Originally posted by KazetNagorraFrom the point of view of the individual, the "average" punishment for someone who stays silent is (10 years + 0.5 years)/2 = 5.25 years. This is because he doesn't know what his "partner" will do, and so he has to assume the outcomes are equally likely.
No, if both defect both get 5 years leading to a collective punishment of 10 years, while if both remain silent both get 6 months leading to a collective punishment of 1 year. The lowest combined punishment is achieved through both remaining silent.
Whether or not they are guilty is not really relevant to the prisoner's dilemma since it's intended as a metaphor for a certain mathematical model describing decision making.
The average for an indivudal who defects is (5 years + 0 years) = 2.5 years.
In addition, there is no scenario where the expected punishment if you stay silent is less than if you defect (I think this makes defecting a 'strictly dominating' strategy; although I am not a game theory expert).
So individuals maximizing the outcome for themselves will all defect.
But as you say, the 'collective' punishment (or the sum of all punishments) under all possible combinations is 10 years (5+5, 0+10, or 10+0) UNLESS both stay silent, where the collective punishment is 1 year.
So to minimize the collective punishment, the two have to simultaneously act against their own self-interest.
Actually, this could probably be adapted to is a nice illustration of why communism doesn't work outside of monestraries (where the option of acting in your own self interest is taken away through religion). Communism also requires a large group to collectively act against their individual self-interest in order to maximize the good of the whole. And note that 'the good of the whole', as in this case, certainly results in less for some than they would have otherwise (a la Warren Buffet).
Originally posted by SeitseAfter due consideration, I think 'bozo' is most apt:
Wajoma is not a clown, he's a retard, which is different.
Please be more careful next time you go name calling, boy ðŸ˜
bozo
"muscular low-I.Q. male," c.1910, perhaps from Sp. bozal, used in slave trade and to mean "one who speaks Spanish poorly." Bozo the clown was created 1940 at Capitol Records as the voice in a series of story-telling records for children ["Wall Street Journal," Oct. 31, 1983].
http://www.etymonline.com/index.php?search=clown
Originally posted by Bosse de NageInteresting. So, what about "dunce" instead?
After due consideration, I think 'bozo' is most apt:
bozo
"muscular low-I.Q. male," c.1910, perhaps from Sp. bozal, used in slave trade and to mean "one who speaks Spanish poorly." Bozo the clown was created 1940 at Capitol Records as the voice in a series of story-telling records for children ["Wall Street Journal," Oct. 31, 1983].
http://www.etymonline.com/index.php?search=clown
http://en.wikipedia.org/wiki/Bozo_the_Clown#.22Bozo.22_as_generic_noun_or_verb
1 edit
Originally posted by telerionRepeatedly on other threads while discussing, amongst other things, socialistic health care KN has answered "prisoners dilemma" as if that all that was needed with out explaining its relevance.
[b]I know a bit about Prisoner's Dilemma problems. I think this discussion could be interesting, but it would be a lot better if you would start with a position to discuss. At this point, I don't quite see where you want to go with this.
Edit: Forget it. I've read through the rest of the thread. I see some people giving honest effort, but clearly you k ussion, quit being lazy and put something forward rather than spurning what others have done.[/b
So I invited him to expand on his non answers, and we got this: "The prisoner's dilemma implies that rational people will make decisions that are not in the best interests of the collective, and also do not provide the ideal outcome for the individual."
I use quotation marks to indicate a quote with added italics or bolding for clarification.
People are irrational so they need to be controlled...
...by...
...well what do you know they need to be controlled by the likes of KN, kmax, and a whole collection of other busybodies who, I'm guessing, feel they are more qualified for some reason.
I put the prisoners dilemma in the same criteria as the life boat situation where there is a set up with a number of "what ifs". I'll stick with reality thanks.
Edit: dang it, why did that come out all in bold, what a bozo.
3 edits
Let me give some oxygen here to Wajoma to keep this going.
In an iterated version of the prisoner's dilemma, the strong result of the one-shot version (where non-cooperation is the only possible NE) disappears. Since the world is dynamic system one can argue that the static version is a worse representation of the world.
Secondly, by Arrow's Impossibility Theorem, we know that voting systems are non-trivial and do not guarantee first-best. Stating that anarchy is not optimal and therefore a government is needed is then fallacious. Why? Because governments do not also meet the criteria that we use to reject anarchy.
Edit - Obviously the argument is different for other type of governments, but they all have their flaws and can be argued to perform worse than democratic ones.
Originally posted by spruce112358spruce: This is because he doesn't know what his "partner" will do, and so he has to assume the outcomes are equally likely.
From the point of view of the individual, the "average" punishment for someone who stays silent is (10 years + 0.5 years)/2 = 5.25 years. This is because he doesn't know what his "partner" will do, and so he has to assume the outcomes are equally likely.
The average for an indivudal who defects is (5 years + 0 years) = 2.5 years.
In addition, ther ...[text shortened]... tainly results in less for some than they would have otherwise (a la Warren Buffet).
How many times are we going to have to read this utterly fallacious argument in these forums?
Originally posted by spruce112358Actually, this could probably be adapted to is a nice illustration of why communism doesn't work
From the point of view of the individual, the "average" punishment for someone who stays silent is (10 years + 0.5 years)/2 = 5.25 years. This is because he doesn't know what his "partner" will do, and so he has to assume the outcomes are equally likely.
The average for an indivudal who defects is (5 years + 0 years) = 2.5 years.
In addition, ther ...[text shortened]... tainly results in less for some than they would have otherwise (a la Warren Buffet).
Correct.
Originally posted by PalynkaIndeed, iterated prisoner's dilemmas ensure that in close-knit, small societies, and within families, people will generally not defect because they are likely to meet again frequently. However, in the globalized economy people and businesses generally don't meet frequently.
Let me give some oxygen here to Wajoma to keep this going.
In an iterated version of the prisoner's dilemma, the strong result of the one-shot version (where non-cooperation is the only possible NE) disappears. Since the world is dynamic system one can argue that the static version is a worse representation of the world.
Secondly, by Arrow's Im ...[text shortened]... nments, but they all have their flaws and can be argued to perform worse than democratic ones.
I don't know about Arrow's Impossibility Theorem, care to elaborate?
I find the Prisoner's Dilemma interesting because it gives a much more massive penalty to the guy who kept quiet when his partner did not than it does to the pair when they keep quiet.
That is, there's more prison time involved when one keeps quiet than when two keep quiet. I wonder if this affects the analysis much.
What if the price of keeping quiet when the other guy did not was exactly twice what each experiences when both keep quiet?
3 edits
Originally posted by AThousandYoung....................Keeps Quiet ....... Confesses
I find the Prisoner's Dilemma interesting because it gives a much more massive penalty to the guy who kept quiet when his partner did not than it does to the pair when they keep quiet.
That is, there's more prison time involved when one keeps quiet than when two keep quiet. I wonder if this affects the analysis much.
What if the price of keepin ...[text shortened]... quiet when the other guy did not was exactly twice what each experiences when both keep quiet?
Keeps Quiet ...... 2,2 ................ 0,4
Confesses ......... 4,0 ............... 3,3
You mean like this ? It wouldn't make any difference to the result, confessing will always give the least amount of prison time, regardless of what the other person does.
Multiple edits to get the table looking somewhat reasonable
1 edit
Originally posted by AThousandYoungThe rational choice for the individual does indeed depend on the expected punishment/reward for each action, so if you change the numbers the rational choice may change (but the relative difference then has to change sign). It's explained here:
I find the Prisoner's Dilemma interesting because it gives a much more massive penalty to the guy who kept quiet when his partner did not than it does to the pair when they keep quiet.
That is, there's more prison time involved when one keeps quiet than when two keep quiet. I wonder if this affects the analysis much.
What if the price of keepin ...[text shortened]... quiet when the other guy did not was exactly twice what each experiences when both keep quiet?
http://en.wikipedia.org/wiki/Normal_form_game
1 edit
Originally posted by KazetNagorraI don't work directly in game theory, and it's been a few years since I took the course for my grad studies; but isn't it the case that in order for "both cooperate" to occur along the equilibrium path in one of these simple PD's, the horizon must be infinite? The problem with finitely repeated games is that they unravel through backward iteration. Generally, assume that there are T periods in which the game is repeated, and consider the T-period game. This is just a one-shot PD so both players want to play "defect." Now consider period T-1. Given that both players will defect in period T, this becomes similar to the one-shot game again. Each player will want to defect. Given that in periods T-1 and T each player will defect, they will do the same in T-2 and so on until they get to period 1, where they will defect.
Indeed, iterated prisoner's dilemmas ensure that in close-knit, small societies, and within families, people will generally not defect because they are likely to meet again frequently. However, in the globalized economy people and businesses generally don't meet frequently.
I don't know about Arrow's Impossibility Theorem, care to elaborate?
There is a NE strategy to the above game that calls for players to play cooperate. It goes something like "Cooperate if both players cooperated in the previous period, otherwise defect." Unfortunately for the prisoners, cooperate will never be played in equilibrium.
One way to get cooperation in a repeated game is if the horizon is infinite and players sufficiently value the future. In this case, the NE strategy "Cooperate if both players cooperated in the previous period, otherwise defect" can lead to cooperation on the equilibrium path. Players choose to cooperate because if players place sufficient value on future payoffs, then the present discounted value of the loss from the infinite stream of future bad equilibriums exceeds the one-period gains from defecting.
There are a lot of other equilibrium concepts that can lead to cooperation equilibria. Take a Perfect Bayesian-Nash equilibrium (PBNE). Unlike a regular NE, a PBNE adds beliefs to the game. Each player assigns to every strategy in each other players strategy space a probability weight which captures the likelihood of playing that strategy. So if you turn back to the one-period PD. Both players cooperate could be a PBNE if each player has a sufficiently high belief that the other will play cooperate. The reason is that the expected payoff (the sum of each payoff weighted by the probability of that event occuring) from playing cooperate would exceed that from playing defect.
Just thought I'd throw that in there. I'm still unsure where I'd take all this in a discussion of freedom though. Again, if I've messed up something above, please jump in and correct me.