The Strategy of Conflict - Thomas Schelling - ****

There are two schools of thought regarding conflict. One school believes that conflict is bad and tries to eliminate it. The other school believes that conflict is inevitable and tries to win.

Most conflict situations are essentially bargaining situations.

Using an ultimatum as deterrence uses different skills than the carrying out of that ultimatum.

As far as deterrence goes, the only way a threat can be "too large" is if it violates suspension of disbelief with its severity - effective deterrent threats are never carried out.
-- That's kind of cool, but I disagree with it. In a situation where the effectiveness of a solution is fixed, you ideally want to solve the problem with the least amount of coercion possible. This makes for better relations and doesn't give your opponents the ability to develop defenses against your stronger weapons.

Being unable to appeal to a higher authority hurts negotiations.

Conflict theory is based on rationality, but irrationality can be a strategic advantage.
The threat of mutual destruction cannot be used to deter an adversary too unintelligent to comprehend it or too weak to enforce his will on those he represents.

We tend to identify peace, stability, and conflict resolution with notions like trust, good faith, and mutual respect. To the extent that this view encourages trust and respect, it's good. When trust and good faith don't exist, and when these qualities can't be cultivated by our pretending they exist, we need to look elsewhere for peaceful solutions.

-- in Chapter 2, the book specifically states that it isn't going to concern itself with mutually advantageous bargaining: instead, it'll focus solely on situations where good for one means bad for the other.
-- that seems like a limited view. A rose is probably not going to be the first thing an expert thistle-spotter sees.

Binding oneself fixes your position, but doing so surrenders control to the other party.
The first person to lock himself into a position establishes the precedent.
If you can demonstrate that the other party isn't commited, or that they've miscalculated their commitment, you may undo or revise their commitment.

Restricting the agenda: saying "these are the only things we discuss" to prevent extracurricular extortion (and non-essential issues)

Imagine there are two projects, each with a cost of 3. Each project is worth 2 to Player A and 4 to Player B. Player A has no incentive to do either of these projects by himself. If Player B wants to get Player A's help, he needs to offer to split the costs - giving Player A a net gain of 1 vs. Player B's net gain of 5.
-- note: if Player A and Player B were directly competing with each other, it would probably be unwise for Player A to double Player B's potential growth (if Player B did both projects by himself, he'd be at +2 instead of +4)

An important limitation of economic problems is that they tend to involve divisible objects and compensable activities. If a drainage ditch that would protect two houses costs $1000, and it's worth $800 to both homeowners, we assume that if they both knew the score, they'd get together and pay the $1000 jointly. But if it costs 10 hours a week to be scoutmaster, and each considers the job worth 8 hours of his time but one man must do the whole job, we can't assume that one of the neighbors will propose a deal to compensate the other for those extra hours.

When you bind yourself to a principle, it looks like you must either choose stalemate or discredit that principle.
-- my way to counter principled negotiations is to convince the party that either:
-- (a.) they shouldn't be applying the principle to this particular case, or
-- (b.) my offers aren't inconsistent with their principles

Last clear chance: a legal doctrine that recognizes that before an event occurs, there are points at which some of the parties can no longer affect it.

The best defense against a threat is to carry out your actions before the threat is ever made. If you can't hasten your actions, you can commit yourself to them to the point where the threat would be unable to alter your course. Finally, the threat has to be communicated - if you make yourself unavailable for messages, you can deter the threat itself.
-- that is unbelievable

If a threat can be broken down into a series of smaller threats, there's an opportunity to demonstrate on the first few transgressions that the threat will be carried out on the rest.
This tactic of decomposition applies to promises as well as threats. If neither party trusts the other on large issues, they might be willing to trust each other on small ones.

What makes many agreements enforceable is only the recognition of future opportunities that will be eliminated if mutual trust is not created and maintained, and that the value of these opportunities outweighs the gains that would come from cheating in the present instance.
-- you can manipulate this dynamic by structuring situations so people have nothing to gain by cheating you

When people can't communicate but need to cooperate, they do so through a common language of equitability, selecting what they believe to be logical points of agreement.

Since agreements typically default to "logical" values, it's important to set the stage in such a way as to give favorable outcomes prominence.

"The initial departure of retaliation from the locality that provokes it may be a kind of declaration of independence that is not conducive to the creation of stable mutual expectations."
-- if your relative weakness in an assessment based on stable expectations is what keeps your opponents from fearing your retaliation, this kind of declaration is the only immediate way to achieve meaningful deterrence.

-- Any bargaining situation can quickly evolve into a hierarchy of extracurricular threats and promises if there are no limits regarding what can be put on the table. Any zero-sum situation you see in reality can be transformed away from its basic dynamic with just a few words.

Mobs evolved away from leadership because leadership could be identified and eliminated: now, "incidents" are the coordinating entities.

Tipping points are the result of communication systems that make it easy to take one course of action and virtually impossible to agree on other solutions.

Coordination isn't about guessing what the other person will do: it's about guessing what he thinks you'll do based on what you'll think he'll do, ad infinitum.

-- giving people a prize for successfully choosing the options they think others will find most attractive is an unbelievably effective way to determine mass appeal

If your neighbor's fruit tree overhangs your yard and you pick all the fruit on your side of the line, he has a pretty good idea of what he's acquiesced to in future if he doesn't retaliate. If, instead, you pick the same amount of fruit randomly from both sides, he won't really know how to react. Was it you who took the fruit? Is he going to have to stop you, or are you not going to take any more?

Game theory needs to take into account communication and enforcement systems to be applicable in real-world situations.

If twenty men are held up for robbery by one man who has six bullets, they can overwhelm him provided they're willing to lose six of themselves. The gunman must keep this kind of agreement from being formed: he must sever their communications and threaten any man from being first.

Prohibition delivers enormous power into the hands of the people who can enforce otherwise-unenforcable contracts.

"The ordinary high-school graduate has to work hard for his money, but he could destroy a hundred times that much if he set his mind to it. Given an institutional arrangement in which he could abstain from destruction for a mere fraction of the value that he might have destroyed, the boy clearly has a calling as an extortionist rather than as a mechanic or clerk. It is fortunate that extortion usually defends on self-identification and overt communication by the extortionist himself."
-- bahaha

The timely destruction of communication may be a winning tactic. When a husband and wife are arguing on the phone about where to have dinner, the argument is won by the wife if she announces where she's going and hangs up.
-- that's not a winning tactic unless the need to collaborate outweighs the desire to prevent the "winner" from reaping the rewards of dysfunctionality. Even if it succeeds, it damages the "winner's" reputation: people recognize that there's not a lot of potential for a collaborative relationship with someone whose idea of negotiation is to get the last word in, destroy all communication, then wait for the other guy to acquiesce.
-- the best use of this tactic is temporary interruption of communication, which takes people out of the moment and gives them time to think about the preceding negotiations. The only time you ever want to be the "hang-up-the-phone" guy is when someone's trying to steamroll over your final offer: there's no discussion taking place at that point.


Children are skilled at avoiding the warning glance from their parents, knowing that if they perceive it the parent is obliged to punish noncompliance. Adults are equally skilled at not requesting the permission they suspect would be denied, knowing that explicit denial is a sterner sanction that obliges the denying authorities to take cognizance of the transgression.
-- the "better to seek forgiveness than beg permission" douchebags

In zero-sum games, the objective is not to be found out. In mixed-motive games, the objective is to make the other party take your behavior for granted.

Those who played uncooperatively against a cooperative partner had an opportunity, on the second play, to respond to the implicit offer of cooperation. Instead of thinking of the act as a cooperative one, however, they tended to assume that the cooperative player either didn't care or didn't quite understand the game.
-- there's something to be learned from this.

A promise's cost is paid when it succeeds, and a threat's cost is paid when it fails. If you promise more than you need to, you pay more than you need to. If you threaten more than you need to, the excess is superfluous - a successful threat is one that's not carried out.
-- again, I would stay away from threatening more than is necessary... they go into some math that agrees

You can't threaten that you MIGHT do something, because that's equivalent to saying that you MIGHT NOT do it - it shows that you're not commited. Furthermore, if you don't carry out a threat that you MIGHT have commited, you're only reinforcing your opponent's belief that when it's put up or shut up you'll shut up.
The only way to threaten that you MIGHT do something is if the final decision to carry out that threat is not entirely under your control.

Brinksmanship is the ability to push your opponent to the point where if you go, he goes with you.

If surprise has an advantage, it makes sense to be the first person to gain that advantage.
The perceived likelihood of being attacked first is a large factor in one's willingness to attack.
-- ideally, you want to structure situations so that surprise attacks confer no advantage: barring that, you want their advantage to be outweighed by disadvantages.

If you have the ability to strike with impunity, your opponent will factor that into the likelihood that you'll attack first and be accordingly more likely to attack you because of it.
An example is submarines. As things stand, submarines can't be stopped from launching their nukes. If technology came out that made it possible to detect and eliminate every sub before its nukes were launched, the situation would be substantially less stable [thanks to the increased effectiveness of surprise attacks]. The U.S. knows this, but we have to research submarine detection methods anyway - we can't afford to be without detection when someone else has it.

It's important that your opponent is not forced to guess whether or not you'll attack him if you have no intention of making a surprise attack. During the Cold War, a general remarked, "Our reluctance to strike first is in fact a military disadvantage to us; but it is also, paradoxially, a factor in preventing a world conflict today."
-- earlier in the book, the author recounts someone's proposal not to execute the Russian spies who represented Russia's only way of independently verifying that we were telling the truth about this

-- when you launch a surprise attack, the goal should be to obliterate the opponent's capacity for reprisal.

An assessment of defensive measures comes out differently if we put primary reliance on deterrence. Chicago cannot be hidden. Similarly, a defense of Chicago that requires the enemy to triple the size of his attack may be a poor prospect; it may mean only that the enemy must invest in a larger initial attack.
-- this is something I know from RTS. If you're building static defenses, odds are you're probably losing.

The ideal limitation for a situation in which people can't be stopped from cheating is high enough to negate the advantage of cheating.

If the enemy can hit anything he can locate, and kill anything he can hit, he has to be made unable to locate it.

Examine the sustainability of measures for increasing alertness. If they're not sustainable, they're not effective solutions against situations where the enemy can attack at any time.

You cannot prove that you won't do something. You can only prove that you can't.
Compliance must be observable, verifiable, and provable.

Tenants are less easily removed by threat of forcible eviction than by shutting off the utilities.


J.M. Keynes, The General Theory of Employment, Interest, and Money

The Art of Coercion