Epsilon Theory: Core Curriculum

I haven’t any right to criticize books, and I don’t do it except when I hate them. I often want to criticize Jane Austen, but her books madden me so that I can’t conceal my frenzy from the reader; and therefore I have to stop every time I begin. Every time I read Pride and Prejudice I want to dig her up and beat her over the skull with her own shin-bone.

          – Mark Twain

Once upon a time in the dead of winter in the Dakota Territory, Theodore Roosevelt took off in a makeshift boat down the Little Missouri River in pursuit of a couple of thieves who had stolen his prized rowboat. After several days on the river, he caught up and got the draw on them with his trusty Winchester, at which point they surrendered. Then Roosevelt set off in a borrowed wagon to haul the thieves cross-country to justice. They headed across the snow-covered wastes of the Badlands to the railhead at Dickinson, and Roosevelt walked the whole way, the entire 40 miles. It was an astonishing feat, what might be called a defining moment in Roosevelt’s eventful life. But what makes it especially memorable is that during that time, he managed to read all of Anna Karenina. I often think of that when I hear people say they haven’t time to read.

          – David McCullough

You need to read more science fiction. Nobody who reads science fiction comes out with this crap about the end of history.

          – Iain Banks

A scholar is just a library’s way of making another library.

          – Dan Dennett, “Consciousness Explained”

I still feel – kind of temporary about myself.

          – Willy Loman (“Death of a Salesman”, by Arthur Miller)


I’ve had dozens of requests to put together a reading list for Epsilon Theory, and I’ve resisted. There’s something uncomfortable about telling people what they should read, of recommending this book but not that one to you because it happened to resonate with me. Also, as Mark Twain said, I haven’t any right to criticize other authors (unless I really, really hate their books!), and I have zero interest in engaging in the all-too-familiar academic exercise of dueling criticism. Been there, done that.

But the roads I’m trying to explore with Epsilon Theory are not the commonly traveled paths of investment theory and practice, and I can appreciate that it would be useful for my readers to have some sort of field guide for this unfamiliar territory. And from a personal perspective, it’s more than just comforting to compile a list like this. Books provide grounding. They help us feel less temporary about ourselves, to use Willy Loman’s memorable phrase, and I’m certainly no exception to that. So with the caveat that this is an entirely impressionistic and non-comprehensive exercise in what has been useful for my personal intellectual grounding, and should be thought of as pointing you to a shelf in the library from which you might want to engage in your own exercise in discovery, here goes ..

Statistics and Econometric Analysis

Edward Tufte is best known today for his books on information display – The Visual Display of Quantitative Information, 2nd ed. (2001), Envisioning Information (1990), Visual Explanations: Images and Quantities, Evidence and Narrative (1997), etc.  These books are amazing in every way, both as resource and as inspiration. But less well known is Tufte’s academic career as a statistician at Princeton’s Woodrow Wilson School, where he collaborated with one of the founders of Information Theory, John Tukey. In 1974 Tufte wrote what I believe is the single best book in explaining both the basic techniques and the meaning of applied statistics – Data Analysis for Politics and Policy. Whether you’re a newbie to statistical applications or a Ph.D in econometrics, this is an incredibly useful book (and although out of print, there are plenty of used copies circulating on Amazon).

I’ve mentioned the work of Gary King, Director of Harvard’s Institute for Quantitative Social Science, in my note Rise of the Machines. If you’re already a fluent speaker of the language of econometrics, his work on ecological inference and likelihood functions is both groundbreaking and extremely useful. But one of King’s great skills is his ability to apply econometric techniques to pretty much any area of inquiry, including fields of study that, for whatever reason, tend to be strangers to this methodology. King’s book with Robert Keohane and Sid Verba – Designing Social Inquiry: Scientific Inference in Qualitative Research (1994) – is an excellent read for anyone who wants to think more rigorously about causality and inference in markets, politics, and our social lives. This is not a statistics book. It’s a how-to-think-about-statistics book.

Game Theory and Information Theory

The classic primer on game theory is Games and Decisions: Introduction and Critical Survey, by Duncan Luce and Howard Raiffa. Originally published in 1957, the 1989 edition is still in print, still assigned in courses all over the world, and is still the most comprehensive work on game theory for non-mathematicians. If you’re really interested in game theory, you must have this book. Other classic books on game theory include Robert Axelrod’s Evolution of Cooperation and Thomas Schelling’s The Strategy of Conflict. These are wonderful books, full of real-life examples of game theoretic applications. But if you only have time for one classic book on game theory, I’d recommend William Riker’s The Art of Political Manipulation (1986). The chapter on Abraham Lincoln and the strategic decision-making involved with the Lincoln-Douglas debates is alone worth the price of admission.

As for information theory, the obvious classic is Claude Shannon’s The Mathematical Theory of Communication, which served as the midwife for this entire field. It’s for the really serious student, though, the equivalent of reading Von Neumann and Morgenstern to learn about game theory. Much more accessible are two books by James Gleick: Chaos (1987) and The Information (2012). Thinking about the world in terms of information and subjective probabilities goes back at least to the 18th century work of Thomas Bayes, and two recent works – Nate Silver’s The Signal and the Noise (2012) and Sharon McGrayne’s The Theory That Would Not Die (2011) do a Gleick-ian job (high praise indeed!) of making this perspective accessible for the non-specialist.

The game and information theoretic research that I’m finding most useful today is not written by economists or political scientists, but by linguists and biologists. There’s a scope and a sweep to this work, as well as an explicit incorporation of evolutionary theory, that I find extremely useful. David Lewis’s book Convention (1969) is a good example of this, as he is trying to explain the development of social coordination solely with the tools of strategic decision-making under informational uncertainty (game theory), without resorting to the deus ex machina of human exceptionalism or consciousness.

There’s nothing special about the human animal in this perspective, no sense in which some unique reasoning capacity has created this convention that is more True with a capital T than that convention. In Signals (2010), Brian Skyrms takes the Convention games developed by Lewis to a new level of usefulness by directly

1, 23  - View Full Page