Hide table of contents

Cross-posted to LessWrong

The Strategy of Conflict, published in 1960 by Thomas Schelling, is:

A series of closely interrelated essays on game theory, [which] deals with an area in which progress has been least satisfactory - the situations where there is a common interest as well as conflict between adversaries: negotiations, war and threats of war, criminal deterrence, extortion, tacit bargaining. It proposes enlightening similarities between, for instance, maneuvering in limited war and in a traffic jam; deterring the Russians and one's own children; the modern strategy of terror and the ancient institution of hostages. (source)

Here’s a free PDF of the book, and here’s the Audible link.[1]

In this post, I’ll: 

  • Share some overall thoughts on the book
  • Share the Anki cards I made for myself when reading the book[2]
    • I intend this as a lower-effort alternative to writing notes specifically for public consumption or writing a proper book review
    • If you want to download the cards themselves (along with other nuclear-war-related cards) to import them into your own deck, follow this link.
  • List some other resources that people could consider reading/watching if they want to learn about game theory and (nuclear) war

My hope is that this post will draw some people's attention to this interesting book, help them decide whether reading it is worth their time, and help them quickly gain some key insights from the book. (See also Should pretty much all content that's EA-relevant and/or created by EAs be (link)posted to the Forum?)

Overall thoughts on the book

I read The Strategy of Conflict because someone told me they think this is a good book to read if one wishes to understand how international tensions might escalate to nuclear war, how many weapons might be used in a nuclear war, and whether countervalue targeting would occur. The book’s concepts, framings, insights, and examples indeed seemed useful for that purpose, as well as for a surprisingly wide range of other issues. 

This was despite me already knowing some game theory - maybe around as much as one learns in a typical undergraduate economics degree? That said, I expect someone who already knows much more game theory than I did would get somewhat less out of the book.

I’ve ranked this as roughly the 9th (out of 44) most useful-seeming EA-relevant book I’ve read since learning about EA. If I wasn’t doing research related to war, I think I might’ve ranked it closer to 20-25th.

There were some sections of the book that I found uninteresting and/or hard to follow - particularly chapter 9, which I ended up partly skimming. This was basically due to some parts of the book using lots of symbols and equations without regularly reminding readers what the symbols stood for or regularly walking readers through the intuition of what’s going on. But perhaps more maths/econ/technically minded people would have no trouble with those parts of the book. 

I also skipped Appendices B and C, as it sounded like they would be hard to follow and would cover relatively technical points about game theory that might be outdated by now anyway (i.e., perhaps the field has already elegantly resolved the issues Schelling was noting).

My anki cards

What follows are the Anki cards I made for myself. Some include direct quotes without having quote marks, while others are just my own interpretations (rather than definitely 100% parroting what the book is saying) but don’t note that fact. It’s possible that some of these cards include mistakes, or will be confusing or misleading out of context. 

The indented parts are the questions. The answers are in "spoiler-blocks"; hover over them to reveal the text. And the parts in square brackets are my notes-to-self.

---

What two legal privileges does Schelling mention that corporations have?

 The right to sue and the right to be sued

[p.43]

---

Schelling says "The mandatory secret ballot is...

 ...a nuisance to the voter who would like to sell his vote, but protection to the one who would fear coercion"

---

What term does Smithies use for the tactic of deliberately exhausting one's annual budgetary allowance so early in the year that the need for more funds is irresistibly urgent?

 Coercive deficiency

[Smithies' book was called The Budgetary Process in the United States, so I'm guessing he used the term in relation to the US government specifically.

Cited by Schelling.]

---

Schelling defines tacit bargaining as...

 bargaining in which communication is incomplete or impossible

[p.53]

---

Schelling notes that the concept of "coordination" that's relevant in tacit bargaining is also often relevant in explicit bargaining. What are 3 common features of agreements which illustrate this?

  • Agreements often involve "mathematical simplicity" (e.g., $2500, 1%, 50/50)
  • Agreements often mimic precedent more than is logically or legally necessary
  • Agreements often follow mediator's suggestions more than is logically or legally necessary

[p.67-68. I'm paraphrasing. I just chose the first 3 examples; he also gave others.]

---

Schelling suggests a big part of bargaining skill is...

 The ability to effectively influence how the problem is formulated, what analogies or precedents the definition of the bargaining issue calls to mind, and the kinds of data that may be available to bear on the question

[p.69]

---

What point from Schelling's discussion of tacit bargaining reminded me of the landmine ban treaty and Treaty on the Prohibition of Nuclear Weapons?

 The claim that it can sometimes be necessary/important for the terms of agreement to be qualitatively distinguishable from the alternatives, and not simply distinguishable by degree

---

Schelling suggests that forces taking/sacrificing any land beyond their side of 38th parallel or the Taiwan Strait could have...

 ...signalled a determination to advance even farther and/or an intention to retreat even farther

[And it thus could've been unstable. P.76]

---

Schelling gives a three-part typology of games of strategy, in which one type is pure-conflict games. What are the other two types?

  • Pure common-interest games / coordination games
  • Bargaining games / mixed-motive games

[p. 88-89. I'm not sure if these are the standard terms nowadays.]

---

Schelling gives a three-part typology of games in which one type is games of luck. What are the other two types?

  • Games of skill
  • games of strategy

---

Schelling highlights 5 areas of asymmetries between the two players in a threat situation which make threats a rich subject for study. These are asymmetries in…

  • The communication system
  • The enforceability of threats and of promises
  • The speed of commitment
  • The rationality of expected responses
  • (In some cases) the relative-damage criterion [note: Schelling highlights that threats can be effectively made even in some cases when execution would harm the threatener more than the target]

[p. 131]

---

Schelling says enforcement of agreements/promises depends on at least two things:

  • Some authority somewhere that can punish/coerce players
  • An ability to discern whether punishment/coercion are called for or not

[I think this idea could be roughly paraphrased as:

  • The ability to guarantee that punishment would happen if noncompliance is detected
  • The ability to guarantee that noncompliance would be detected

P.132]

---

Schelling defines a "strategic move" as:

 "A move that influences the other person's choice, in a manner favourable to one's self, by affecting the other person's expectations on how one's self will behave"

[p. 160. "One constrains the partner's choice by constraining one's own behaviour".]

---

What does Schelling suggest operates similarly to the "threat" of inadvertent war?

The risk that limited war could escalate to general/all-out war (with that escalation not necessarily depending on deliberate decisions by leaders)

[~p.190-195]

---

Schelling suggests compellent threats often take the form of administering punishment ___, whereas deterrent threats often take the form of administering punishment ___.

until the other person acts (in the desired way); if the other person acts (in the undesired way)

[p.196

Punishment could include things like increased risk of general war.

He wasn’t saying that compellant threats always take the form of administering punishment until the other person acts. They could also take the form of administering punishment if the other person doesn’t act in the desired way.]

---

What are 2 key points about how the "threat" represented by risk of inadvertent war might deter aggressive acts and crises?

  • Such acts and crises would probably increase the risk of inadvertent war (and both parties know this, giving them incentive to avoid such acts and crises)
  • The risk isn't fully controllable by the "threatener", making it more credible as a threat

[Schelling, p. 189-190.]

---

The risk of inadvertent war can operate like a threat. If provocative acts are taken despite this "threat" yet war still doesn't result, this…

isn't strong evidence the "threat" was a bluff; the actors involved may have just gotten lucky

[Just as surviving a single round of Russian Roulette isn't strong evidence against the gun being loaded.

Schelling, p. 189-190.]

---

The risk of inadvertent war can operate like a threat. Does this require that either actor intended to use that risk as a threat?

No

[Schelling, p. 189-190.]

---

Schelling says there's a distinction between nuclear and other weapons in the sense pertinent to the limiting of war. He also says that, if he's right about that but nevertheless the US wants maximum freedom to use atomic weapons, then we ought to, in the interest of limiting war, ...

destroy or erode the distinction between nuclear and other weapons as best we can

[This is to make it less likely that limited US use of nuclear weapons is charged with excessive symbolic significance, as evidence of the US rejecting all relevant limits and distinctions.

P. 264]

---

Schelling argues that a limitation on the number of missiles two adversaries can have might be more stabilising the larger the number permitted. He gives two reasons for this:

  • That would increase the number of missiles each side should expect each side would have left over if it was struck first (since each missile has a <100% chance of taking out an enemy missile)
  • That would increase how many more missiles either side would need to add to its stockpile in order to be capable of a first strike that leaves their adversary with less than some specified number of missiles for retaliation

[The first effect increases the expected size of retaliation, thus deterring first strikes.

The second effect makes it harder to cheat on the agreement by disguising/hiding extra missiles, and to break the agreement and race to achieve a dominant number of missiles.

P.236

He doesn't suggest that this is a complete analysis or his all-things-considered view. This doesn't seem to account for risks of inadvertent or accidental escalation. 

I think MIRVs hadn’t been developed when he first wrote this. I think the existence of MIRVs makes this false, since each MIRVed missile can carry multiple warheads and thus might take out more than 1 enemy missile in expectation.]

---

Schelling highlights 9 institutional and structural characteristics of bargaining situations that may affect how commitment tactics are/can be used. These are…

  • Use of a bargaining agent
  • Secrecy vs publicity
  • Intersecting negotiations
  • Continuous negotiations
  • A restrictive agenda
  • Possibility of compensation
  • Mechanics of negotiation
  • Principles and precedents
  • Casuistry

[p.28-35. That's a very long answer. It'd be good if I could somehow categorise or chunk these ideas.]

Some other resources on similar topics

  • Six years later (in 1966), Schelling published the book Arms and Influence
    • I haven’t read this, so I can’t compare it to The Strategy of Conflict myself
    • But it’s possible it’s more up-to-date and useful
      • An interesting study I read kept referencing that book and didn't reference The Strategy of Conflict
      • On the other hand, I think I recall another source referencing The Strategy of Conflict as an essential classic without mentioning Arms and Influence
  • I recently finished the 24-part lecture series Game Theory with Ben Polak from Yale.
    • I found that useful, interesting, and surprisingly engaging (the lecturer is good at avoiding dryness and bringing the topic to life)
  • In 2008, RAND produced the monograph Dangerous Thresholds: Managing Escalation in the 21st Century
    • The whole text is available as a free PDF at that link
    • This is focused on war (especially but not only nuclear war), but still includes a fair amount of theory, at least in Chapter 2
    • I only read the summary and the first two chapters, but found it very interesting and useful
      • I might read the rest later
        • The main reason I’ve hesitated is that it’s from 2008, and the topics beyond Chapter 2 seem like the sort of things that may have changed notably in the 13 years since then
    • Let me know if you'd find it useful for me to share the Anki cards I made when reading that (it wouldn't take me long)
  • See also Books / book reviews on nuclear risk, WMDs, great power war?

Note: This post contains only my personal opinions, and was written in a personal capacity rather than in my capacity as a researcher for Rethink Priorities.

[1] I generally prefer listening to audiobooks rather than reading physical books/ebooks, but in this case I downloaded the free PDF into my iPad’s Kindle app. This was because a few reviewers of this book on Audible suggest that they think they’d have been better off with a physical book rather than the audiobook, because the book uses a fair few equations and graphs.

[2] See here for the article that inspired me to actually start using Anki properly. Hat tip to Michelle Hutchinson for linking to that article and thus prompting me to read it. Note that some of the Anki cards that I made and include in this post violate some of the advice in that article - in particular, the advice to try to ensure that questions and answers each express only one idea. 

Comments4


Sorted by Click to highlight new comments since:

Meta: Does this sort of post seem useful? Should there be more posts like this?

I previously asked Should pretty much all content that's EA-relevant and/or created by EAs be (link)posted to the Forum? I found Aaron Gertler's response interesting and useful. Among other things, he said:

Eventually, we'd like it to be the case that almost all well-written EA content exists on the Forum somewhere.

[...]

I meant "quite EA-relevant and well-written". I don't especially care whether the content is written by community members, though I suppose that's slightly preferable (as community members are much more likely to respond to comments on their work).

[...]

A single crosspost with a bit of context from the author -- e.g. a few sentences each of summary/highlights, commentary, and action items/takeaways -- seems better to me than three or four crossposts with no context at all. In my view, the best Forum content tends to give busy people a quick way to decide whether to read further.

And I read a lot of stuff that I think it could be useful for at least some other EAs to read, and that isn't (link)posted to the Forum. So Aaron's comments, combined with my own thinking and some comments from other people, make me think it'd be good for me to make linkposts for lots of that stuff if there was a way to do it that took up very little of my time.

Unfortunately, writing proper book reviews, or even just notes that are geared for public consumption, for all of those things I read would probably take a while. 

But, starting about a month ago, I now make Anki cards for myself anyway during most of the reading I do. So maybe I should just make posts sort-of like this one for most particularly interesting things I read? And maybe other people could start doing that too?

A big uncertainty I have is how often the cards I make myself would be able to transmit useful ideas even to people who (a) aren't me and (b) didn't read the thing I read, and how often they'd do that with an efficiency comparable to people just finding and reading useful sources themselves directly. Another, related uncertainty is whether there'd be any demand for posts like this. 

So I'd be interested in people's thoughts on the above.

As an avid Anki user, I actually prefer a deck over a review or a summary, other things equal. Most people don't use Anki or other spaced-repetition software, though, so they are unlikely to share this preference.

Incidentally, could you make the deck available as an apkg file?

Here's a deck of all my nuclear-war-related cards so far. Most aren't from Schelling, but if you just want those I think it should be easy enough to delete the rest (all the Schelling-related ones says Schelling somewhere).

I haven't shared an Anki deck before, so let me know if it doesn't work.

(Edited to add: I've now edited the post itself to include a link to this deck.)

I started reading Arms and Influence at the library.  I would it say it further applies the concepts of SoC, specifically to war or nuclear threat scenarios but SoC would be the first principles to analyze conflict.  Interestingly, in the forward I read, Schelling mentions he realizes some of the things are out of date but left it in for historical context. 

Curated and popular this week
trammell
 ·  · 25m read
 · 
Introduction When a system is made safer, its users may be willing to offset at least some of the safety improvement by using it more dangerously. A seminal example is that, according to Peltzman (1975), drivers largely compensated for improvements in car safety at the time by driving more dangerously. The phenomenon in general is therefore sometimes known as the “Peltzman Effect”, though it is more often known as “risk compensation”.[1] One domain in which risk compensation has been studied relatively carefully is NASCAR (Sobel and Nesbit, 2007; Pope and Tollison, 2010), where, apparently, the evidence for a large compensation effect is especially strong.[2] In principle, more dangerous usage can partially, fully, or more than fully offset the extent to which the system has been made safer holding usage fixed. Making a system safer thus has an ambiguous effect on the probability of an accident, after its users change their behavior. There’s no reason why risk compensation shouldn’t apply in the existential risk domain, and we arguably have examples in which it has. For example, reinforcement learning from human feedback (RLHF) makes AI more reliable, all else equal; so it may be making some AI labs comfortable releasing more capable, and so maybe more dangerous, models than they would release otherwise.[3] Yet risk compensation per se appears to have gotten relatively little formal, public attention in the existential risk community so far. There has been informal discussion of the issue: e.g. risk compensation in the AI risk domain is discussed by Guest et al. (2023), who call it “the dangerous valley problem”. There is also a cluster of papers and works in progress by Robert Trager, Allan Dafoe, Nick Emery-Xu, Mckay Jensen, and others, including these two and some not yet public but largely summarized here, exploring the issue formally in models with multiple competing firms. In a sense what they do goes well beyond this post, but as far as I’m aware none of t
LewisBollard
 ·  · 6m read
 · 
> Despite the setbacks, I'm hopeful about the technology's future ---------------------------------------- It wasn’t meant to go like this. Alternative protein startups that were once soaring are now struggling. Impact investors who were once everywhere are now absent. Banks that confidently predicted 31% annual growth (UBS) and a 2030 global market worth $88-263B (Credit Suisse) have quietly taken down their predictions. This sucks. For many founders and staff this wasn’t just a job, but a calling — an opportunity to work toward a world free of factory farming. For many investors, it wasn’t just an investment, but a bet on a better future. It’s easy to feel frustrated, disillusioned, and even hopeless. It’s also wrong. There’s still plenty of hope for alternative proteins — just on a longer timeline than the unrealistic ones that were once touted. Here are three trends I’m particularly excited about. Better products People are eating less plant-based meat for many reasons, but the simplest one may just be that they don’t like how they taste. “Taste/texture” was the top reason chosen by Brits for reducing their plant-based meat consumption in a recent survey by Bryant Research. US consumers most disliked the “consistency and texture” of plant-based foods in a survey of shoppers at retailer Kroger.  They’ve got a point. In 2018-21, every food giant, meat company, and two-person startup rushed new products to market with minimal product testing. Indeed, the meat companies’ plant-based offerings were bad enough to inspire conspiracy theories that this was a case of the car companies buying up the streetcars.  Consumers noticed. The Bryant Research survey found that two thirds of Brits agreed with the statement “some plant based meat products or brands taste much worse than others.” In a 2021 taste test, 100 consumers rated all five brands of plant-based nuggets as much worse than chicken-based nuggets on taste, texture, and “overall liking.” One silver lining
 ·  · 6m read
 · 
Cross-posted from Otherwise. Most people in EA won't find these arguments new. Apologies for leaving out animal welfare entirely for the sake of simplicity. Last month, Emma Goldberg wrote a NYT piece contrasting effective altruism with approaches that refuse to quantify meaningful experiences. The piece indicates that effective altruism is creepily numbers-focused. Goldberg asks “what if charity shouldn’t be optimized?” The egalitarian answer Dylan Matthews gives a try at answering a question in the piece: “How can anyone put a numerical value on a holy space” like Notre Dame cathedral? For the $760 million spent restoring the cathedral, he estimates you could prevent 47,500 deaths from malaria. “47,500 people is about five times the population of the town I grew up in. . . . It’s useful to imagine walking down Main Street, stopping at each table at the diner Lou’s, shaking hands with as many people as you can, and telling them, ‘I think you need to die to make a cathedral pretty.’ And then going to the next town over and doing it again, and again, until you’ve told 47,500 people why they have to die.” Who prefers magnificence? Goldberg’s article draws a lot on author Amy Schiller’s plea to focus charity on “magnificence” rather than effectiveness. Some causes “make people’s lives feel meaningful, radiant, sacred. Think nature conservancies, cultural centers and places of worship. These are institutions that lend life its texture and color, and not just bare bones existence.” But US arts funding goes disproportionately to the most expensive projects, with more than half of the funding going to the most expensive 2% of projects. These are typically museums, classical music groups, and performing arts centers. When donors prioritize giving to communities they already have ties to, the money stays in richer communities. Some areas have way more rich people than others. New York City has 119 billionaires; most African countries have none. Unsurprisingly, Schill