Introduction To Combinatorial Analysis John Riordan

A Working Look at Riordan's Combinatorics Text

I ran into John Riordan's Introduction to Combinatorial Analysis back when I was trying to get a handle on generating functions for a scheduling problem at work. The book is from 1958, and it shows it in the best way possible, which is that the mathematics hasn't aged a day. What has aged is the typesetting and the assumed background. You are expected to be comfortable with basic algebra and not shy away from summation notation. The book covers the usual territory: binomial coefficients, Pascal's triangle identities, generating functions, recurrence relations, the inclusion-exclusion principle, partition theory, and some basics of permutation groups. Riordan's approach is distinctly American combinatorics from that era. He builds things from the ground up using algebraic manipulation rather than abstract structures. This means you learn to actually compute things, not just categorize them.

Why People Still Reference Introduction To Combinatorial Analysis John Riordan

The reason this book keeps coming up in discussions is that later texts tend to assume you already understand the mechanical tricks Riordan takes the time to teach. When someone writes about exponential generating functions without explaining why you divide by n!, they are skipping steps that Riordan walks through explicitly. The result is a gap in intuition that only practice fills, and Riordan provides a lot of practice through worked examples and problems. I encountered a specific issue when using techniques from the book on a problem involving restricted permutations. The chapter on inclusion-exclusion gives clean theoretical coverage, but the worked examples assume small cases where you can manually expand everything. My problem involved about 40 constraints, which made direct application impractical. The workaround was to code up the inclusion-exclusion formula as a recurrence rather than expanding it symbolically. Riordan mentions this direction in passing near the end of that chapter but does not develop it into an algorithm. You have to take that hint and run with it yourself. One thing beginners consistently miss is the difference between ordinary and exponential generating functions and when to use each. The book introduces both early, but it does not spend enough time making the boundary obvious. The rule of thumb that actually works is this. If the objects you are counting are unlabeled and you care about their aggregate structure, use ordinary generating functions. If you are counting labeled structures where the labels matter for the enumeration, switch to exponential generating functions. This is not a rigorous theorem, but it has kept me from wasting hours on the wrong framework.

Another counter-intuitive point involves the manipulation of divergent series. Riordan treats generating functions formally, which means you can work with expressions like 1/(1-x) even when x=2 makes the series diverge. The algebra still produces correct coefficient extractions. This bothers people who learned analysis first. The practical benefit is that formal manipulation lets you derive recurrences faster than trying to establish convergence at every step. The risk is that you can accidentally apply an identity outside its domain. I once derived a closed form that looked elegant and then realized it only held for a restricted range of parameters. Checking the radius of convergence after a derivation takes about thirty seconds and prevents embarrassing errors later. The book also has real limitations. The treatment of analytic combinatorics is nonexistent because it predates Flajolet and Sedgewick by decades. If you need singularity analysis or saddle-point methods, Riordan is not going to help you. The combinatorial species framework is absent. The coverage of modern topics like combinatorial enumeration via computer algebra is simply not there. For those, you need more recent sources. Another weakness is the problem set. Some problems are straightforward applications of the preceding section. Others are genuinely difficult and the hints are thin. I spent roughly two hours on a single problem in the generating functions chapter that boiled down to recognizing a particular convolution identity. The answer in the back of the book is correct but skips the intermediate algebraic step that took me the longest to find. Having a solutions manual would have saved me time, but one does not widely exist for this edition.

If you are working through this book, the most efficient path is to focus on chapters two through five for a solid foundation. Chapter two on binomial coefficients alone is worth the price of the book. Chapter three on generating functions is where most of the practical techniques live. Chapter four on recurrence relations gives you the machinery to convert enumeration problems into solvable forms. After that, you can dip into partition theory and the later chapters as needed for your work. The book is available through several channels. It is in the public domain in some regions due to its age, which means you can find scanned copies on archive.org and similar repositories. Wiley still lists it in their catalog, and used copies circulate on Amazon and AbeBooks in the five to fifteen dollar range depending on condition. I recommend the Dover reprints if you can find them, since they are cheaper and the pages hold up reasonably well. My practical advice for using this material is to keep a notebook of identities you verify yourself rather than just reading them. Riordan states many results without full proof, assuming you will work them out. The ones you derive on paper are the ones you will remember when you actually need them under time pressure. This approach typically cuts down future reference time from twenty minutes of flipping through pages to about two minutes of locating the identity you already proved.

There is no single perfect textbook for combinatorial analysis, and Riordan's book occupies a specific niche. It is strong on computational technique and weak on modern structural perspectives. If your work involves enumerating concrete objects, solving recurrences, or doing the kind of counting that appears in algorithm analysis, this book will serve you well. If you need category-theoretic frameworks or asymptotic enumeration tools, look elsewhere. Knowing the gap between what the book covers and what you actually need is more useful than any general recommendation.