View Full Version : Game Theory
Rob Oplawar
November 18th, 2008, 04:36 PM
No not videogames;
http://en.wikipedia.org/wiki/Game_theory
Do you guys wanna have a quick discussion about game theory?
A good example in popular culture is the climax to The Dark Knight:
Two boats have the ability to destroy the other thus ensuring its own safety; if neither destroys the other there is a possibility of either being rescued or both boats being destroyed.
The only rational decision here is to immediately destroy the other boat, as it is the only way that guarantees survival.
I was just talking to snaf about the concept of a relativistic bullet. The idea is, with the amount of energy it takes to power a planet you could accelerate a small (read: truck-sized) projectile to very nearly the speed of light. A projectile such as this would be virtually undetectable and even if you could see it coming it would be impossible to get out of the way in time.
A projectile of that sort of mass moving close to the speed of light would have enough energy to effectively kill all life on a planet by sheer force of impact.
Some extremely paranoid people offer this as a doomsday scenario, with some pretty large assumptions:
Assuming every civilization will continue to advance technologically, AND
Assuming it is possible to produce a relativistic bullet,
THEN eventually every civilization that survives long enough will have the ability to produce a relativistic bullet, AND
Every civilization that lives long enough will have the ability to destroy any other civilization.
This being the case, there are two rational choices:
Hide so that no other civilization ever knows you're there, although this is only a temporary solution, as by a similar argument you could say that you will eventually be found, OR
Immediately destroy any civilization you encounter with your own relativistic bullet.
There are no options which guarantee survival, but preventing other civilizations from destroying you by the only means possible gives the best chance of survival.
So, if you accept this argument, you'd expect any sufficiently advanced race we encounter to immediately destroy us to protect themselves.
oooOOOOOooo!
edit: There are actually some pretty big holes in this argument; +rep to the first to point them out
Discuss (not just the relativistic bullet, but game theory in general).
ima_from_America
November 18th, 2008, 07:45 PM
:tinfoil:
Sounds like a type 1 on the Kardashev scale.
Anton
November 18th, 2008, 07:54 PM
If you think of it as if everyone was irrational and automatically assumed their lives were at risk then yes this theory would work..maybe.
But not everyone is irrational and sometimes in order to grow, we must seek others.
I don't know what the hell I'm talking about. :)
Rentafence
November 18th, 2008, 08:19 PM
Who says that each party is going to feel like the other is hostile against them in the first place? Both parties would surely want to compromise with each other instead of blowing the shit out of each other. Like the US and Russia in the cold war, both had nukes pointed at each other but where always reluctant to fire because of the mutual destruction that would ensue.
E: Any advanced life that would encounter a less advanced species wouldn't feel threatened by them, so the only time this could apply would be two equally advanced races, but they would forever be at a standoff because they both have the ability to destroy each other.
Phopojijo
November 18th, 2008, 08:30 PM
Who says that each party is going to feel like the other is hostile against them in the first place? Both parties would surely want to compromise with each other instead of blowing the shit out of each other. Like the US and Russia in the cold war, both had nukes pointed at each other but where always reluctant to fire because of the mutual destruction that would ensue.
E: Any advanced life that would encounter a less advanced species wouldn't feel threatened by them, so the only time this could apply would be two equally advanced races, but they would forever be at a standoff because they both have the ability to destroy each other.
In both of those cases you assume that the opposing yet equal player will be able to counter-move before your move is complete.
In the US vs Russia event... counter-nukes can fly before the first nukes land.
In the advanced races thing... you assume that neither race can completely and utterly decimate the other before a possible counter-attack.
Game Theory gets to be complicated... hence how John Nash was schizophrenic.
Rentafence
November 18th, 2008, 08:33 PM
But if 2 races were so advanced that they could propel something with the mass of a truck at light speed, they must be colonizing other planets and exploring space. In which case the attack by Civilization A can be observed by Civilization B because they aren't all on one planet, thus counter attacking. Or maybe I should just go with the idea that one player can destroy the other before they know what the fuck happened.
A strange game. The only winning move is not to play. :aaaaa:
Hotrod
November 18th, 2008, 08:58 PM
If the other races out there were hostile, AND they had a weapon like that, AND they had the ability to come over here, why aren't we all dead?
If the bullet went that fast, why wouldn't it disintegrate? The forces acting against it would be so big, they would destroy the bullet instantly, unless it had some sort of protection.
Rob Oplawar
November 18th, 2008, 10:15 PM
A strange game. The only winning move is not to play.
wtc, I can't spread rep. I guess there are a total of like 5 people on this forum that I ever rep, so it's always "you must spread some rep around first"
What makes it interesting is that there is no mutual assured destruction in the simple case. Moving very near the speed of light, you would be physically incapable of detecting it pretty much until it impacted.
I'm having trouble finding a link to where I first saw this problem- I think the term "relativistic bullet" is incorrect. Can anybody else find a more formal description?
Zeph
November 18th, 2008, 10:34 PM
Why would you want to use all the energy to accelerate something like that to the speed of light. Just toss a moon at them. It uses much less energy and has the same effect minus relative surprise (IE, you can dodge a moon). Or why not just lob some gamma rays at them?
Rob Oplawar
November 18th, 2008, 11:58 PM
The paranoid people have arguments in favor of fearing imminent destruction, but in my case, it's more of an interesting thought experiment, a hypothetical situation.
It gets more interesting when the concept of reputation is involved; in the case of the cold war, even if we were able to strike quickly enough to completely obliterate Russia before they could retaliate, where would such an action put us in the global community?
Or in the case of a relativistic bullet, if other civilizations could observe the fact that you had destroyed a planet, would that cause them to find you to be a threat and directly target you?
I personally think the goal of human civilization should be to spread as rapidly as possible throughout the galaxy, to prevent the possibility of total extermination. That being the case, it would then make sense to assume that other species would have similar goals. In that case the game becomes boringly simple, because cooperation is pretty obviously the rational choice upon encountering another race.
Anybody know of any other interesting applications of game theory?
Kornman00
November 19th, 2008, 12:44 AM
I refuse to play :bushcolbert:
f4WkvC8jMb4
Sever
November 19th, 2008, 12:45 AM
...the goal... [is] to spread as rapidly as possible...Isn't that the goal of all life? Game theory can be deduced from even the most basic micro-ecosystems - organisms are instinctively compelled to reproduce and spread, to increase their survival as a whole.
Rob Oplawar
November 19th, 2008, 02:57 PM
It's a self-reinforcing system. By that logic you would (and I do) expect that eventually the galaxy will be populated by a single race, having wiped out all competition, assuming the galaxy is teeming with intelligent life.
To reiterate: by definition, the forms of life that are most capable of surviving and reproducing will be the most common, and will tend to push out other forms of life to make room for themselves. Those that do this aggressively will tend to be most common by definition, because they will deliberately eliminate others so that they can spread further. So either every species in the galaxy will be willing to limit its own growth in the interest of cohabiting with every other species (unlikely) or at least one species will continue to aggressively expand until it controls the galaxy.
Of course, the concept of community and society causes this argument to break down somewhat, because the galactic community might collectively prevent such spread, forming a superspecies intent on aggressively protecting the survival of the whole.
paladin
November 19th, 2008, 04:15 PM
Its funny, our guest speaker in class used the Dark Knight situation while explaining this very aspect of games. It really was interesting.
Rob Oplawar
November 19th, 2008, 04:28 PM
You might apply game theory to a curved exam in a class:
Options: don't study, get a mediocre score on the exam.
Study, get a good score on the exam.
The best option for the community is for everybody to not study (and instead use the time to have fun), so that the average score is low, and so with the curve the majority of people get a good grade.
Of course, the best option for the individual is to study so that they get a good score on the exam; this guarantees them a good grade (otherwise, they might score low when others scored high, and so with the curve they get a bad grade), but hurts the community, forcing everybody to either also study or suffer a bad grade.
:-3
edit: This is a kind of interesting chicken and egg problem- Assuming that all or most other students will make a rational decision and that rational decision is to study and score high, then the rational decision is to study and score high. But if the rational decision is to not study then the students can be expected to score low and the rational decision is to not study. Lol. So it comes down to how you weight the benefit of not studying (and enjoying the time) vs getting a good grade. I think in almost all cases (assuming all students were equal) it would come to an equilibrium on either study or don't study, but there might be a weighting that causes the unbalanced state I described above where the rational decision depends on what the rational decision is.
Of course, when other players (ie students) start weighting the benefit of free time vs good grades differently, things get really complicated.
Dwood
November 19th, 2008, 05:15 PM
Orson Scott Card did this "Game theory" thing in his book, Ender's Game.
When you fight an enemy that has hurt you, or wants to hurt you, then what you do is hurt them hard enough that they can't hurt you any longer.
And that was how Ender won. :3
* won't give the ending away *
Abdurahman
November 19th, 2008, 05:34 PM
Wasn't there a lot of use of the game theory in the SAW movies? Correct me if i'm wrong, but I never watched the movies, so I don't know exactly.
Powered by vBulletin® Version 4.2.5 Copyright © 2024 vBulletin Solutions Inc. All rights reserved.