1996-07-20 - Game Theory and its Relevance to Cypherpunks

Header Data

From: tcmay@got.net (Timothy C. May)
To: cypherpunks@toad.com
Message Hash: 2d6fa04f2192f0e89aba49da995ecae7d2a6819b38f6857b53923f8029463af3
Message ID: <ae168dd4030210043a7a@[205.199.118.202]>
Reply To: N/A
UTC Datetime: 1996-07-20 23:27:10 UTC
Raw Date: Sun, 21 Jul 1996 07:27:10 +0800

Raw message

From: tcmay@got.net (Timothy C. May)
Date: Sun, 21 Jul 1996 07:27:10 +0800
To: cypherpunks@toad.com
Subject: Game Theory and its Relevance to Cypherpunks
Message-ID: <ae168dd4030210043a7a@[205.199.118.202]>
MIME-Version: 1.0
Content-Type: text/plain



Game theory is terribly important to Cypherpunks.

At 8:11 PM 7/20/96, Perry E. Metzger wrote:
>Arun Mehta writes:
>> >It may sound "stupid" to you...I suggest you read up on evolutionary game
>> >theory. Sometimes one has no choice but to respond to an arms buildup.
>> >Unilateral disarmament rarely works.
>>
>> There are surely alternatives to the extremes of unilateral
>> disarmament and an arms race? For instance, slowing down  the
>> race?
>
>Don't ask the question as though it is theoretical. Don't try
>answering it as though your personal values have any bearing, because
>there is indeed an objective answer here. Try doing some game theory
>simulations and see how well unilateral disarmament works.
>
>Oh, and don't give us stuff about how humans are above evolutionary
>pressures or nonsense like that, because we aren't any more above such
>pressures than we are above the laws of physics.


Moreover, folks should pay careful attention to the way the _appearance_ of
strength is critical. I don't mean bluffing, thought that sometimes has a
place, I mean the "scaring off" of would-be
challengers/attackers/intruders.

For example, one well-known variant in evolutionary game theory is the
"game of chicken," immortalized in the race-to-the-cliff in "Rebel Without
a Clue^H^H^H^H^HCause." While such games sound foolish to many--and such a
cliff race is not something I'd ever partake in--they are quite common and
confer evolutionary advantages to the winners.

To put it simply, the leader of a pack, whether dogs, humans, whatever, may
face challenges from other pack members. If he can convince them that he'll
win, that he's the craziest motherfucker on the planet, the challenges are
reduced. A weak-appearing leader is of course attacked more quickly. (These
are separate issues from the actual strength of skill of the participants,
but it's a basic fact that a leader who can avoid as many challenges as
possible will likely last longer.)

Seen another way, there is considerable game-theoretic advantage in being
seen as "crazy." If the other guy veers off long before impact or going
over the cliff, because he think his opponent is crazy enough to ignore
"rationality," much is gained by the victor.

Avoiding fighting is often the most important consideration, and arms races
often accomplish this goal! Even the "MAD" policy of "mutually assured
destruction" has game-theoretic justification. (Indeed, this is virtually a
truism, given the role game theory and the RAND Corporation played in the
devising of the MAD strategy.) While humanists and liberals may cluck at
the admittedly horrible consequences of MAD, were it ever implemented, it
is solidly grounded in these "games." Fortunately, the goal of MAD was to
not have to be used, and it appears now to have worked quite well (albeit
at high cost).

The application of these ideas to gun ownership is pretty clear, even to
those who have not studied or thought much about these topics: a person
contemplating a crime will be more likely to do so with an unarmed person.
Duh. And less likely against someone he suspects is armed. Still less
likely againt a "gun nut." Seeing an "NRA" sticker in the window of a car
parked outside is likely to make him think twice, fearing a "gun nut" is
inside. Being perceived as a gun nut does have some advantages. Hence the
availabilty of signs saying: "Trespassers will be Shot," and, my favorite:
"I have a .45 and a shovel; I doubt you'll be missed." (5 points to
whomever first identifies the movie this was in)

(One of the interesting speculations is what role this thinking played in
the development of U.S. paramilitary S.W.A.T. teams, the "black-clad
ninjas" we have been talking about recently. Creating a terrifying image,
an image of crazed indifference, is a useful thing. Certainly the S.S.
understood the power of their frightening uniforms and the "myth" of their
bloodthirstiness. (Note: I am not saying their bloodthirstiness was a myth,
but that they deliberately cultivated this image. Intimidation works,
game-theoretically.)

Game theory is at the confluence of economics (costs), psychology
(motivations, rewards), evolution (who survives to reproduce), sociobiology
(essentially another name for evolutionary game theory), and other fields.
All educated persons should know the basics of a bunch of related things:

-- "the prisoner's dilemma" (another famous example)

-- the iterated forms of this and other games

-- the concept of payoff matrices

-- the game of chicken

-- the nature of arms races

-- random reinforcement

-- the concept of "defection"

-- the ideas of positive-, zero-, and negative-sum games.

More advanced stuff, e.g, Nash equilibria, is real useful to know about,
but is not part of any ordinary conversations I have seen on the Net in
many years of participating.

In particular, game theory has a lot of usefullness to crypto and
Cypherpunk themes. "All security is economics." Lots of obvious
connections. Most Alice-Bob situations are essentially multi-party games,
with all the related stuff about spoofing, expectations, reputations,
cheating, etc. (Granted, many of these situations are far beyond the formal
games that have been analyzed in detail to date. That is, it is not clear
how crytpographic protocols affect such games; or at least this is a
fertile area for further study.)

A good understanding of game theory can be gotten in a few weeks by reading
some of the main introductory texts. This will not qualify oneself as an
expert, but will provide one with insight into why many of us make the
points we make about economics, self-reliance, payoff outcomes, and so on.

Useful sources:

Axelrod, "The Evolution of Cooperation" --> the best place to start, IMO

Hofstadter, various chapters in "Metamagical Themas"

Poundstone, "The Prisoner's Dilemma"

Kahn, "On Thermonuclear War"

And of course various textbooks on game theory. (But read the popular
accounts first, else the mathematical rigor will be meaningless and
confusing.)

The usefullness to cryptography, especially to the more outre stuff we are
most interested in, has largely remained unexplored. I get the impression
that few academic cryptographers know much about it. I predict that
important insights will come as these fields come into more contact.

--Tim May

Boycott "Big Brother Inside" software!
We got computers, we're tapping phone lines, we know that that ain't allowed.
---------:---------:---------:---------:---------:---------:---------:----
Timothy C. May              | Crypto Anarchy: encryption, digital money,
tcmay@got.net  408-728-0152 | anonymous networks, digital pseudonyms, zero
W.A.S.T.E.: Corralitos, CA  | knowledge, reputations, information markets,
Licensed Ontologist         | black markets, collapse of governments.
"National borders aren't even speed bumps on the information superhighway."









Thread