Talk:Prisoner's dilemma

what?

A purely selfish strategy will not "cheat" on its opponent for purely self-interested reasons first.

Why is that sentence in the "Nice" section? How is a purely selfish strategy "nice?" 107.77.195.33 (talk) 04:39, 24 November 2023 (UTC)

I also don't understand the logic of that sentence. How is a purely selfish strategy not purely selfish? 107.77.195.33 (talk) 04:42, 24 November 2023 (UTC)

I didn't find many mentions of "purely selfish" in decent sources after a quick search, although I think it means does not consider the other party at all, that is, it always chooses the best outcome--defect. I am changing this to "This strategy will not "cheat" on its opponent for purely self-interested reasons first" for now. Winston (talk) 02:02, 19 June 2024 (UTC)

Yeah, this section is a bit convoluted. "Nice" simply means the algorithm will not defect first, compared to algorithms which will try to defect. Winston (talk) 01:50, 19 June 2024 (UTC)

what person?

For example, if a population consists entirely of players who always defect, except for one who follows the tit-for-tat strategy, that person is at a slight disadvantage because of the loss on the first turn. In such a population, the optimal strategy is to defect every time. 107.77.195.33 (talk) 04:45, 24 November 2023 (UTC)

Christmas truce in "See also"

No inline citation is given for why the Christmas truce is mentioned at the bottom of this article, and mentioning it in this article seems inappropriate - can anyone explain? As I understand it, the Christmas truce was a stag hunt and therefore should instead be mentioned in that article:

- if both sides cooperate, everybody involved gets to live that day and enjoy the truce;

- if both sides betray, then the war continues as usual with some casualties being expected;

- if one side cooperates with the other betraying, then the cooperating side suffers massive casualties due to an ill-fated attempt to cross no man's land, while the betraying side can shoot the enemy with impunity, but does not enjoy a truce that day. TheDestroyer111 (talk) 08:53, 7 January 2024 (UTC)

As it turns out, the mentioning of the Christmas truce was introduced in a June 2010 edit with no reason given. TheDestroyer111 (talk) 19:45, 2 February 2024 (UTC)

The grammar is wrong.

The Prisoners' Dilemma redirects to The Prisoner's Dilemma. The plural possessive, however, is the correct form (regardless of how many times the mistake has been repeated in the past) and the singular possessive is wrong.

While there's only one dilemma (it's the same dilemma), it must be faced by multiple prisoners, otherwise the entire thought experiment is meaningless. Therefore multiple prisoners face the dilemma in question. It's the prisoners' dilemma, not any one prisoner's dilemma.

This grammatical atrocity must be ferreted out and stopped anywhere it can, and redirecting from the correct grammar to the wrong grammar is enabling the reign of the kingdom of the Doofuses.

Next step: figuring out how to make Chrome stop underlining the correct form. 2601:1C2:5000:8CC7:F5EF:AC6C:C8:5638 (talk) 03:35, 18 February 2024 (UTC)

Hi, What you are commenting is not a grammar issue, it is just a preference for naming the game. Grammatically both expressions are correct. We could also argue that the dilemma is not really about prisoners but about criminals or defendants, and argue in favor of changing the name of the game accordingly. Now, i personally prefer singular, as it puts emphasis on the point of view of one (arbitrary) criminal/defendant/prisoner, and the concept of Nash equilibrium also puts emphasis on the issue with (isolated/distributed) individual decisions. I checked the corresponding wiki page in about 10 other languages and singular is used in all. 37.67.33.80 (talk) 06:36, 7 May 2024 (UTC)

Plus/minus and payoff/cost

I suspect that many may be puzzled comparing the "payoff" matrix in the Prisoner's dilemma#Premise, displaying years in prison as payoffs, and the one in the Prisoner's dilemma#Generalized form, displaying, as expected, a more desirable payoff with a greater value than a less desirable one. We ask for instance in a general prisoner's dilemma that $R>P$, where R is the payoff for cooperating=staying silent, and P that for defecting=testifying; this is inconsistent with payoffs being years in prison. The matrix displaying years in prison should rather be called "cost matrix", or be qualified with some other negative word like "punishment/penalty/sentence/...", and then the page should have a word of explanation to relate payoffs and costs. Alternatively we can display negatives of years in prison, and keep calling it "payoff matrix": this is consistent with the definition of a general prisoner's dilemma. 37.67.33.80 (talk) 05:46, 7 May 2024 (UTC) [I inadvertently posted this without being logged in, my username is plm203.]

Moral philosophy

I was reading today about the applications of prisoner's dilemma in moral philosophy, and noticed that there is no section about it on this page! I just added a new section, along with some opinions. Feel free to expand and discuss

I still can't reconcile these 2 statements in the current article

I still can't reconcile these 2 statements in the current article:

"For cooperation to emerge between rational players, the number of rounds must be unknown or infinite.

"Another aspect of the iterated prisoner's dilemma is that this tacit agreement between players has always been established successfully even when the number of iterations is made public to both sides."

~~~~ Mathiastck (talk) 01:17, 23 July 2024 (UTC)

I think the idea is that when real people play repeated rounds they do not follow the rational / dominant strategy, and co-operation still emerges. I can believe this is true, but it would probably be better supported by a reference. Mgp28 (talk) 08:01, 23 July 2024 (UTC)

I can find references in which co-operation emerges in a finite game (I'll leave this reference here as an example) but the word always needs some justification. I'm tempted to rephrase without the always but equally good if there is a citation to support it. Mgp28 (talk) 08:32, 23 July 2024 (UTC)

References

Suppose a prisoner testified but perjured their testimony with false extraneous details.

Just saying. Hope no policy has been broken by my use of this. Tathagata1st (talk) 17:16, 11 September 2024 (UTC)

Lacks reference

The following paragraph lacks reference.

"In contrast to the one-time prisoner's dilemma game, the optimal strategy in the iterated prisoner's dilemma depends upon the strategies of likely opponents, and how they will react to defections and cooperation. For example, if a population consists entirely of players who always defect, except for one who follows the tit-for-tat strategy, that person is at a slight disadvantage because of the loss on the first turn. In such a population, the optimal strategy is to defect every time. More generally, given a population with a certain percentage of always-defectors with the rest being tit-for-tat players, the optimal strategy depends on the percentage and number of iterations played." 202.92.130.118 (talk) 02:52, 25 September 2024 (UTC)