Her prisoner’s dilemma with Barack Obama:
…you can recast the choices as:
If Hillary and Obama expose each other’s role in the foreign policy debacle, then both face political ruin and possible criminal liability, if any laws were violated.
If Hillary can pin it on Obama or Obama can pin it on Hillary then one walks and the other takes the rap.
If Hillary and Barack can cut a deal, then both walk or emerge with minimal damage.One of the assumptions of the prisoners dilemma is that they are isolated, precluding collusion. In this case since the parties are meeting, collusion is not only inevitable, but guaranteed.
On the other hand, you know what they say about honor and thieves.
They can both just blame it on Bush. The media and their sheep (but I repeat myself) will buy it. Neither one of these people has ever taken responsibility for anything in their entire lives. Why would they start now?
Hell they can both blame it on Mickey Mouse. Can anyone imagine Jim or Bobby-won altering their viewpoints one iota as a result of ANYTHING they might say?
Re “Nurse Ratched” vs. “Il Dufe,” as Lincoln said in another context, “Go it, husband! Go it, bear!” It’s like if Ivy Starnes got into a screaming cat-fight with Wesley Mouch. (Some of you should get the reference.)
As phrased by Mr. Fernandez, this is not a prisoners’ dilemma. The prisoner’s dilemma requires that behavior X (e.g., squealing) improves each party’s outcome, but if both parties choose X, then the outcome for both parties is worse than if each chose ~X. In other words, for each player p, F_p(X, ~X) > F_p(~X, ~X) [it’s better to squeal if my comrade squeals] and F_p(X,X) > F_p(~X, X) [it’s better to squeal if my comrade does not], but F_p(X, X) < F_p(~X, ~X) [if we both squeal, it's worse than if we both did not]. The condition F_p(X,X) < F_p(~X,~X) is central to the prisoner's dilemma.
The summary of the payoff matrix presented by Mr. Fernandez in the article is correct, but he does not then map this matrix to the Clinton-Obama dynamic. Mr. Fernandez meaningfully alters the payoff matrix by making (X, X) [both parties squeal, and both parties walk] no worse for either party than (~X, ~X) [both parties remain quiet, and both parties walk].
Perhaps the author misstated the consequences of (X, X). Perhaps the first to squeal actually does have an improved outcome. In this case, one player must go first, so (X, X) is no longer an available move in the game. There is, therefore, no dilemma: both players should rush to squeal as quickly as possible, and the first one wins.
Also, just to be hopelessly pedantic, there is no requirement for non-collusion in the prisoner's dilemma: the prisoners may communicate freely with each other without altering the payoff matrix of this game.
Also, just to be hopelessly pedantic, there is no requirement for non-collusion in the prisoner’s dilemma: the prisoners may communicate freely with each other without altering the payoff matrix of this game.
No, that’s incorrect. The inability to collude (or at least to know what the other prisoner will do) is a key part of the game.
The prisoner’s dilemma matrix is algebraically identical to the matrix in the two-player tragedy of the commons, and that game has no restriction on communication.
The Futurama episode “Murder on the Planet Express” ends by placing Fry and Bender into a prisoner’s dilemma, despite the fact that they were sitting next to each other on the couch.
Within a game, communication can alter the payoff matrix by allowing one party to communicate novel information to the other, thereby altering the game itself: If you tell me, “If you squeal, then my friends will kill you”, then you have altered the payoff matrix such that we are no longer playing PD. Of course, this only matters if I didn’t already know this; if I already knew this, then we were never playing PD (see my upcoming paper, “Why Crooks Love It When the Mob Kills Squealers” for further analysis ;).
I figure the use of isolation in the story of the prisoner’s dilemma is intended to prevent exactly this issue; i.e., to force the prisoners to play by exactly this matrix. I guess the question of whether communication matters in PD depends on where you put the boundaries of the game. If the matrix is the game, and the players cannot change the payoff matrix by communicating novel information, then communication has no effect. If the matrix can be altered by the players, then we’re playing a much bigger meta-game, and communication might matter.
As I said, the prisoner’s dilemma isn’t a dilemma if you have certain knowledge of what the other will do. In the iterated game (with the computer tournaments that Axelrod hosted at Michigan), one can only attempt to infer it from past exchanges, and that is the only communication available.
Even if you go first, and I go second, and I know exactly what you did before I take my turn, we still both squeal. (Regardless of your move, it’s better for me to squeal).
This is true whether you do or do not know that you are going first. If you don’t know that I will find out what you did, then you are playing “normal” PD (and your best move is to squeal). If you do know that I will find out what you did, you’ll squeal to protect yourself from the worst outcome: (don’t squeal, squeal) really sucks…
———
If you go first, then I find out what you did, and I go second, but then you have a chance to change your mind after you find out what I did, we still both squeal. (If you don’t squeal, I will, and when you find out what I did, you’ll squeal. If you squeal, so will I, and you’ll stick with squealing).
If you go first, and I go second, then you get to change your mind, and then I get to change my mind, ad infinitum, we still both squeal. (Exactly the same as above. The state (squeal, squeal) is the only stable play).
——-
IIRC, “tit-for-tat” does fairly well in iterated PD (IPD) tournaments, but only because it picks up a few points against random strategies. “Tit-for-tat” still loses or ties against “always squeal”, though: if TfT picks “don’t squeal” as its first move, it loses; if TfT picks “squeal” as its first move, it ties.
No, TfT does well in an environment with other TfTers. It will not do well by itself in a group of defectors. But if there are enough TfTers for it to interact with, eventually they will take over. That’s why Axelrod called it the “evolution of cooperation.”
Perhaps though, the game seems less like a “dilemma” to me, if I know what you are going to do (or did). My choice becomes obvious.
Kinda loses the “Lady or the Tiger” vibe if I already know you decided to squeal.
Of course, my move is still the same, regardless of what you did. I squeal, and so do you.
To back up what our host said, here’s the Wikipedia entry: http://en.wikipedia.org/wiki/Prisoner%27s_dilemma
If you prefer a more scholarly approach, here is a page at Stanford University: http://plato.stanford.edu/entries/prisoner-dilemma/
Please note that the Stanford example states that the two prisoners are “placed in separate isolation cells.”
Well, the way suspects are played off against each other is they are kept in separate interrogation rooms, but even if they are in communications with each other, they can still cheat.
Where it gets interesting is the Ms. Clinton is playing the Prisoner’s Dilemma — against herself. Does she go with the new tough Clinton or the old loyal-to-the-administration Secretary Clinton?
What makes this different than the Prisoner’s Dilemma game is the fact that the bad consequences will be mitigated by the MSM.
When the MSM concludes that Obama can and must be thrown under the bus to save Hillary, then any bad consequences that Hillary might suffer as the result of Obama’s actions will be mitigated.
And I think I’m seeing that MSM conclusion in action already, though not dialed up to maximum power.