The Issue of CSM Elections or the Condorcet Paradox

Communications Center General Discussion
22

Damn OP. Well thanks for trying.

How about a compromise: establish parties. like the British do. A Nullsec party, Carebear party, Ganker Party, Pirate Party, etc. Make it more parliamentary.

The CSM is like all the worst elements of the US system and zero benefits (no “branches”).

I say, give us a KING who says “let’s take Eve into the future and make it a success and it’s my way so at least one of us will be happy”. Then we can vote with our wallets, which is in the end the only vote that ever mattered, ever, in any situation.

1 Like
24

look I understand I hurt your feelings and will review my previous posts so I can see what I did and do it again.

1 Like
26

image

27

Both the destruction and creation as loses and profit are important. Nothing will be possible to achieve in this game by having only destructions and loses. The more ISK and assetes will possess every character in this game, the more content it will produce for it. For example, remember the WWB (Casino War) and those barons Eep and Lenny behind it and how they founded a campaign against CFC (I’d like to say not against Goonswarm). It’s debatable how they got their money, because it’s anti EULA or Policy Terms, but the truth is that it allowed entire campaign to happen. The same can be said about mentioned Olmeca Gold. C5 and C6 WH allowed her to make enough money for her multibox fleet set firstly and after a successfull campaign to invest even more ISK into a Cap multibox fleet. As she explained in “Whaler” video, because a ganked industrialist had enough ISK, it allowed him to place a bounty on Goon’s Supers and Harvester Drones which was announced by SLYCE CEO Riotrick. An example when Bounty Office works in game.

IDK, I didn’t saw any good suggestions about how to rebalance ISK income and outcome. No need to reduce the loot drop and bounties on combat NPC sites, exploration sites, which can be performed and are performed as usually by a single character. It’s frustrating to get a garbage loot from a site, which requires a 1.5B ISK vessel to clear it in 40-50minutes! Let the in game market aka demand and supply to determine the site profity and the level of interest to run them. To build or buy a Cap ship requires much more ISK. Actually it’s nonsense to stack ISK for a Carrirer by doing H-Sec missions, right? Capitals are N-Sec tools. The Assets and ISK to build or buy them respectively should be gained in 0.0 systems. It must be a challenge for them as it is in C6 WH. A 5/10 DED Expedition is also a challenge for a single character in H- and L-Sec. It’s hilarious when a drone Cruiser (no ammo spent) can complete the 0.0 Heaven sites AFK! Put some Dreadnoughts and Carriers there, with better bounty and loot, between those Heaven’s spawns. Make it harder for capital pilots and botters and near impossible for sub-cap vessels. Why this part of the game remains as it was 10 years ago? Instead they nerfed VNIs, Gilas and etc.

Any player not affiliated to a big N-Sec Bloc should not worry about their candidates. They’ll have a candidate in CSM14. You should be interested into your play style, whatever it is: Industry, Exploration, FW, Ganking, Abyss, Incursions. Olmeca Gold is a new guy, with a new play style, and it’ll be interesting to see and listen what he’ll offer for the game. What he’s doing is very hard to acomplish for even a good PVP player. I’d like to see any suggestings about how to reduce a bit the playstyle rewards by improving its accessibility, considering he has a decent CSM program.

You mentioned “I highly recommend anyone who does not wish to see yet another CSM entirely dominated by NullSec/GoonSwarm interests”, but offered the rank voting:

Olmeca Gold -> Mike Azariah -> Steve Ronuken -> Exookiz -> Manic Velocity -> Xenuria

Xenuria is a good guy, but if Olmeca, Mike, Steve, Exoo and Manic will fail then Xenuria will win, being a Goonswarm member.

28

We need a different voting model to have a candidate per different game playstyle. In a PR-STV voting model it’s practically impossible.

Imagine a scenario when CCP announces that every alliance must be submitted by one candidate. It will not stop the second candidate from any alliance to join to a FW allaince or to be in a NPC Corp by recieving votes from his N-Sec Bloc as currently. He’ll be a H-Sec representative on paper only.

EVE players have only one option to somehow influence on CSM election results by rank voting up to 3 favorite candidates. The more different candidates you’ll put into your ranked votes the lower chance your favorite candidate will have.

We just can increase the Quota with our votes. It will make more harder or much expensive for a Bloc group to have multiple seats in CSM14.

29

I’d prefer to use the “groups by interest” terminology here.

For example, let’s say that we have a represintative from The Imperium, Legacy and PanFam at CSM. They all are N-Sec dwellers with the same problems and playstyle actually. What one representative can offer that another one can’t related to N-Sec life?

The Imperium in 2016 can offer: add Jump Fatigue, boost Rorquals, while the PanFam representative can offer in 2016: remove Jump Fatigue, nerf FAXes. In 2019 The Imperium representative can offer (under CCPLS banner): nerf Jump Fatigue, add Jump Bridges, nerf Rorquals, nerf FAXes, while the PanFam - vice versa: buff FAXes, improve Jump Fatigue. The Legacy can offer in 2019: add Shield Implants, new Shield Rigs, add nullification to combat Interceptors.

2 Likes
30

Therein lies the heart of the issue: are any of these changes good for the game?
Do any of these changes add conflict of thwart it?
Does anybody who is not in these groups even understand these changes?

This is why I think the CSM should be tossed out. Just get some good people in charge who think in terms of “good game” and let them direct what happens. If they screw it up, they get fired.

All I see is CCP hiding behind the CSM in the same way the US congress lets the executive and judicial branches do what they feel like then hide behind that and claim it was not their doing if results of things don’t go well or if the endeavor or ruling become unpopular. “Don’t blame CCP. Blame the nullbears they voted for it!”

I’m tired of it. Rule by vote is a farce. Find people who are really good at what they do, who deserve to be in charge, and put them in charge. The game needs people who really know games, gamers, and design well. CSMs are players, and players for the most part are only good at playing games.

3 Likes
31

Do you have a link or an article or something that supports this claim that CCP hides behind the CSM?

32

Given your low tier communications skills would it matter?

1 Like
33

Yes. It would tell me that you have actual support for your claims and not just pulling things out of your rear end.

1 Like
34

You really ask to much.

35

I don’t have to take orders from gammas. Go get some proof that proves me wrong.

36

The point of Condorcet’s Paradox and more broadly Arrow’s Theorem or even more broadly Nakamura’s Number is that sometimes voting can provide dubious results. Not that all elections produce dubious results. For example, if there is a Condorcet Winner, then Condorcet’s Paradox does not hold. Similarly, if there is a candidate that all voters prefer then Arrow’s theorem implies that candidate should win.

And these results only apply to deterministic voting models. These results do not apply to probabilistic voting models.

The probabilistic voting theory , also known as the probabilistic voting model , is a voting theory developed by professors Assar Lindbeck and Jörgen Weibull in the article “Balanced-budget redistribution as the outcome of political competition”, published in 1987 in the journal Public Choice , which has gradually replaced the median voter theory, thanks to its ability to find equilibrium within multi-dimensional spaces.

The probabilistic voting model assumes that voters are imperfectly informed about candidates and their platforms. Candidates are also imperfectly informed about the utility preferences of the electorate and the distribution of voters’ preferences.

Unlike the median voter theorem, what drives the equilibrium policy is both the numerosity and the density of social groups and not the median position of voters on a preference scale. This difference explains why social groups which have a great homogeneity of preferences are more politically powerful than those whose preferences are dispersed.

1 Like
37

In social choice theory, Arrow’s impossibility theorem , the general possibility theorem or Arrow’s paradox is an impossibility theorem stating that when voters have three or more distinct alternatives (options), no ranked voting electoral system can convert the ranked preferences of individuals into a community-wide (complete and transitive) ranking while also meeting a specified set of criteria: unrestricted domain , non-dictatorship , Pareto efficiency , and independence of irrelevant alternatives . The theorem is often cited in discussions of voting theory as it is further interpreted by the Gibbard–Satterthwaite theorem. The theorem is named after economist and Nobel laureate Kenneth Arrow, who demonstrated the theorem in his doctoral thesis and popularized it in his 1951 book Social Choice and Individual Values . The original paper was titled “A Difficulty in the Concept of Social Welfare”.[1]

…when voters have three or more distinct alternatives (options)…

In almost all practical scenarios not all voters prefer a single candidate. The CSM voting model is based on PR-STV voting model. In a STV election the quota is the minimum number of votes a candidate must receive in order to be elected. Extra votes are shared to next candidates in voting list. The Quota defines the winner. Moreover, The PR-STV model doesn’t elect a single candidate for one seat, but multiple candidates for multiple seats (in our case 10 seats).

CSM elections, in my opinion, is a communistic voting procedure to define a new Eden dictatorship order by using a democratic STV voting model. Her’s an example, not perfect, but a good one to demonstrate how its working:

For example, in an election for 3 seats, 2 parties (A as Goons and C as NC) present two candidates and a third party B as Manic, Steve or Olmeca presents a single candidate. All party A voters prefer party B to party C (of course, none Goon voter will vote for an NC member). All party C voters prefer party B to party A (almost nobody from NC will vote for a Goon candidate). All party B voters prefer party C to party A. The following 1000 votes are casted:

  • 220: A1 > A2 > B;
  • 200: A2 > A1 > B;
  • 190: B > C1 > C2;
  • 250: C1 > C2 > B;
  • 140: C2 > C1 > B.

The quota is 250. C2 is eliminated first and candidates A1, B and C1 are elected.

Things are way more complex and broken in EVE Online than in RL, as explained above, because big Bloc groups will use the “command dictature” to vote with their main and ALT accounts.

Let’s suppose, based on previous example, group A will give 220 + 200 = 420 votes to candidate A1 and the same amount of 420 votes to candidate A2 by using their Omega ALT accounts. Group C will give 250 + 140 = 390 votes to both C1 and C2 candidates in the same way. Candidate B will get 190 + n (n << 190) like 220 votes. The Quota = ((2x420 + 2x390 + 220)/4) = 460 will be higher than maximal amount of votes that any candidate has. B is eliminated first and his 120 votes will be distributed between C1 and C2 like 390 + 60 = 460. Depending by STV algorithm, votes from A2 will be distributed to A1 or vice versa. Candidates C1, C2 and A1 or A2 are elected.

In RL there are many tricks to confuse voters, like propaganda. In some countries opponents forming dummy candidates to steal votes from main and most popular candidate. Lets say the top candidate’s name is John Smith. The dummy candidate name is registered as John Smart. This trick is created for pensioners and people with poor eyesight. I saw things like this.

Your argument " if there is a candidate that all voters prefer" will work in “green zones” where there isn’t a real working economy, where there isn’t a dispute of interests between social and economical groups. Such “green zones” are some small countries, where everything is easy to control and count and big money are pumped there from around the world, like Switzerland, Luxembourg, Liechtenstein, Monaco, Singapore and some cities like Vancouver, San Diego. In “green zones” people aren’t going to work and to make money. They going there with a lot of money already made, in less fortunate places, to live a luxury life. Well, Vancouver isn’t a very luxury city, but its very polarized. Maybe 60% of habitants are families of rich China intrepreneurs. Clean ecology, clean air, study at prestigious universities, the development of foreign languages, new connections … to return and continue the family business. Everyone is going to live more than a hundred years. Sigh! What voting you are talking about?

38

Yes I read Arrow’s theorem (the first time was like 25 years ago)…maybe you should read the part that those results only apply to deterministic voting. If we live in a world of probabilistic voting then those results do not apply, at least nobody has shown that they do as far as I know.

First off, that is not my argument. Second, a Condorcet winner is not a candidate that everyone prefers. For example, using P to indicate preference the following preference ordering over three voters for three candidates leads to a “cycle”:

Voter 1: a P b P c
Voter 2: b P c P a
Voter 3: c P a P b.

In this case, in a simple majority voting there is no winner if each voter votes according to his or her preferences. Further, there is no clear winner in any pair wise matching. For example:

a vs. b implies that a wins ( 1 and 3 vote for a, while 2 vote for b).
b vs. c implies that b wins ( 1 and 2 vote for b, while 3 votes for c).
a vs. c implies that c wins (1 votes for a, 2 and 3 vote for c).

Hence a cycle appears if voting is done via pair wise contests where pairs are determined randomly. If one of the voters can control the pair wise matchings and also knows the voters preferences they can control the outcome (dictatorship).

This is a specific example/special case of Arrow’s theorem.

However that distribution of preferences is but 1 of 18. In fact of the 18 different possible preference orderings that is the only ordering that exhibits the problem of cycling/dictatorship. For example, suppose instead preferences are,

Voter 1: a P b P c
Voter 2: a P b P c
Voter 3: c P a P b.

That is voters 1 & 2 share the same preference orderings over a, b, and c. Then A is the Condorcet “winner” in that in any pairing with candidate a then a wins.

a vs. b:
Both voters 1 & 2 vote for a so a wins.
a vs. c:
Both voters 1 & 2 vote for a so a wins.

As for b vs. c, that doesn’t matter in that whichever one is the victor it will go up against a and lose given the results above. Thus candidate ‘a’ is the Condorcet winner even though only 2 out of 3 voters prefer ‘a’ as their “top choice”.

Third, this is actually part of Arrow’s theorem, it is the Pareto criterion or Pareto efficiency. If every voter prefers A to all other candidates then A wins. It is the last assumption shown in the wikipedia page.

Pareto efficiency, or unanimity
If every individual prefers a certain option to another, then so must the resulting societal preference order. This, again, is a demand that the social welfare function will be minimally sensitive to the preference profile.

The point of Arrow’s theorem (or more generally the Nakamura Number) is not that every elections is going to result in a bad outcome, but that sometimes voting yields a bad outcome because voting does not give us a preference ordering that is like those of individuals, at least not all the time. Arrow’s theorem tells us we should stop looking at voting as a deterministic process, but should instead think of it in terms of game theory (see the Nakamura Number link, voting is actually an example of cooperative game theory). Further, that institutions matter. Since voting can go “off the rails” sometimes we want institutions that can limit the damage when “going off the rails” occurs.

1 Like
39

This hasnt helped tell me who I shouldve voted for.

40

The CSM is pretty pointless - as far as voting goes. There should be a way to distribute the power so it isnt always sitting with the SOV folks. SOV isnt the end game of Eve.

There really should be a way to even that out between the types of space - HS, LS, NS. How to get there, i dont know but the CSM in its current form is leading to much in one direction.

41

I would love to see that. Therefore it will never happen.

42

Why? The type of space isn’t the problem. The fact players in all other places than null don’t organize themselves is the reason why CSM is run by null blocks.

1 Like
43

It’s not need to cover the generic concepts and theorems with a lot of specific terminology and even more support theorems that consider partial cases of the general theorem. The concepts as “deterministic” or “probabilistic” should be used by professionals in calculation models to get a specific result. In reality, most voting processes are deterministic, while when calculating the probabilities of something in science or in game choice - should be used the probabilistic model mainly.

Do you think that Goonswarm voting order for CSM as Aryth>Merckelchen>Innominate>Xenuria>…Vily>? is probabilistic?

Do you think that Legacy voting order for CSM as Vily>Cornak>Dunk Dinkle>? is probabilistic?

Do you think that republican’s vote for Trump is probabilistic? There weren’t scenarios, in the republican camp, like this: Ah, today is a voting day! I should vote for somebody. Let me think. IDK, I’ll vote for Trump because I like his tie.

Theory: Deterministic formal models assume that voters make decisions with certainty while probabilistic models assume that voters’ intentions are uncertain. The classic deterministic model assumes that individuals choose the candidates whose positions are nearest their ideal points with certainty. In contrast, probabilistic models posit that the likelihood of voting for a particular candidate increases as distance decreases, but to vote against the nearest candidate remains possible.

There is another, more practical, definition to notice when the Arrow’s theorem criteria can happen - ordinal utility. We must rely on ordinal utility not on deterministic concept to define the usage of Arrow’s theorem areas of application.

In economics, an ordinal utility function is a function representing the preferences of an agent on an ordinal scale. The ordinal utility theory claims that it is only meaningful to ask which option is better than the other, but it is meaningless to ask how much better it is or how good it is (as CSM voting happens).

The ordinal approach is based on the fact that the preferences of an individual regarding the alternatives proposed for selection cannot be measured quantitatively, but only qualitatively, that is, one alternative is worse or better than the other.

Ordinal utility contrasts with cardinal utility theory: the latter assumes that the differences between preferences are also important.

Within the framework of the cardinal approach, which implies the quantitative measurability of preferences, Arrow’s theorem in general does not work.

Or, in more public terms

Arrow has said “Most systems are not going to work badly all of the time. All I proved is that all can work badly at times.”

You mentioned the Nakamura number. Its calculation is like a quantum mechanics in social statistics and mainly optimised for games. In real elections we should consider the human factor. CSM elections, as many real elections, uses the STV voting model. STV voting model uses the Quota concept, as I mentioned, which “adds” the cardinal utility and Nakamura number ideas to the voting model indirectly. It’s a rough definition of course. Anyway, the STV voting model is vulnerable as I demonstrated for elections between Goonswarm and NC few posts above.

In general, all these things as CSM elections looks “fishy” as I mentioned in the OP. Just a rhetorical question. Do you think that Exookiz got more direct votes than Sort Dragon? No! He got the same amount of votes. But he got more shared votes from Bloc lists also: from Innominate, Xenuria, Dunk Dinkle, Matterall?

My expectation for CSM14, a month ago was:

  • Goownswarm 2 + [3]: Aryth, Innominate, [Merkelchen], [Xenuria], [The Judge];
  • NC 1 + [1]: Killah Bee, [Matterall];
  • TEST 1 + [1]: Vily, [Cornak Firefist];
  • Horde 1 + [1]: Gobbins, [Arqui Nurbs];
  • BRAVE 1: Dunk Dinkle;
  • DARKNESS 1: Sort Dragon;
  • ProviBloc 1: Carbon Alabel;
  • H-Sec 1 + [2]: Steve Ronuken, [Mike Azariah], [Manic Velocity];
  • L- J-Space 1 + [?]: Olmeca Gold, Mathew Dust, Exookiz, Blood Ruin and e.t.c.

Official data:

Summary

Round 1 beginning - 44 candidates remain
32994 votes, 3000 quota
Initial talley:
7200 “Aryth”
5089 “Gobbins”
3232 “Olmeca Gold”
3011 “Vily”
2196 “Killah Bee”
1500 “ExookiZ”
1500 “Sort Dragon”
1389 “Dunk Dinkle”
937 “Stitch Kaneland”
920 “Steve Ronuken”
717 “Cornak Firefist”
682 “Matterall”
476 “Mike Azariah”
466 “The Judge”
446 “Merkelchen”
324 “Innominate”

…
Actions:
Elected: “Aryth”
Elected: “Gobbins”
Elected: “Olmeca Gold”
Elected: “Vily”
Transfer from “Aryth”:
Votes: 7200.000000, Factor: 0.583333, Excess: 4200.000000
3857.000000 votes to “Merkelchen”
233.916667 votes to “Innominate”
24.500000 votes to “Xenuria”
22.166667 votes to “The Judge”
15.166667 votes to “Killah Bee”
13.416667 votes to “Steve Ronuken”
5.833333 votes to “Dunk Dinkle”
5.250000 votes to “Stitch Kaneland”
4.083333 votes to “Matterall”
4.083333 votes to “ExookiZ”
3.500000 votes to “Sort Dragon”
…

The “cycle” exists in this scenario.

Voter1: a > b > c;

Voter2: b > c > a;

Voter3: c > a > b.

Imagine this scenario in real situation: 3M voters prefer a>b>c and 3M voters prefer b>c>a and 3M voters prefer c>a>b from overall 9M voters.

Here’s another example - 8 voters and 3 candidates:

  • 3 voters: A > B > C;
  • 2 voters: C > A > B;
  • 2 voters: B > C > A;
  • 1 voter: C > B > A.

A > B: 5 voters prefer A to B, B > A:3, A > C:3, C > A:5, B > C:5, C > B:3. Here’s a “cycle” between A > B, C > A and B > C.

Now lets suppose that 8 voters voted for 3 candidates differently:

  • 3 voters: A > B > C;
  • 2 voters: B > A > C;
  • 1 voter: B > C (we are allowed to vote for different amount of candidates in any order);
  • 2 voters: C > B > A.

A > B:3, B > A:4, A > C:5, C > A:2, B > C:6, C > B:2; B > C, B > A and A > C => B > A > C. Candidate B won, because candidate A didn’t won the quota.

So, “voting order” matters and it’s determined by ordinal utility mainly, because votes are deterministic by default in STV systems. Due to Arrow’s theorem and ordinal utility, Sort Dragon with 1500 initial votes lost while Merkelchen with 446, Innominate with 324 votes and Steve Ronuken with 920 initial votes won. An example how the N-Sec Blocs voting. On the other side, Exookiz got a big support from N-Sec Blocs while Olmeca Gold didn’t get a single vote from them. So, entire 0.0 Coalition of 33k + 17k +17k +8k + … players didn’t like Olmeca Gold, but he won! He won just due to the quota and direct votes, while Exookiz won due to significan support from transferred excess votes. Now enjoy an elite speech: "IDK what you are talking about! Look! A WH candidate won! Don’t whine!

Therefore, the Condorcet Paradox - it is a stick with two ends.