Talk:Chicken dilemma

This is the discussion page (the "Talk:" page) for the page named "Chicken dilemma". Please use this page to discuss the topic described in the corresponding page in the main namespace (i.e. the "Chicken dilemma" page here on electowiki), or visit Help:Talk to learn more about talk pages.

Merged chicken dilemma articles

Prior to today, there was an article called "Chicken Dilemma Criterion" (which I'll call "A") and one called Chicken dilemma (which I'll call "B"). They needed mergeing. The Chicken Dilemma Criterion ("A") was a more complete article, but Chicken dilemma ("B") had a better title. So here's what I did:

Moved Chicken dilemma ("B") to Chicken dilemma criterion
Moved Chicken Dilemma Criterion ("A") to Chicken dilemma
Copied the content from "B" to the top of "A", thus merging them into "AB" at Chicken dilemma
Replaced Chicken dilemma criterion ("B") with a redirect to Chicken dilemma ("AB")
Did some copyediting work

This article needs much more work, but I think I'm done working on it for now. -- RobLa (talk) 04:13, 24 December 2019 (UTC)

Chicken dilemma section title

Kristomun, why did you change the section title from "Definition" to "Definition of chicken dilemma criterion"? I can understand doing that when the article is brand-new, but I suspect that you may have broken that link (i.e. if someone on Reddit linked to the Definition section, their link will no longer go to the right section). In general, I think we editors ought to establish a practice of being reticent to touch section names. BetterVotingAdvocacy (talk) 17:40, 24 March 2020 (UTC)

Because it's not a definition of the chicken dilemma. Electowiki has quite a few old pages where the threshold for revision is, I think, lower than on an ordinary Wikipedia page. For instance, some pages have an informal style and were originally written by Mike Ossipoff. As I understand it, the chicken dilemma page is one of these, so I did the edit with clarification in mind.

I see your point, though; if you think it's serious, then we could change "Definition of..." back into "Definition" and then change "Formal definition" into "Formal definition of...", as it's much less likely that anyone would have linked directly to the first subsection of a section. I think we're fine in any case (and the current change is a little more elegant), but if you disagree, feel free to do the change. Kristomun (talk) 20:35, 24 March 2020 (UTC)

User:Kristomun, thanks for cleaning up some of Mike Ossipoff's writing. Mike was very influential in my thinking about electoral reform, but he seemed to prefer raw stream-of-consciousness writing. The sheer quantity of his prose made him less persuasive to a general audience than he would have been with shorter, well-constructed prose. I did a litte more restructuring of the article (consolidating the two definitions into a "Definition" section), but I didn't do a deep rethink of the article. We shouldn't feel bashful about trimming back his work, and restructuring it some more. It doesn't seem like it's necessary to have a "CD" and "CD2" definition, and those names are kinda terrible section titles. I hope we keep the titles as simple as possible so that hashtag links are easier (e.g Chicken dilemma#Definition), which is why I changed "Definition of..." back into "Definition".

User:BetterVotingAdvocacy, you'll notice that I added a section title to this part of the conversation. Could you make a point of adding section titles when you start a new conversation? Thanks in advance. -- RobLa (talk) 02:28, 25 March 2020 (UTC)

Chicken dilemma vs center squeeze

I just realized that I tend to think of chicken dilemma and center squeeze as being the same problem. Is center squeeze just a flavor of chicken dilemma? -- RobLa (talk) 06:55, 6 February 2022 (UTC)

I think chicken dilemma is more the "anti-center squeeze", i.e. they're opposites but related. In a method that fails chicken dilemma, there is sometimes an incentive for a wing of the majority group to not approve or rank all the group's candidates. (E.g. Nader>Gore voters may get Nader elected by not approving of Gore). This is an incentive to pull away from the center, which goes wrong when too many voters do so. On the other hand, center squeeze is a failure to find the center where it exists, and incentivizes compromising to pull towards the center.

Both problems may cause instability if the voters try to compensate in order to get the true center elected. But the chicken dilemma happens when the voting method has a broad support bias (like Approval), and center squeeze when the voting method has a narrow support bias (like IRV). Kristomun (talk) 14:28, 6 February 2022 (UTC)

Let's say that we find an election with the following scenario

39:LarryLeft>MalloryMiddle>RonaldRight
21:MalloryMiddle>LarryLeft=RonaldRight
19:RonaldRight>MalloryMiddle>LarryLeft
21:RonaldRight>LarryLeft=MalloryMiddle

... where the candidates are on the classic "left-right" political spectrum. Wouldn't MalloryMiddle be in danger of being squeezed out of the middle of this election with only 21 first-place votes? -- RobLa (talk) 00:22, 7 February 2022 (UTC)

Using symmetric completion so that Rob LeGrand's voting calculator lets us use IRV:

39:LarryLeft>MalloryMiddle>RonaldRight
11:MalloryMiddle>LarryLeft>RonaldRight
10:MalloryMiddle>RonaldRight>LarryLeft
19:RonaldRight>MalloryMiddle>LarryLeft
11:RonaldRight>MalloryMiddle>LarryLeft
10:RonaldRight>LarryLeft>MalloryMiddle

Every method on his list elects MM, except Carey and IRV (where there's a tie between LL and RR). So this is center squeeze. Now, truncation may be effective for a few methods, e.g.:

29:LarryLeft
10:LarryLeft>MalloryMiddle>RonaldRight
11:MalloryMiddle>LarryLeft>RonaldRight
10:MalloryMiddle>RonaldRight>LarryLeft
19:RonaldRight>MalloryMiddle>LarryLeft
11:RonaldRight>MalloryMiddle>LarryLeft
10:RonaldRight>LarryLeft>MalloryMiddle

and LarryLeft wins in Baldwin and Raynaud. If the MalloryMiddle voters also truncate, then RR wins. So there's a chicken dilemma in Baldwin and Raynaud (as far as I can see). My characterization from before might have been wrong: Left is not the true center here; so a chicken dilemma might either be something biased too far in favor of the center, or the method originally settling on the true center but being able to be pulled away from it by truncation right up until a swing the other way happens (the crash in the chicken dilemma).

But IRV has no problem with this, it just sticks to its tie all the time. Condorcet,IRV would let the LL voters get away with their truncation, but then no truncation on the MM voters' part will restore the win to MM.

So in that sense, although the same election can demonstrate chicken in some methods and center squeeze in others, I don't think you can get both at the same time for the same method. Kristomun (talk) 09:56, 7 February 2022 (UTC)

To put it differently, center squeeze initially happens without strategy: the voters may compensate by compromising. Chicken dilemma happens when voters want to get a better result by strategizing, but if they all do it, then something bad happens. It's the use of strategy that's the problem in both cases: compromising strengthens two-party domination, while truncation leads to bad outcomes when some faction does too much of it. Kristomun (talk) 10:00, 7 February 2022 (UTC)

Thanks, those examples are helpful. I tried the top one out on Eric Gorr's calculator: (<https://www.ericgorr.net/condorcet/>), with the "tell me everything" setting as well as all of the available methods. Since I considered Ross Perot "in the middle" in 1992, my mental model seriously conflated "chicken dilemma" and "center squeeze". -- RobLa (talk) 08:17, 8 February 2022 (UTC)

Snowdrift

The article as currently written compares the "chicken dilemma" to "snowdrift"

...a situation that resembles the classic "chicken" or "snowdrift" game...

This makes the assumption that people know what the "snowdrift" game is. I have no idea what this refers to. Could someone educate me? -- RobLa (talk) 22:54, 28 November 2022 (UTC)

IIRC, it's a game involving two cars stuck in the snow, where each driver may either shovel snow (which costs effort) or just sit in his car (which doesn't). If both just sit in the car, they get nowhere, so that gives a negative outcome for both. If only one shovels the snow, then he has to pay for the effort while both drivers get out of the snowdrift. Finally, if they both do so, they share the effort and reward.

Neither player shoveling is like neither player swerving in chicken (big loss for both). Both shoveling is analogous to both swerving (moderate gain for both). Only one shoveling is analogous to only one swerving (loss to the one who shovels/swerves, greater payoff to the freeloader).

So the structure of the payoffs is the same as in Chicken, the actions just have different names. See also https://wiki.snowdrift.coop/about/snowdrift-dilemma. Kristomun (talk) 14:36, 2 December 2022 (UTC)