The Issue of CSM Elections or the Condorcet Paradox


I know, it’s a game, but I’d like to share with you some scientific knowledge related to mathematics in general and Set theory in particular. Their principles are widely used in statistics and to measure different real social processes, including elections.

We know about CSM’s positive and negative qualities as an ingame institute and about nasty scandals around its members recently. There are more and more players in EVE community over years, which would like to cancel the CSM ingame institute at all. Despite this, CCP is interested to keep the CSM running and more different and good neutral players (non linked to big Null-sec Alliances) decided to participate into CSM14 campaign this year. This means a higher quota level, higher competition, and more distributed votes between candidates.

The CSM election uses the Single Transferrable voting system (PR-STV). This system allows you to order candidates by preference and have your vote move between them as they are either elected, or eliminated, to ensure that your vote still retains impact.

As CCP stated in their voting guide: “VOTE ORDER IS IMPORTANT!”, it’s very true and very important. I’d like to add also that: “THE CANDIDATE GROUPS ARE ALSO IMPORTANT!”, considering that some big Blocs have two or more candidates to represent them. Based on these two principles, vote for your favorite candidate in the first place and share your preferences with addition few candidates representing a group of your part of the game (H-, L- or J-space) ONLY. It will allow you to focus your vote power for a candidate representing your game space and will not distribute your vote, based on how the PR-STV system works, to a candidate which represent an opposite ingame group.

The PR-STV voting system allows to vote for multiple candidates through ranked voting in multi-seat organizations. There are other modern voting systems like STV, BC-STV, SGT. They all are models of democratic voting system and have the Condorcet paradox of voting.

The Condorcet paradox (also known as voting paradox or the paradox of voting) in social choice theory is a situation noted by the Marquis de Condorcet in the late 18th century, in which collective preferences can be cyclic, even if the preferences of individual voters are not cyclic. This is paradoxical, because it means that majority wishes can be in conflict with each other: Majorities prefer, for example, candidate A over B, B over C, and yet C over A. When this occurs, it is because the conflicting majorities are each made up of different groups of individuals.

For example, lets implement social preferences as A > B > C, which means that voters prefer candidate A to B and candidate B to C. Lets 15 voters in total voted next way for their candidates

  • 6 voters for A > C > B;
  • 5 voters for B > C > A;
  • 3 voters for C > B > A;
  • 1 voters for C > A > B.

Therefore, 6 + 1 = 7 voters prefer A > B and 5 + 3 = 8 voters prefer B > A by comparing A and B. 6 voters prefer A > C and 5 + 3 + 1 = 9 voters prefer C > A by comparing A and C. Based on Condorcet paradox, the society’s preferences show that B is better than A and C is better than A no matter that A got most votes, which can be expressed in the form of three judgments: C > B; B > A; C > A, which can be combined into one judgment C > B > A. If there is only one seat then C will be chosen.

The old Condorcet paradox was reviewed and mathematically demonstrated for modern social groups by economist and Nobel laureate Kenneth Arrow in 1951.

In social choice theory, 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.

Both the Condorcet and Arrow’s paradox defines the chaotic nature of the voting system. New voting system models like PR-STV have addition concept ideas implemented into elections, like quota and distributed votes, to minimize its chaotic nature. The PR-STV model underlies the concept of controlled chaos. The majority can have a win condition defined by Quota, while the minority or individual candidates are elected by chaotic serialization.


Therefore, the quota is a core parameter for win condition in PR-STV voting system. In an STV election the quota is the minimum number of votes a candidate must receive in order to be elected. Any votes a candidate receives above the quota are transferred to another prefered candidate. Sources differ as to the exact formula for the Droop quota, but there is a generic formula to calculate it:

Q = (Total valid poll / (Seats + 1)) + 1,

Q result should be rounded to lowest integer number.

To get some very rough values of ingame Q value based on active online players, we have to calculate it when the amount of online players is maximal - 35k. By taking into consideration that maybe 35% of those accounts are Alpha (0.65*35000 = 22750 potential Omega accounts), we can calculate a normalized amount of votes to win the elections.

Q = (22750/11) + 1 = 2069

It’s still possible that some Alpha accounts will sign a one month subscription to vote. The Council of Stellar Management 14th consists of 10 members as was announced previously. Due to the number of candidates, CCP can increase the number of seats to 12, maybe 13. All these unknown for players parameters in the formula like: the amount of active Omega accounts, the amount of players with Omega accounts interested to vote, the amount of “shadow” Omega ALT accounts which will be signed to add consolidated votes, the number of seats, makes difficult to calculate the Q more accurate, but it will be in range between 1300 and 2100. The win condition, in the worst scenario, is to have like (35000/11) + 1 = 3182 or above votes!


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 cast:

  • 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.

The candidates which represent the FW-, J-space, piracy gameplay have not only the quantity disadvantage, but the political disadvantage also. Based on the explained scenario above, nobody from group A and many from group C will not give their 3rd ranked vote for a FW candidate just because they don’t care and for a J-Space or Whaler candidate just because they don’t want extra troubles to be promoted or implemented into game by them. H-Sec players have the same attitude. Therefore, these candidates can win a seat by quota and have very low chance to win by vote distribution. All pirate and PVP players not affiliated to N-Sec Blocs, corporations, alliances should work on a coalition level during these elections, to set a determined voting order for their 2 or 3 candidates only to get a chance to win at least one spot, if CCP will not increase the number of seats somehow.

PS. Data provided by Forex Corp - New Eden research operations.


To he popular voting system it is then…


TL;DR: (of the OP)
The requirements to beat out a large, organized, voting block, with access to one of the largest subbed populations, range from ‘next to impossible’ to ‘minimal at best’.

‘Democratic’ principles rather fall by the wayside when the situation is effectively “We have a democratic voting process here, all the wealthy landowners with Internet access get to cast an e-vote for each person that lives on their plantation. Anyone who doesn’t fit that category is welcome to walk 20 miles to the nearest voting booth. In that manner we look after everyone’s interests.”

It’s not a good analogy, but basically, one group can dominate the voting, and dominates representation on the CSM, and that domination has apparently contributed to the fact that the game has tilted more and more in their favor, to the point where pilots across EVE are running to Null to join them and get in on the fat, safe Null farming, which allows them to dominate each new CSM even more.

So yeah, distill all the math, all the possible ways to theoretically organize an opposition, and the case remains: “Broken unbalanced system is broken and unbalanced, and will remain so, no matter how many times you use the word ‘democratic’ to make it seem fair and objective.”

Tyranny of the Majority , also known as “Mob Rule”.


Interesting post about voting systems. Never heard of the Condorcet Paradox until today, thanks for that! I’m one of the players that doesn’t really care much about CSM but it was an interesting read as a Canadian (our current prime minister promised to change our voting system but decided not to after winning the election).

It seems like selecting representatives that accurately reflect the ideas of the masses is difficult to do. Makes me wonder why CCP does it instead of just having players vote yes/no/don’t care on issues and ideas.


This is actually not something you should even want to do. A game is not a democratic institution, it is a business run according to certain principles. Game design, game mechanics, cash store management, effort/reward balance, game economy… these are all primary game/business issues that you specifically do not want to have influenced by special interest groups and/or people who don’t understand the design tradeoffs involved.

The proper way to go about this is to hire professionals who understand these issues (something CCP will hopefully do someday, instead of simply assuming “Hey we’re still in business, so we must be good at this right?”), develop some ideas, and then run those ideas past a group of testers who represent varying areas of the game, not of the largest power-bloc who can garner the most votes because they have a lot of subbed/plexed/alt accounts.

You want people who are going to tell you what can be exploited and is non-workable in all the areas of the game, not just the primary safe ISK/loot generating system. You most especially want to represent the needs and interests of players new to the game, or those moving from ‘fairly new’ to ‘interested enough to start paying’. That is your entire next generation of income sources, and CCP has been botching their interests for far too long (hence the sizeable decline in player population).


Can I say ‘ahem’, to people who say that you need a large voting block.

Yes, makes it easier. Isn’t required.

One thing to remember for other people: By not voting, you make the vote of those who do, stronger. (as they have less competition).

Vote for people you’d like to see elected. Then put in the people you’d be less annoyed about.

(If the person you want gets elected, a little of your vote will spill over to the next person on your list. If they don’t, well, at least you get someone you don’t loathe)


Instead of staying secluded away from the rest of the entire game, why don’t you highsec people reach out and advertise to pirates and nullsec players and win their votes? It’s silly to only advertise to the handful of people who stay in highsec, completely ignore the entire rest of the game, and then whine about how the numbers aren’t lining up in your favor.

This might be of interest, long but interesting:

1 Like

You convinced me at the first paragraph. I haven’t read the rest.

The only true solution to this problem is to stop having a CSM. The CSM is no longer the instrument it was introduced for (which was already a knee jerk move). It has turned into a clown circus with no actual use other than fortifying the null sec forts and using insider information.

A good CCP with the help of a good CSM would have made very different design decisions over the years, they haven’t. They’re both to blame for the situation in regards to cap proliferation, stagnation and the barren landscape when it comes to different play styles. The proof is in the pudding.


You just missed my “point idea”. Indeed, in a PR-STV voting model there isn’t possible to “beat out” an organization with a domination amount of votes due to the Quota and other factors. This isn’t need even due to democratic concepts - majority must have a vote. It’s even harder to do in a system where the “command dictature” using the democratic voting mechanisms (where everyone will vote as a commander will say) supported by an unknown amount of clone voters (which in RL is called falsification). Indeed, about this is this thread “The Issues of CSM Elections”.

Despite this, we can increase the Quota by voting for different/other top candidates, which will increase the “Condorcet competition” among top groups. Therefore, we can minimize the CSM’s seat monopoly of one organization, as was demonstrated in the OP.

As I stated in OP and as how the economist and Nobel laureate Kenneth Arrow mathematically demonstrated, 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. IT can be called the theorem of impossibility of democratic choice.

This whole set of knowledge is very professional oriented. I read a lot of books about economics, statistics, but I came across these selective definitions by chance from other sources a few years ago. Now I learned how these STV, PR-STV models works in addition. You can learn more about this from Arrow’s book “Social Choice and Individual Values”. The original paper was titled “A Difficulty in the Concept of Social Welfare”.

The Condorcet paradox as Arrow’s paradox, is true as 2 + 2 = 4, as PI = C/d - the ratio of a circle’s circumference C to its diameter d, Einstein’s theory of relativity.

From a philosophical point of view, according to Condorcet, any society contains a fundamentally conflicting state. Since it is impossible to determine the opinion of the majority which satisfied every member of this society.

It’s actually true in general. Ingame things are different due to the available election tools and the nature of voters in particular. A scenario when one true person can vote multiple times symmetrically from his accounts.

Here are my predictions. The CSM14 has 10 seats, if I’m not wrong, and the competition is very high even between big ingame Blocks. Every alliance, with consolidates votes, can vote for their main candidate by granting him a seat in CSM14. It is not even an election, but a bureaucratic routine with a predetermined result. With squre brakets I marked the “candidates for insurance” in case if the main candidate will fail by some miracle. Somebody should take his spot. Below are presented candidates by groups, which will define the voting model of participants:

  • 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.

If all candidates, marked with square brakets, will fail, than L- J-Space groups has only one spot in CSM14. It’s a hard win condition for them. I mean, there isn’t possible to have different spots and representatives at CSM for L-Sec, FW, J-Space individually.

In my opinion, you and Suitonia won CSM13 not due to quota, but because you got a lot of shared votes from different groups: Imperium, PanFam, Legacy and H-Sec industrialists and dwellers. For example, Whalers and Wormholers can’t have this advantage.

Goonswarm can vote next order: Aryth > Innominate > Merkelchen > Xenuria > Steve Ronuken > Vily with their main GoonWaffe playerbase. They can give 4000 votes to Aryth easily supported by their ALT accoutns. This doesn’t mean that their entire 35k playerbase will vote. The Judge can do something also, it depends by how much ISK he has from the CO2 Vault. There are a plethora of harvesters ready to sell their votes for ISK.

Horde isn’t such a political polarized Block as other, but Gobbins still has very good chances to win the quota compettition as Killah Bee has, considering that the North Block hasn’t a consolidated power base anymore. They can go together as Killah > Gobbins > Matterall > Arqui Nurbs or individually Killah > Matterall and Gobbins > Arqui Nurbs. Here new voters coming into play - Winter Coalition, which obviously will not vote for Legacy, Imperium and pirate organiszations, but can support NC/Horde by boosting the quota more than H-Sec can.

In Legacy the situation looks similarly to how it looks in Northern Pandemic, with one exception. Looks like Dunk Dinkle is more motivated and interested to win a seat in CSM14 than any other candidate. Similarly to CSM13, they can give him a spot as they give it to Jintaan even after Creecher was banned. Therefore, we can expect T2 Salvage drones soon and maybe Faction ones.

The most interesting are spots for Sort Dragon and Carbon Alabel. Sort will not have the support from entire PanFam as previously, but, in my opinion, there are some supporting expectations from chinese community in general and Ranger Regiment in particular. After BlueDonut War ProviBloc was stomped very hard. They built their first Keepstar and, I suppose, they’ll try to defend their game play style in CSM now harder than before. I have no idea how ProviBloc members will vote this time.

Most H-Sec player base don’t care about CSM at all. Like half of them are Alpha and other half have mainly one Omega account.

Therefore, L-Sec, J-Space and FW dwellers have a chance to get a spot in CSM14. Their player base should vote based on a predifined model like Olmeca Gold > Mathew Dust > ExookiZ or vice versa. Their main accounts are Omega as their ALT accounts. They can project their votest from: Hard Knocks, Lazerhawks, HLN Sholupen, InnerHell, Snuff, GanBangTeam, Siege Green, Banderlogs, PurpleHelmetedW, PIRAT, FEDUP, Urshakhan, Scary Wormhole People, Bombers Bar. All these organizations doing PVP mainly, piracy, are Wormholers in one or anothe way and are Whalers. I’d say that a big part of russian playerbase, from different parts of space, have this play style nowadays. They can have one representative.

I don’t want to trigger any CSM candidate. This is how it is.

1 Like

As a firm supporter of “We need more destruction, more emergent gameplay, more reason to take RISKS and less ability to bot/farm trillions of ISK safely under supercap/titan umbrellas; oh, and for God’s sake maybe even make the game more interesting/engaging for newer players”, my votes(s) are going:

Olmeca Gold -> Mike Azariah -> Steve Ronuken

I’ll do a little more research to see if any of the other candidates have decent cred in those areas.


Researched more candidates. Currently focusing on those who have a decent blend of good ideas, strong organization skills, clear communication, and proven experience in various expects of the game. My own vote order will be:

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

These candidates have demonstrated high value to the EVE experience in their gaming history.

I highly recommend anyone who does not wish to see yet another CSM entirely dominated by NullSec/GoonSwarm interests to place these names in their voting list. You can vote for 10 candidates per account that you have Omega status on, during the June 10th to June 17th voting period. Make sure your Omega status is active during that window, and please don’t wait until the last minute.

CSM: How to Vote


Human behavior can certainly be predicted in terms of very likely, but not with 100% mathematical certainty.

In the words of the philosopher Björk:

How about… You reach out and advertise to other groups instead of just complaining?

The entire thread reads like a version of “The Cold Equations”.

…which, come to think of it, is very EVE -like in its own way.

Or, when all stop voting we finally get rid of the abomination which is csm.

What if as a minimum there must be at least 1 CSM member that represents each type of space voted in, this could at least balance thing’s a little.

If a single part of space has no one representing them, then that would be a tragedy.

Null / Low / J-Space / High

But then again where do people like Stitch fall, who represent small gang/ solo PvP instead of a single type of space, things start getting complicated :confused: tbh I feel 10 spots might be too little for todays Eve.

I Wonder how many people would bet that all 10 CSM would be representatives of Null Sec.

The voting blocs will get around any such requirement without much effort.

CCP cares only for nullsec, it flounders in its development of EVE in other areas of game play, precisely because CCPs vision for the game was nullsec, the rest of the game is just window dressing as far as CCP is concerned.