Game Theory, Nuclear War and Nobel Prizes

[welsh]

This is kind of interesting for some folks. Game Theory has been used to help predict what would happen in nuclear war and Thomas Shelling was a big name. Whether he deserves the Nobel for his work, I am not sure. Still, for those interested in how conflicts evolve, this might be interesting. You might want to check his book which does a good job of discussing deterrence and compellence strategies.

Also of interest here, Shelling supposedly was one of the advisors during the Vietnam War, suggesting how much bombing should be done to compel the North Vietnamese to come back to the peace negotiations.

Peace prize?

Economics focus

War games

Oct 13th 2005
From The Economist print edition

A big pay-off for two game theorists

THIS year's Nobel prize for economics might almost have doubled as the prize for peace. On October 10th, three days after the International Atomic Energy Agency and its director-general, Mohamed ElBaradei, won their laurels for monitoring the misuse of nuclear power, the economics prize was bestowed on two scholars whose best work was also done in the shadow of the mushroom cloud.

Robert Aumann, of Hebrew University, and Thomas Schelling, of the University of Maryland, are both game theorists. Game theory is now part of every economist's toolkit and has been recognised by the Nobel award before, when John Harsanyi, John Nash and Reinhard Selten shared the honour in 1994. It is the study of what happens when the calculating, self-interested protagonist of economic fable meets another member of his kind. In such encounters, neither party can decide what to do without taking into account the actions of the other.

Therefore the key of game theory is anticipating what rational minds would do when forced to interact. Are they more likely to cooperate with each other or be antagonistic. Classic game for most people is the Prisoner's Dilemma.

http://pespmc1.vub.ac.be/PRISDIL.html

During the cold war, two protagonists that captured game theorists' imaginations were the United States and the Soviet Union. How each of these nuclear adversaries might successfully deter the other was the most pressing question hanging over Mr Schelling's classic work, “The Strategy of Conflict”, published in 1960. The book ranged freely and widely in search of an answer, finding inspiration in gun duels in the Old West, a child's game of brinkmanship with its parents, or the safety precautions of ancient despots, who made a habit of drinking from the same cup as any rival who might want to poison them.

Which is actually one of the cool aspects of the book. How does the theory work in similar circumstances.

Mr Schelling's back-of-the-envelope logic reached many striking conclusions that appeared obvious only after he had made them clear. He argued that a country's best safeguard against nuclear war was to protect its weapons, not its people. A country that thinks it can withstand a nuclear war is more likely to start one. Better to show your enemy you can hit back after a strike, than to show him you can survive one. Mr Schelling invested his hopes for peace not in arms reductions or fall-out shelters but in preserving the ability to retaliate, for example by putting missiles into submarines.

Therefore the notion of second strike capacity matters more, but note how this also suggests that countries will arms race.

All-out thermonuclear warfare is the kind of game you get to play only once. Other games, however, are replayed again and again. It is these that fascinate Mr Aumann. In a repeated encounter, one player can always punish the other for something he did in the past. The prospect of vengeful retaliation, Mr Aumann showed, opens up many opportunities for amicable co-operation. One player will collaborate with another only because he knows that if he is cheated today, he can punish the cheat tomorrow.

Those interested in the application of Game Theory might also look to Robert Axelrod's Evolution of Cooperation. Game Theory has also been used to help explain the origins of constitutional rules, property rights and even relationships.

Mutually assured co-operation
According to this view, co-operation need not rely on goodwill, good faith, or an outside referee. It can emerge out of nothing more than the cold calculation of self-interest. This is in many ways a hopeful result: opportunists can hold each other in check. Mr Aumann named this insight the “folk theorem” (like many folk songs, the theorem has no original author, though many have embellished it). In 1959, he generalised it to games between many players, some of whom might gang up on the rest.

WHich is perhaps the virtue- it's not good morals that lead to cooperation, but selfishness. Morality is uncertain- some folks are more than others. But if people are placed in a situation where their self- interest leads to cooperate, then you've got something.

Mr Aumann is loyal to a method—game theory—not to the subject matter of economics per se. His primary affiliation is to his university's delightfully named Centre for Rationality, not its economics department. Trained as a mathematician, he started out as a purist—pursuing maths for maths' sake—but soon found his work pressed into practical use. Between 1965 and 1968, for example, he co-wrote a series of reports for the United States Arms Control and Disarmament Agency. The Russians and Americans were pursuing gradual, step-by-step disarmament. But the military capabilities of each superpower were so shrouded in mystery that neither side knew precisely what game they were playing: they did not know what their opponents were prepared to sacrifice, nor what they themselves stood to gain. Without knowing how many missiles the Russians had, for example, the Americans could not know whether an agreement to scrap 100 of them was meaningful or not.

A side issue then is to make sure each side knows what the other side has. There are weapon systems that stabilize and destabilizing systems.

For example having bombers might mean that you have a second strike potential that the enemy knows. But having a bomber that can slip into your airspace undetected might be destabilizing as it can take out your weapons system before you have a chance to use it.

In such games, Mr Aumann pointed out, how a player acts can reveal what he knows. If Russia were quick to agree to cut 100 missiles, it might suggest its missile stocks were larger than the Americans had guessed. Or perhaps the Russians just wanted the Americans to think that. Examples of such deception are not limited to the cold war. Some have speculated that Saddam Hussein pursued a similar strategy in the run-up to the invasion of Iraq in 2003. Despite apparently having no actual weapons of mass destruction left, he offered only the most grudging co-operation to weapons inspectors. The Iraqi dictator perhaps wanted to conceal the humiliating fact that he had nothing much to hide.

The trick, therefore is not the game, but how you play it?

Messrs Aumann and Schelling have never worked together, perhaps because the division of labour between them is so clear. Mr Aumann is happiest proving theorems; Mr Schelling delights in applying them. Mr Aumann operates at the highest levels of abstraction, where the air is thin but the views are panoramic. Mr Schelling tills the lower-lying valleys, discovering the most fertile fields of application and plucking the juiciest examples.

In one of his more unusual papers, Mr Aumann uses game theory to shed light on an obscure passage in the Talmud, which explains how to divide an asset, such as a fine garment, between competing claimants. You should give three-quarters to the person who claims all of it, and the remaining quarter to the person who claims half of it, the text instructs, somewhat inscrutably. Fortunately, the Nobel committee had no need for such a complicated rule in dividing up its prize. Between its two equally deserving laureates, it split the SKr10m ($1.3m) fifty-fifty.

Cool, $1.3 mil. Sweet!

 

[El Chupacabra]

Another game theorist, Anatol Rapoport, is known for his critics of the aggressive strategies in actions based on the game theory and popularization of the cooperative ones (books like Fights, Games, and Debates (1960) or Strategy and Conscience (1964)). He recalls in the latter some of the scary calculations that were made during the game theory-based simulations of an atomic conflict. The strategists wondered, for example, what would be the maximum acceptable human casualties count worth a victory. One of them wrote something like:

If 180 million (whole population of the U.S.A.) would be too much of a price to pay to punish the aggressor, what price would we be willing to pay?... I consulted the problem with many Americans and after about 15 minute-discussion the declared accectable bodycount ranged between 10 and 60 million, but in most of the cases, closer to the latter.

 

[SimpleMinded]

Another game theorist, Anatol Rapoport, is known for his critics of the aggressive strategies in actions based on the game theory and popularization of the cooperative ones (books like Fights, Games, and Debates (1960) or Strategy and Conscience (1964)). He recalls in the latter some of the scary calculations that were made during the game theory-based simulations of an atomic conflict. The strategists wondered, for example, what would be the maximum acceptable human casualties count worth a victory. One of them wrote something like:

If 180 million (whole population of the U.S.A.) would be too much of a price to pay to punish the aggressor, what price would we be willing to pay?... I consulted the problem with many Americans and after about 15 minute-discussion the declared accectable bodycount ranged between 10 and 60 million, but in most of the cases, closer to the latter.

you can kill 179,999,999 americans so long as I get to be the one that stays alive .

Interesting info Welsh. I'm interested in reading more on Game Theory now.

 

[4too]

Pearls Before Swine

Pearls Before Swine





All this "And Or Nand Nor" logic grinding has an appeal that transcends the academic. Yet any wisdom gleaned from these constructs can be easily trivialized by the selfish, short term gain, self interest. Immediate gratification is the most likely choice of the 'rich and famous' of any age.


Macnamara 'confessed' that the force level necessary to succeed in Vietnam was politically, ... inappropriate. So that campaign became a sally in the great siege of Cold War One. Any 'Game Theory' applied was to get the most 'bang for the buck' when bombing the Northern Communists into the, pre Marxian, Stone Age.


After the bloom has faded, ""Shock And Awe"" appears to be a mere sales slogan and as hollow as ""Mission Accomplished"". Smoke and mirrors is no substitute for 'boots on the ground''. Declaring a ... ''Liberation'' in Iraq with out forging a foundation of true security could create as shallow a victory as the Cold War One coupe that placed the Shaw in power in Iran. What kind of draconian compromise was dictated when cooperating with the military that was pointing to the '"'de je vu, all over again"' of another campaign featuring 'insufficient forces'.

On the economic front, the stock market, that 'once and future' refuge for all private and public pension schemes, appears to be less about cooperation and more of a circus side show with cooperate carnival barkers shilling the next, quarterly, 'quick kill'. Easy money.

One can game out all the virtue one wants, and bang the tambourine of enlightened self interest. Those that comply with our know animal natures, will cross examine redemption with the time honored expedient of: "'If you can't eat it, drink it, or f'ck it, ... p'ss on it.""

Pearls before swine.



W.:

... WHich is perhaps the virtue- it's not good morals that lead to cooperation, but selfishness. Morality is uncertain- ...

""Saint"' Machiavelli advised his little prince that 'love' was too fickle an emotion to illicit in the governing of the state. Fear was the sure thing.

Think of all the fear and anxiety coloring all commercial and political communication in this modern age. Humbuggery of hooks. Products galore are marshaled to protect our threatened sex appeal. Political platforms are based more on fear of some "'other"' [ your personal phobia inserted here] and less on constructive cooperation.

The schematic progressions of Game Theory seem to be incorporated into social logic trees, only if it bears a predetermined fruit. No matter how strange, fear is the most familiar fruit. The forbidden fruit of knowledge has disastrous side effects. We risk our Eden like animal innocence, the cold comfort of conformity: The fight or flight certainty of a fear actuated life style.

Game Theory and wisdom,
Pearls before swine.




4too

 

[welsh]

Well Simpleminded, it is very interesting stuff.

A good read in Yussof Cohen's book Radicals, reformers, and reactionaries : the prisoner's dilemma and the collapse of democracy in Latin America- a very readable book on game theory.

More complicated might be The Rebel's Dilemma and the Cooperator's Dilemma.

Axelrod's discussion of cooperation using Tit-for-Tat strategey is pretty classic.

Brian Skyrms wrote a few interesting pieces on - The stag hunt and the evolution of social structure, and Evolution of the social contract

There's a good book on cooperation and strategic interaction in the Civil Rights Movement, but I can't remember the name. (edit- Dennis Chong- Collective Action and the Civil Rights Movement).

A friend of mine, Charles Kromkowski used game theory as part of his explanation for the Revolutionary War, the Civil War and another crisis in between- Recreating the American republic : rules of apportionment, constitutional change, and American political development, 1700-1870.

If you are interested in history and the use of game theory, you might want to look at Analytical Narratives, edited by Bates and others.

When I studied game theory under Charlie Holt he used to say that the interesting thing about game theory is not necessarily what it predicts but trying to understand why people don't play the game like you thought you could. The basis is rational action, but people are sometimes less rational than we think, or hope for. It might be the game is less perfect or people are more complex than rational choice theorists would think.

Thus the real problem of game theory, it can predict, but it cannot foretell.