Talk:Chicken (game)

From Wikipedia, the free encyclopedia

	This article is part of WikiProject Game theory, an attempt to improve, grow, and standardize Wikipedia's articles related to Game theory. We need your help! Join in | Fix a red link | Add content | Weigh in
???	This article has not yet received a rating on the assessment scale. [FAQ]
???	This article has not yet received an importance rating within game theory.

Should this article be in the category Mathematical Games rather than Game Theory? -- Birkett 16:17, 25 Aug 2004 (UTC)

Under mathematical games, I understand games that require some knowledge of mathematics in order to play or solve them. In game theory, mathematics is used to study (dynamics of, optimal strategies for, ...) games. It could be interesting to add these (or other) views to the Mathematical Games page. The game of chicken is one of the simple games studied by game theory, so I think it is in the correct category. If there are more doubters though, you can put them in both categories, but I guess that other games like the Prisoner's Dilemma and Rock Paper Scissors and the likes should be then treated similarly. --Anthony Liekens 13:41, 26 Aug 2004 (UTC)

Contents 1 Chicken != Tragedy of the commons 2 Nash equilibria 3 Hawk Dove/Chicken 4 Peace war game

[edit] Chicken != Tragedy of the commons

I've never seen this version of the tradgedy of the commons. It really isn't like the prisoner's dilemma dynamic that is usually applied. While the verbal description given in this page is Chicken-like (ie a discoordination game) it's not really representative of the usual meaning of the model. Pete.Hurd 00:45, 24 April 2006 (UTC)

Yeah, I think at least the reference to TotC should be taken out, and unless sourced probably that whole section. --best, kevin [kzollman][talk] 01:28, 24 April 2006 (UTC)

OK, I'm on it. Pete.Hurd 01:54, 24 April 2006 (UTC)

ummm, while we're on the topic, the "Chicken and the prisoner's dilemma" section feels like a bit of a loss too. Chicken is a discoordination game and the PD is a game with a dominated strategy, very different. I'm trying to wrap my head around what the text in this section is trying to explain. Right now it leaves me more confused than enlightened. Anyone care to make the case for keeping? Pete.Hurd 00:46, 8 August 2006 (UTC)

[edit] Nash equilibria

Article says:

The Hawk Dove game has three Nash equilibria; 1) the row players chooses Hawk while the column player chooses Dove, 2) the row player chooses Dove while the column player chooses Hawk, and 3) both players play a mixed strategy where Hawk is played with probability p, and Dove is played with probability 1-p.

Isn't this wrong? It seems like 1) and 2) are not Nash equilibria, since the Dove player can improve his return by switching to Hawk (assuming C < V). Actually Hawk/Hawk is a Nash equilibrium. Right? (I'm no expert.) --Jorend 12:49, 27 July 2006 (UTC)

That's why the standard assumption is that V<C, the reverse is a boring case where there's no conflict of interest, and always attack is the strict Nash. Given the standard V<C assumption Hawk-v-Hawk is the worst possible outcome, and therefore not a Nash. Cheers, Pete.Hurd 00:42, 8 August 2006 (UTC)

[edit] Hawk Dove/Chicken

Hawk and dove has a strategy and payoff matrix sufficiently different from those for Chicken that the two games should be put back on separate pages. -- Mbhiii 18:07, 23 October 2006 (UTC)

Can you say more the ways in which you think they differ? The example of chicken given here is an instance of the general payoff matrix provided for Hawk-Dove. --best, kevin [kzollman][talk] 05:32, 24 October 2006 (UTC)

Those considerations alone seem sufficient to separate them. Otherwise, just put all 2x2 payoff matrix games together, as a sub-category of Payoff matrix. It is what they simulate in real life that makes them so different, and instructive. Hawk and dove could be a point of reference throughout a normal lifetime and could help one deal with unavoidable aggression, if played with proper strategy, for example the Hanseatic League vs. the Vikings. Anyone playing Chicken will have an expendable life, but since its real life analogues are avoidable, it represents no real threats. --Mbhiii 18:00, 26 October 2006 (UTC)

Thanks for your response. In game theory, it is common to group classes of games together that have similar properties. Both Chicken and Hawk-Dove have two asymmetric nash equilibria, and each actor has a preferred one. As a result they both have similar properties. Importantly for our purposes, ANY game with this sort of structure will be called Hawk-Dove or Chicken. Just because we use different payoffs in our examples in the article, it does not mean these are the only examples of each game. Any payoff matrix you provide for Hawk-Dove will also satisfy the conditions for being a game of chicken, and vice versa. For an example, see GameTheory.net's entry on Hawk-Dove. The first sentence is "The hawk-dove game is also commonly known as the game of chicken". --best, kevin [kzollman][talk] 19:49, 26 October 2006 (UTC)

Understanding all that, there still seems to me to be more to Hawk-Dove than its game theory properties. It is an instructive model for those wishing to find a way to deal with aggression. You have shown clearly its game theory properties, but to accomplish the other purpose, I've spun off Peace War Game with links to Peace and War, writing "- not a wargame, as such, rather a simulation of economic decisions underlying war." --Mbhiii 13:22, 31 October 2006 (UTC)

Mbhiii - I've read the new article. I have suggested its content be merged back here. I don't disagree that it would be good to have more in this article, and the stuff you added to Peace-war game would be a good addition here. But, please, understand that minor difference do not warrant a new article. I don't mean to be a pain, but I have yet to hear you give a reason why there need to be separate articles on the two games. I understand they have different descriptions, but I don't see why that requires two different articles. This article is not overly long, so including addition material here will not hurt this article. Having two separate articles is harmful because we will have to repeat much of what is said or users will have to flip back and forth between the two. --best, kevin [kzollman][talk] 16:39, 31 October 2006 (UTC)

It doesn't, necessarily, require two different articles, but the fact that you see the game's real-world instructive content and historical relevance as "minor" shows our differences. To me, there is more at stake here than correct game theory sub-categorization. Peace-War discovered something surprizing about the economics of war, and there should be a special place for it. People seeking to read about it, directed there from Peace or War, will want to do so directly without reading a lot of game theory. If they want to read more about game theory, those links are provided. The title game for this article, Chicken, does not rise to that level of seriousness or adequately represent those considerations. Its payoff matrix makes it irrelevant, never mind its game theory similarities.

What about "Iterated prisoner's dilemma" a main part of Prisoner's dilemma, do you want to merge that too? Iterated-prisoner's-dilemma/Peace-War/Hawk-Dove is sufficiently different in instructive content, historical relevance, and potential real-world analogues to be set apart from Chicken. If you want to merge them all together into one article, do it under one of those three titles, with Chicken as a curiously divergent sub-category. Thanks, Mbhiii 21:58, 31 October 2006 (UTC)

Chicken is not iterated prisoner's dilemma, they have very different game theoretic properties (nash equilibria, learning dynamics, etc.) As a result it would be inappropriate for them be in the same article. There is not much repetition in Prisoner's dilemma and Chicken, illustrating my point. On the other hand, a complete article on the Peace war game would discuss its nash equilibria, learning properties, etc. This would be redundant with the information in this article. As I suspect the two of us will not reach an agreement, I will solicit suggestions from other members of Wikipedia:WikiProject Game theory to see how they feel. --best, kevin [kzollman][talk] 22:26, 31 October 2006 (UTC)

I agree with Kevin here. When you say "just put all 2x2 payoff matrix games together, as a sub-category of Payoff matrix. It is what they simulate in real life that makes them so different, and instructive"[1]. I could not disagree more, the purpose of having these articles in the game theory project is to explain their game theoretical properties. If you think that all games in the abstract should all just be lumped together whilly-nilley and then have sparate articles that describe their applications to different topics, I think that's odd. It sort of runs against the idea of game theory: that the abstract model captures something essential about the real world. Ignoring the essential properties of the different games (and by extension grouping those that are essentially identical) in their abstract sense is to abandon the aim of capturing the essence of the "real life" problem in a model.

I'm not opposed to the idea of a separate article for the hawk-dove game, there is a great deal of published research on that model (work by Mike Mesterton-Gibbons, Philip Crowley, Carl Bergstrom, Eshel & Sansone, Rufus Johnstone leaps to mind for starters), an article describing that body of work would be a great thing. The old Hawk and Dove article [2] just wasn't such a thing, it merely rehashed the essential basic fundamentals of the model (and where it wasn't merely duplicating material copied from the Chicken (game) article, it presented material in an ideosyncratic way, see my comments [3] for e.g.). I see similar problems in the Peace war game article as in the old Hawk and Dove article, it's ideosyncratic to the point that I'm not convinced that it accurately represents the state of scholarship "in the field". For instance: "It is a variation, also called Hawk and Dove, of the iterated prisoner's dilemma in which the decisions (Cooperate, Defect) are replaced by (Peace, War)." seems to confabulate ideas which ought to be clearly distinct. Prisoner's dilemma is a totally different game from Hawk-Dove (aka chicken).

It stands to reason that articles covering the topic of game theory should have different articles for different games, and that games that are the same be covered in the same article. I am however, opposed to merging. Until reliable sources are produced to support the Peace war game article I suggest it be viewed as original research. If sources do support, then I'm still opposed to merging, it's different enough that I think it's like the extensive form game/game tree situation. Pete.Hurd 04:59, 1 November 2006 (UTC)

Needless to say that even if games have similar but different properties it might still be ok to put them in the same article, if discussing the differences in context makes more sense than discussing them separately. For naming issues we have redirects. On Peace/War vs. Prisoners Dilemma, I don't see anything in Peace/War that is structurally different, it sounds much more like PD repackaged for business seminars. And that's not something we need to cover in a separate article. ~ trialsanderrors 03:42, 15 November 2006 (UTC)

[edit] Peace war game

The talk page for Peace war game redirects here (which is probably appropriate due to the suggested merge) but then the cleanup-rewrite template also redirects to this page. So, in the hopes that someone came looking for what needs cleaning up in that article, see the section above... Pete.Hurd 04:24, 9 November 2006 (UTC)

• Note: the science citation index finds no hits for "peace war game", googling "Peace war game" computer simulation turns up no hits other than the wikipedia article. Pete.Hurd 18:48, 9 November 2006 (UTC)

As a math/phil undergrad in the 60s and a prob/stat grad student in the 70s, I recall occasional, but excited or reverential, mention of the "Peace War Game" by faculty in those departments (U of Ky, Harvard, Stanford), some of whom were engaged in playing it on computer or by mail, and always by that name, never by any of the other names mentioned above. In a corporate training session, in the 80s, (on departmental competition and cooperation) it was referred to again by that name. I've written what I remember of those discussions. --Mbhiii 13:51, 6 December 2006 (UTC)