Requested Prisoner’s Dilemma by William Poundstone (1992) from the library after someone on the BoingBoing forums mentioned it, in context of the very early Cold War. I enjoyed it and learned a lot, even though I did not find its two major topics to gel as well as the author had intended.

As I grew up in the 1980s, critical analyses of 20th-century nuclear-arms policies that terrified me as a child intrigue me as an adult. The forum-commenter accurately related how this book closely examines to the especially strange days immediately following World War II, when the US and the USSR began to sink into inevitable, mutual enmity — but only the former possessed nuclear weapons to a significant degree. This led to many highly ranking or respected voices in the US and western Europe (such as the secretary of the US Navy) to advocate a “preventive war” with the USSR. This would have involved delivering an existential ultimatum to them as soon as possible, with an actual resolve to follow up with thermonuclear devastation and subsequent occupation should they decline to accept.

Against this historical setting, Prisoner’s Dilemma also offers a biography of John von Neumann — “Johnny” to both his friends and to the speaking voice of this book — who, among many other world-changing accomplishments in a life cut short, invented game theory as a tool for expressing or predicting the actions of rational actors in mutual competition. Circa 1950 (much as today), the most well-known mathematical/philosophical parable of game theory is the prisoner’s dilemma. Professional think-tankers of the time, including von Neumann’s colleagues at the RAND Corporation, could not help but see parallels between the questions raised by this puzzle and those faced by the United Stated during the early years of the Cold War, before MAD took hold as ad-hoc international policy.

Though I came at this book with a working knowledge of game theory, it taught me a lot about its important underpinnings without needing to resort to any explanatory tools more complicated than well crafted prose. I have a much stronger understanding now of von Neumann’s “minimax” theory, and how it applies to game-theory payoff matrices — including that of the prisoner’s dilemma itself. I could probably also fake my way through an explanation of Nash equilibria and saddle-points, if I had to. More interestingly, I have a richer understanding of “one-shot” versus iterated dilemmas, now, and the separate metaphorical uses of both.

I furthermore learned a lot of fascinating history about the post-war formation of the transatlantic nuclear threat, such as the fact that the USSR apparently relied less on espionage to puzzle out how the US had made its bombs, and more on simply waiting for fallout from US nuclear tests to drift on over, ready for analysis and deduction about the nature and the relative proportions of the bombs’ constituent isotopes.

However, I found the book lacking in connective tissue between the chapters about game theory and those looking at mid-century nuclear policy. While the book makes clear that study the prisoner’s dilemma arose from this geopolitical backdrop, it didn’t seem to draw any direct causal lines between the two topics, in either direction. I expected a tying-together where I would learn how nuclear decision makers used the prisoner’s dilemma as a predictive model, but this never quite came about. Even though Prisoner’s Dilemma ultimately felt like two tangentially related books accidentally interleaved with one another, I still found both very well written examinations of interesting topics, and I’m glad I read the whole thing.


Next post: Plerd levels up

Previous post: Very good winter loafers