mb w/r/t interpretation of rules of debate on the Board, questions of 'fact' lie with the Chairs (altho mb Appeal from Decision of the Chair can throw it to the procedural tribunal?), but questions of 'law' (eg what the rules mean) lie with the procedural tribunal

this would reduce the power of appeal from decision of the chair, and prevent spectacles where the legislature appeals a decision when the decision was clearly correct


need to make the Board election procedure able to deal with constraints such as e.g. at least N independent directors, at least N directors with financial expertise, at least N directors of each gender, etc.


these ppl might be interested in using it:

jeffrey berns and blockchains LLC bought 67k mostly undeveloped acres to make a town which is a "distributed collaborative entity" where "ownership rights and voting powers will be recorded in a digital wallet" according to the nytimes. Berns is a lawyer.

"Every resident and employee will have what amounts to an Ethereum address, which they will use to vote on local measures and store their personal data."

"Mr. Berns believes that one of the big problems has been security. People have been terrible at holding the private keys that are necessary to get access to a Bitcoin or Ethereum wallet.

He wants to address that with a custom-built system where people’s private keys are stored on multiple digital devices, kept in vaults, so that no one device can gain access to the keys. He has already purchased vaults that are burrowed into mountains in Sweden and Switzerland, and he plans to build additional vaults in the mountains in Nevada."

-- [1] (archived)

"After the land purchase, the company made no significant announcements, though it did hire the founder of MyEtherWallet?, Kosala Hemachandra, as its "chief blockchains officer" in June." -- [2]


learn from this:


" Stage c

c ollective decisions In the second part (stage c ), we asked participants to make collective decisions. First, they were instructed to find other members in their gr oup according to a numerical code found in page 1. Each group had six members, and all participants were seated next to each other in two consecutive rows. The speaker announced that there were two possible roles in the group: player or moderator. Each gro up had five players and one moderator. Each participant could 16 find their assigned role in page 1 (e.g., “You are the moderator in group 765” or “You are a player in group 391 ”). Players were instructed to reach a consensus and report it to the moderator in a maximum of 6 0 seconds. Moderators were given verbal and written instructions to not participate nor intercede in the decisions made by the players. The role of the moderators was simply to write down the collective decisions made by the players in their group. Moderators were also instructed to write down an ‘X’ if there was lack of consensus among the group. Groups were asked to answer four of the eight questions from stage i1 (see Supplementary Table 1 ). The speaker read the four questions again, and a nnounced the moments in which time was over. "

" Participants first answered individually, then deliberated and made consensus decisions in groups of five, and finally provided revised individ ual estimates. We found that averaging consensus decisions was substantially more accurate than aggregating the i nitial independent opinions. Remarkably, combining as few as four consensus choices outperformed the wisdom of thousands of individuals "

and the average of the revised individuals estimates (phase III) was even better, i think


"Posner and Weyl do give one example of what I would call a decentralized institution: a game for choosing who gets an asset in the event of a divorce or a company splitting in half, where both sides provide their own valuation, the person with the higher valuation gets the item, but they must then give an amount equal to half the average of the two valuations to the loser. "


tangentially relevant:


defendent decides whether 'jury' (full tribunal) is used, or just judges

or, have a 'trial' first with a panel of 3 judges, and only if they disagree, does it go to a jury

3 judges + 6 jurors = 9; needs 6 to convict


no parliamentary system can survive if it does not provide a mechanism for quick decision-making in times of war and emergencies


if the person who would exceed term limits or serve in multiple offices gets >=2/3 of the vote, they can do this even without the order of a majority of the Chairs. This is to prevent the owner of a company from having control taken away by surprise due to some unexpected procedural issue.


recruit this guy to ask for advice:

Richard A. Arenberg

Arenberg worked for Sens. Paul Tsongas (D-Mass.), Carl Levin (D-Mich.) and Majority Leader George Mitchell (D-Maine) for 34 years and is co-author of the award winning "Defending the Filibuster: The Soul of the Senate-Revised and Edited Edition." He is a visiting professor of political science and international and public affairs at Brown University.


the election protocol for the Doge of Venice is interesting: but see which disputes the actual claim that range voting was used

the Doge seems to be sort of a cross between a king and an executive officer. The governance of Venice is purported to have been pretty good. So maybe this is worth taking a look at.

Seems to me that the primary virtue of this procedure is anti-corruption through randomizaiton; it's hard to predict who the final electors will be so it's hard to get to them ahead of time. Some evidence for this is that the electors were sequestered while they debated. But for similar reasons, it also prevents any one faction from being certain of a win.

how can we use this? i think in the election of Prime Minister (CEO). If we used it on the legislature we would lose the ability to trace and hence to have cross-councils. If we used it on the Chairs they would lose their legitimacy. We could also use it on judges, i guess, but right now i'm thinking just the CEO.

imo the full procedure is too complicated to retain legitimacy in the modern world, and even the reduced 7-round one given in the paper TODO is too complicated. Also it requires too many people.

so how about this variant: the 7-seat Board elects a commission of 7*3 = 21 distinct people, where each person must be elected by a vote of >=2/3 of the Board. The 21 people are reduced via sortition to 7. Those seven elect another commission of 21 distinct people by a vote of >=2/3. And those 21 select the CEO via single-seat triscore voting with a final threshold of 3/5.

When term limits are put on hold this procedure is too (to allow a supermajority shareholder to determine the CEO).

Note which states that the nominees are considered in random order and the first one to get >=2/3s wins (is this only in the final round or in earlier rounds too?).

Unlike just letting the Board decide, this procedure yields a non-incumbent result at unpredictable times, so we need a way to decide when to use it. I say the Board can vote out the incumbent with a simple majority vote. It's possible that this procedure doesn't result in an election, though (if one of the stages fails to elect the required number for the next stage, or if no candidate gets >=3/5 of the final vote), in which case no one wins and the incumbent remains.

hmm.. actually maybe that's not so good; that means that as long as the incumbent enjoys at least >1/3 support, on average their supporters can refuse to vote for anyone in the commission-electing rounds. And then the legislature will just vote to replace again the next day, until randomly we end up with both 7-person rounds having unusually few of the incumbent's supporters. Otoh that outcome isn't actually so bad. (mb let's actually build it in; after we have a selection, the legislature then votes whether or not to accept it by majority vote; if it rejects, explicitly say that the legislature can then vote to go back and do it again?)

This also makes me think that the supermajority votes in the original Doge of Venice elections may have been intended as majoritarian, so as to promote power being kept by whichever group of faction(s) currently holds it (which makes sense as this was an 'aristocratic' system).

I guess a problem remains that in the first round, when the legislature itself is voting by >=2/3s to select 21 people, the >1/3 supporters of the incumbent can just block every time, meaning that this protocol allows the incumbent to stay in office until 2/3s of the legislature is tired of them (which is only 5/7 so it's not THAT crazy...)

We could always make it multiseat triscore in the first round, which doesn't fail (i think) or at least can't be made to fail easily by a small faction.

The thing is, if a majority faction has a candidate in mind, they can keep redoing this until their candidate wins, and if a large enough minority faction likes the incumbent, they can keep refuse to vote for anyone except their own people and thereby block the procedure from electing a replacement. So this procedure only has teeth when (a) a large enough supermajority dislikes the incumbent but (b) there is no majority preference for any single replacement.

Which seems like this is pushing us to use simple majorities rather than supermajorities. So, if any legislative faction has a majority, they can get their way (which seems OK for choosing a prime minister). And when a majority dislikes the incumbent, but there is no majority for any single replacement, now is when this procedure 'has teeth'; the procedure has some probability of coming up with a replacement which is preferred to the incumbent even by the Board, although it's not clear who this person will be.

another issue is that an incumbent friendly commission of electors could simply refuse to end debate. We can deal with that by giving the legislature the power to disolve them at will.

we can better serve the purpose of resisting corruption by making the first commission elected by the Boardelectors rather than the Board (which is more similar to the original Venetian procedure in which a large electorate elected the first commission); in fact, like the Venetians, just start by sortitioning that larger body down to 7.

what if the 7 who are selected happen to be mostly incumbent supporters? They can just refuse to elect a candidate, or to elect someone who is just as unacceptable to the Board as the incumbent. We could add complexity to force them to at least elect someone, but i think this is a lost cause -- if they like the incumbent, they like the incumbent, and they'll find some way to do what e wants. So let's just let them do their thing -- forcing them to filibuster, or to elect someone else besides their guy, is 'punishment' enough for not doing their "duty" of honestly searching for a good, broadly-supported, replacement for the incumbent.

if the office of CEO is vacant, then whoever is currently filling in for that role (acting CEO) (the understudy of the previous CEO) is in the role of incumbent in the final vote, and if they win, they become the real CEO. If the office is truly vacant (there is no one even filling in, and it's totally unclear who should), then the first Chair acts as the incumbent, but if separation of powers is in effect, they remain only an acting CEO, and in this case the CEO replacement procedure is automatically invoked every Boardmeeting until a real CEO is found.

so the procedure i'm currently advocating is:

the Board can vote to challenge the incumbent CEO. When they do so, 7 out of the Boardelectors are chosen by sortition. They elect a temporary commission of 21 seats via multiseat triscore (with a final runoff vote of the chosen replacement against debating more and then redoing the vote). That commission, after debating, elects a potential replacement CEO by single-seat triscore (with the incumbent not eligible) (with a final runoff vote of the chosen replacement against debating more and then redoing the vote); note that this procedure might not terminate, so if these commissions haven't elected a potential replacement CEO by the beginning of the next Boardmeeting, they are dissolved. This potential replacement is then sent to the Board, who vote in their own runoff between the incumbent and the potential replacement. The Board may re-challenge the incumbent (or new) CEO once per Boardmeeting. if the office of CEO becomes vacant, the previous CEO's understudy immediately becomes CEO. However if there is no understudy and the office becomes vacant, the Chair becomes acting CEO and the CEO replacement procedure above is automatically invoked every Boardmeeting, with Chair as incumbent, until a CEO is found (if separation of powers is not in effect then a vote for the incumbent by the Board makes the Chair the CEO).


what are the advantages of the above procedure?

with this expansion of their role, we might change the name of the Boardelectors to just Electors. Which sounds better anyhow.

also, i forgot how many Boardelectors there are, but we should make the commission number the same as that, instead of just 21.


maybe a simpler variant would be:


yknow we could actually use the board-tracing system to have local law; when there is a civil suit between two citizens of the same traced faction, their jury is only composed of others from the same faction. And you could let each faction make its own rules that only apply to its citizens.

One issue here is that each citizen is actually in two factions at once due to the two-term structure of the Board. Do these additional rules only apply when the same rule is passed by both factions? Yeah, probably, because it's already annoying for citizens to have to obey these 'dynamic factions' that they didn't directly join.

A simpler thing would be to do the jury-only thing without the additional rules. Does the jury-only thing only trigger if both citizens match on both of their factions? Yeah, probably.

When two citizens are of different factions, we could still select the jury so that it only consists of other citizens from those two factions? i think that's a bad idea though; in that case everyone on the jury has a 'side'.


not even precedent can override plain meaning of the rules even if the rules were not followed in the past precedent does not compel or allow them to be disregarded


"Britain’s brilliant method of picking Supreme Court justices, explained"

skimmed this but it probably deserves a read


interesting scifi take on fundamental rights:


" When, finally, during the Kronstadt rebellion, the Soviets revolted against the party dictatorship and the incompatibility of the new councils with the party system became manifest, he decided almost at once to crush the councils, since they threatened the power monopoly of the Bolshevik party. The name 'Soviet Union' for post-revolutionary Russia has been a lie ever since, but this lie has also contained, ever since, the grudging admission of the overwhelming popularity, not of the Bolshevik party, but of the soviet system which the party reduced to impotence. ... More recently, his- torians have pointed to the rather obvious similarities between the councils and the medieval townships, the Swiss cantons, the English seventeenth-century 'agitators' - or rather 'adjustators', as they were originally called - and the General Council of Cromwell's army, but the point of the matter is that none of them, with the possible exception of the medieval town, 7 * had ever the slightest influence on the minds of the people who in the course of a revolution spontaneously organized themselves in councils.

Hence, no tradition, either revolutionary or pre-revolutionary, can be called to account for the regular emergence and re- emergence of the council system ever since the French Revo- lution. If we leave aside the February Revolution of 1848 in Paris, where a commission pour les travailleurs, set up by the government itself, was almost exclusively concerned with ques- tions of social legislation, the main dates of appearance of these organs of action and germs of a new state are the following: the year 1870, when the French capital under siege by the Prus- sian army 'spontaneously reorganized itself into a miniature federal body', which then formed the nucleus for the Parisian Commune government in the spring of 1871; 75 the year 1905, when the wave of spontaneous strikes in Russia suddenly de- veloped a political leadership of its own, outside all revolution- ary parties and groups, and the workers in the factories organ- ized themselves into councils, Soviets, for the purpose of repre- sentative self-government; the February Revolution of 1917 in Russia, when 'despite different political tendencies among the Russian workers, the organization itself, that is the soviet, was not even subject to discussion'; 76 the years 1918 and 1919 in Germany, when, after the defeat of the army, soldiers and wor- kers in open rebellion constituted themselves into Arbeiter- und Soldatenrate, demanding, in Berlin, that this Ratesystem be- come the foundation stone of the new German constitution, and establishing, together with the Bohemians of the coffee houses, in Munich in the spring of 1919, the short-lived Bavarian Rdterepubli\'^ the last date, finally, is the autumn of 1956, when the Hungarian Revolution from its very beginning produced the council system anew in Budapest, from which it spread all over the country 'with incredible rapidity'. 78

The mere enumeration of these dates suggests a continuity that in fact never existed. It is precisely the absence of con- tinuity, tradition, and organized influence that makes the same- ness of the phenomenon so very striking ... It is true that wher- ever the revolution was not defeated and not followed by some sort of restoration the one-party dictatorship, that is, the model of the professional revolutionary, eventually prevailed, but it prevailed only after a violent struggle with the organs and insti- tutions of the revolution itself. " --

"Odysse Barrot formulated with rare precision the chief difference in terms of French history between the new form of government, aimed at by the Commune, and the old regime which soon was to be restored in a different, non- monarchical disguise"; then a French quote which translates to something about federation rather than 'one and indivisible'

" the intimate connection between the spirit of revolution and the principle of federation. In order to prove what Odysse Barrot felt to be true, we must turn to the Febru- ary Revolution of 1917 in Russia and to the Hungarian Revolu- tion of 1956, both of which lasted just long enough to show in bare outlines what a government would look like and how a republic was likely to function if they were founded upon the principles of the council system. In both instances councils or Soviets had sprung up everywhere, completely independent of one another, workers*, soldiers', and peasants' councils in the case of Russia, the most disparate kinds of councils in the case of Hungary : neighbourhood councils that emerged in all resi- dential districts, so-called revolutionary councils that grew out of fighting together in the streets, councils of writers and artists,


hmm.. a person both lives in a neighborhood, but also works.. my 'constituencies' get part of the way there because people are free to self-organize into them, rather than top-down drawing fixed geographic districts; but i had previously thought that each citizen could become a constituent of only one 'constituency' -- but perhaps this should be generalized into a system where each citizen has one 'vote' that they can fractionally allocate between as many constituencies as they like.

most people have a home, a workplace, and a political affiliation, so people should be able to split their vote at least 3 ways.


some ideas:


the thread starting with had a few good points on 'loser pays' court systems:


consider adding a third legislative house, citizen assemblies/deliberative polling either by sortition, or stratified by board tracing (which means it only can include the first layer of delegates)

now something passes if it passes a majority of those houses which choose to vote on it

a new deliberative polling group is convened whenever the other two houses disagree. there is also a standing deliberative polling group that can create its own proposals, or take up and vote against a resolution passed by another house.

in the past, similar efforts seem to have involved groups of 32 to 200 citizens, chosen in a demographically stratified manner. i would have much smaller groups, 7 to 24, chosen either randomly or using board tracing.


yknow, i'm thinking of moving to a system with 1 Chair and 3 co-CEOs. Not as good for vision but better for balanced execution -- with 3 ppl instead of 1 it's less likely that one person's biases will cause them to miss something imoprtant -- every now and then you see a CEO make a really boneheaded decision because they blew off something important. Also, as Dana pointed out long ago, if we want the Chair to be the most respected official, that doesn't work so well if there are more Chairs and fewer CEOs.

this proposed change makes the system simpler because, instead of some Chair powers being individual and some by majority vote and some by unanimous vote, all co-CEO powers can be by majority vote. It also makes the initialization procedure simpler, because the proto-Chair can just be the actual Chair.

(In fact it could be made even simpler -- if two CEOs agree, the third one may not even have to be consulted (not sure if that's a good idea though -- requiring a formal vote means that at least the third one gets notified about what's going on as soon as possible))

the downside is that it doesn't allow for 'visionary' CEOs, and related, a single startup founder wouldn't adopt a system that requires em to work through a committee of 3 subordinates. I guess my answer to that is, the single founder would control (or at least be voted for by) a majority of the votes in the governance system, and could occupy multiple seats at once -- so the board could elect that founder to 2 seats at once, in which case they are effectively the only CEO. So really what it says is that you can have 1 CEO if the faction supporting them has 2/3s of the votes.

another downside is that this is now different enough that you can't really see existing corporations or other groups switching to it without them really being true believers.

another downside is that, by that theorem/example (all agree that A and B imply C, and 1/3 believes only A, 1/3 believes only B, 1/3 believes A and B and C), now the co-CEOs, as a group, have logically inconsistent beliefs

still, at this moment, all things considered, i like it.

dunno what the techical name for a position with 3 co-holders is. Trimumvirate? Triarchy? Troika? Triune? Triple? Tripartite? Triadic? Triad? Trio? Troika is shortish, but it has an association with particular historical institutions. Maybe just don't refer to this by any name. Or maybe Three Kings, Triple CEO, or Triceo (sounds kinda like Triceratops!). Kinda goes well with the name Triscore Voting.

Actually trio and triad are both pretty short.


two more downsides. First, everything might take longer, if the 3 co-ceos insist on having a discussion before every little decision. Second, it's hard to negotiate or to work under them, because if you are talking to just 1 of them, they could promise something that the other two don't agree with; in theory if two of them promise you something you are good, but in that case, if the third doesn't agree, your promise won't be upheld if only one of the two promisers defects. Similarly, it's harder for the board to hold them accountable, because every now and then there will be some byzantine power play that will cause the alliances between co-CEOs to switch and therefore the joint policy of the co-CEOs to change, and each co-CEO will try to shift blame onto the time they weren't in the ruling pair for failures.

and one more upside. The separation of negative power from positive, and the parliamentary (vs presidential) system of having the CEO a creature of the board is sort of all part of the goal of reducing the power of the most powerful single actor (the CEO) so as to provide less of a lever for dictatorial behavior. Splitting the CEO position into 3 sort of drives a stake through the heart of this; there is no longer any single position which provides a structural platform for dictatorship -- the only remaining unitary role in the whole system is the Chair, who has only negative and parliamentarian power (and most importantly, the legitimacy that comes with the popular vote). (and i guess the downside of that is that, in governments at least, the positions of PM/President provide a honeypot for some people who in earlier times would have raised armies to take over the country).

and one more downside. The rest of the world probably isn't organized to deal with 'multisig' arrangements like this one, so e.g. banks will want to know 'is this person authorized to sign on behalf of the company?' and may not be setup to handle the answer 'no, but any two of these three people can'. For these purposes you could always allow (or even require) the co-CEOs to appoint a 'COO' who is a corporate officer in their own right (but who is subordinate to the co-CEOs and whose decisions can be overruled by them); most organizations large enough to use this system would want a COO anyhow. In that case, is the COO just the real CEO and the co-CEOs are just an executive committee? Not really, the difference is that the co-CEOs are envisioned to be involved in execution (in 'the thick of it'), not just policy-setting.

One thing to wonder about is: if there are 6 board members, 3 in each class, surely many organizations will adopt a tradition of the 3 senior board members being co-CEO. This is disappointing because now professional 'politicans' are being placed in an important 'management' position. But that's unlikely to happen in traditional corporations (because the board members don't want the full-time job of co-CEO), and it's no worse than the status quo in governments (where politicians end up being the PM anyhow). Also, surely often what will happen is that the leader of the most powerful faction will form a durable alliance with the leader of a smaller faction such that they are both CEOs and both always vote together in whichever way the powerful leader wants; in this case the co-CEO structure is effectively as if the powerful leader were a single CEO (because the opposition co-CEO is always outvoted and hence is of little consequence, although as noted above at least they get some operational visibility into what's going on). The co-CEO structure still has some impact though, b/c this dictatorial arrangement requires the dictator to keep control of the allied leader, which requires em to keep the alliegence of the allied leader's faction; so it still requires the dictator to maintain a strong level of consensus support, and of course, if a dictator has strong enough consensus support, no structural impediments can help in any case. I don't suppose this 'allied puppet' arrangement could possibly occur if the would-be dictator's faction plus the allied faction has less than 50% of the Board (because otherwise the other 50+% would get 2 co-CEOs among them), so compared to the one-CEO status quo (in which case the would-be dictator would probably simply be the CEO), this is not more dictatorial.

co-CEOs remind me a little of the 5 Spartan Ephors, although this analogy is of little use because i can't find much on the internet about whether that governance arrangement was thought to have worked out well.

one thing to keep in mind: when thinking about the increased time spent by the co-CEOs having discussions/negotiations amongst themselves, i like to imagine "hey, they can just have quick votes and be done with it". But that's not human nature for many people; their feelings will be hurt when another co-CEO disagrees with them and rather than briskly and coldly move on, they may demand time spent smoothing things over; this dynamic will also lead to many co-CEO teams making an effort towards unity and consensus decision-making. So the deadweight of the increased discussion time may be much larger than i predict. Otoh at least some of this time these discussions will lead to better understanding of the issues and will lead to the decision quality gains that i'm hoping for, by preventing co-CEOs from blowing off important issues.

people will say that oh, the best CEOs won't want to be a co-CEO, so you lose out in attracting talent. I agree that the people most suited for a single CEO job might not want to be a co-CEO, but it's an open question whether better decisions are made by the best willing single CEOs, or by the best willing co-CEOs.


the multiple offices and term limits seem good for most purposes, but in pseudononymous groups a person could maybe evade the multiple offices thing just by getting a second pseudonym. Still, it's a hoop to jump thru, and also other voters might not respect the new pseudonym as much, so it might help.

however it's annoying to have that extra question on the ballot about allowing these to be evaded. Maybe you could replace this by just having higher voting thresholds to evade it. Like, if you are being elected to role X more than Y times, or if you win role X and also role Z, then you don't actually get role X unless you have a supermajority of support, or something like that.


if the Chair is also elected co-CEO, then they could keep their Chair position, but the choice of co-CEOs would have to be ratified by a 2/3s vote of the Board (which is 5/7). If it fails ratification, then a new slate of co-CEOs must be selected, with the Chair excluded from candidacy. Similarly if the same person scores two co-CEO seats, 2 Procedural Tribunal seats, etc.

If a person becomes Chair while already holding other positions, they must pass a 2/3s vote for each of those other positions to keep it.


one way to simplify the whole system would be to just cut out the Boardelectors

so then elections become simpler:

ballots just have two things:

The Board delegates delegate in rounds. At the beginning of each round, the average number of votes held by Board delegates is noted, and end of each round, any remaining delegates with less than the begin-round average are eliminated and then the next round starts; unless less than 3 or equal to 3 would be left after elimination, in which case, the top 3 win. The 7th Board seat is randomly chosen weighted by number of original ballots naming that Delegate.

Initial ballots are secret, but delegations by Delegates are public.

Chair candidates are selected by signature campaigns. Candidates are listed on the ballot in descending order of signatures collected. Only the top 7 candidates (by # of signatures collected) appear on the ballot.

Chair voting is:

Boardmembers may be expelled from the Board by a 2/3s (weighted) vote of all original Delegates (or mb just: all original Delegates who did not indirectly delegate to them!), in which case, the original Delegates that did indirectly delegate to them hold a new round of delegation to replace them.

(i guess adding back in a layer of "Boarddelegates" would not be that complicated here; you could use 23, as i did before, or you could just say that the penultimate set of Boarddelegates, that is, just before the second-to-last elimination, are the "Boarddelegates" -- but i'm not sure that it's necessary -- otoh it may make the recall process more deliberate)


a simple majority of the Board could Appeal a Decision of the Chair to the Procedural Tribunal -- or maybe a supermajority? no probably just a simple majority; the Tribunal may not agree, after all, especially if we say that the Chair's decision is like a defendent, so it takes a supermajority to 'convict'


had a discussion with an experienced executive on the 3 CEO issue. They think it would be a problem, and the biggest problems would be (my wording, not theirs):

Other lesser problems would be: waste of the personal time of the co-CEOs; confusion on the Board due to the conflicting narratives presented by the co-CEOs, which would be intentionally skewed more than with a unitary CEO due to the power struggle.

They think 99% incidence of power struggles will be observed. As for the impact of these disadvantages, on a scale of 1 to 10, where 1 means that the co-CEO innovation will not cause any noticable problems at all, and 10 means that an organization which otherwise would have been wildly successful would fail terribly as a result of this innovation, they put the impact of this at 7.

So that sounds pretty bad. So maybe i'll abandon this idea.


i sent them this email:

a modification to the idea:

So when the Board is unified, there can be a single CEO (assuming the unified Board wants this), and this is has the same outcome as the traditional single-CEO system.

But when the Board is divided into factions, there are 3 CEOs. Contrast with the traditional single-CEO system's solution to a divided board: the traditional system creates winners (the Board members in the plurality faction) and losers (all of the other Board members), and gives all (executive) power to the winners and no power to the losers. In the proposed system, instead of the plurality faction getting everything they want, on various specific issues various factions get what they want, depending on which 2 CEOs ally on that issue.

So in essence, when there is disunity, rather than resolving that disunity at the Board level, the disunity is propagated down to the CEO level. The benefit is that all factions on the Board share influence on executive decision-making, rather than creating winners and losers. The cost is that the disunity now infects one level below, causing the costs we discussed (e.g. politicking between the 3 CEOs).

The idea can be modified by adjusting the threshold for 1 CEO to be permitted (or otherwise changing the criterion).

they replied:

> You have a point. Augmented by the fact that sometimes, under one CEO, if Board divided, they essentially make a strategic decision without complete information. The faction that differs from the side that the CEO favors does not have a working CEO representing that point of view. So they are dealing a little blind.

so mb do it the way that i said there.


had a short discussion with an experienced admin. Asked them if they would rather the US have 1 president or 3, they said the more the better (so that they can check each other, it was implied). Asked them what about in a company. They said it doesn't matter, the CEO is a figurehead and the VPs make the real decisions.


one thing we could do is say that by DEFAULT, there are 3 co-CEOs, but with a 2/3s vote the Board can set a single CEO. This makes sense because if 2 of the 3 co-CEO slots were held by the same person, that person would effectively be a single CEO anyways (although the other co-CEO would effectively have 'observer status' at least). I like this because you can still get rid of the 3 co-chairs and then you have the possiblity in a non-polarized group of having one chair and one CEO, which both 'scales down' and also fits existing practice. But i dislike it because i think the possibility of getting rid of your co-CEOs entirely will lead to more power struggles between co-CEOs. Also, in a group with a lot of discord, wouldn't you prefer 3 co-chairs over 3 co-ceos? Actually i'm not sure, maybe you would prefer 3 co-ceos...

or, instead of a 2/3s vote, could just say that if the proportional voting system elects the same person to 2 seats

or, could say that if the first winner has the 'support' (meaning a non-zero on the 0,1,2 ballot, or a non-minus-one on the -1,0,1 ballot) of more than 2/3.


mb if there is a tie in the board election for CEO(s), the incumbent(s) stay (mb a tie in ANY of the 3 seats if 3 CEOs)


should look into this b/c this guy likes it :

looks a bit random to me... and the wards are strange, although mb could use my constiuencies/asset voting to make the wards...

here's another one that guy seems to like:

here's a crazy one (that i found through Google):


two tweaks to triscore voting: