming, and marginal analysis. It is hard to find the breadth of topics covered in that book elsewhere - whether all in one book or in any combination of books.
So I found it perplexing that this 5th edition dedacted some materials and topics covered in older editions. Gone are the interesting discussion of descriptive statistics for sets of data that do not easily conform to any of the standard probability distributions (e.g., where median is the best measure of the 'average' and substitutes must be used for the more common parameters such as standard deviation). I have a hard time finding anything coherent much less accessible on those topics elsewhere so it is a shame that they were left off of the 5th edition.
The only new material (not previously present) is a short blurb on public-key cryptography. While that topic is interesting to me and the limited discussion was illuminating and mathematically sound, it seemed a rather quixotic choice to put in when some interesting materials in previous editions were left off and new material that would have been more useful to the targeted audience have yet to be added.
What I mean by useful material that have yet to be added is that both the 1998 edition and this edition don't have some materials that I would think naturally ought to be added. E.g., the section on finance & investment mathematics is mostly devoted to various discussions on interest rate/time value of money & basic probability. I think adding material on CAPM (although CAPM Beta is defined in the book's very helpful glossary section), option/derivative pricing, financial portfolio optimization, and other topics in financial mathematics/engineering would make a great and natural addition to this book.
Some problems common to both the prior and current edition of this book are the occassional (relatively rare) typos. They are usually minor (although they are most annoyingly frequent in the section on time value of money / interest rates).
Another flaw in both the older and newer editions is that there are gaps in the expository material that don't make much sense. To be fair, this book is designed to be a brief intro/overview into a wide swathe of topics so it wouldn't be reasonable to expect that the author go into great detail on every topic. However, there are instances - e.g., the example on mixed strategies in game theory - where one or two additional sentences would help novices to understand (e.g., how did you get the the mixed strategy probabilities? author should have added a couple of more lines about how the system of equations are interrelated with one another when determining mixed strategies).
Having said all of that, let me reiterate that BOTH the old and the new edition of The Economist Numbers Guide is a wonderful resource for people interested in business mathematics. The sections on decision-making and forecasting are especially of value since they are so wonderfully explained here and a comparable set of explanations are hard to find elsewhere.
In future editions, I just hope that the author heeds my advice about bringing back some topics in older editions, correcting a few errors & lapses, and adding some material that would fit in with what has otherwise been an excellent series of books.