Equality by lot

Asset or approval voting =?= sortitional selection

32 Responses

1. 

   Terry Bouricius, on July 8, 2013 at 4:03 pm said:

   Score Voting (also called “Range Voting”) is derived from an anti-majoritarian economic utility approach to assessing the public preference. Since it would be off the topic of this Blog to explain and refute in any detail the claims of activists working for Score Voting reform here, suffice it to say that even if one accepted that voting were appropriate for filling an executive office, when trying to find a single winner from among more than two choices, EVERY voting method INEVITABLY has mathematical paradoxes, flaws, opportunities for strategic manipulation and pathologies …Score voting included. If people are curious about the arguments against Score (Range) Voting, here is the final paragraph of a short essay I wrote several years ago about it:

   “Most people will never be willing to abandon the principle of majority rule in favor of Bayesian regret (the core concept of Range Voting advocates). If 55% of voters prefer candidate A and 45% prefer candidate B in a two candidate race, Range Voting promotes the concept that perhaps B should win if the 45% feel very strongly about their choice, while the 55% are only lukewarm about their choice. This characteristic of Range Voting, the fact that it expressly rejects the notion of majority rule, means that it will never be adopted for government elections.”

   full essay is here:

   Click to access rangevoting.pdf

   LikeLike
2. 

   Terry Bouricius, on July 8, 2013 at 4:14 pm said:

   Note that “Approval Voting” is sort of a simplified version of Range (Score) Voting. Since a candidate that nobody thinks it he best choice (and would come in last place in a traditional election) can WIN under Approval Voting rules, it also has serious problems. A short analysis is here:
   http://www.fairvote.org/why-approval-voting-is-unworkable-in-contested-elections#.UdrmFjuG3Eo

   LikeLike
3. 

   Clay Shentrup, on July 8, 2013 at 4:29 pm said:

   Terry Bouricius is a long-time proponent of Instant Runoff Voting. His link contains a cacophony of misleading arguments. See refutation of pretty much all of it here, from people with advanced math degrees and election science expertise.

   http://www.electology.org/bullet-voting
   scorevoting.net/RichieRV.html

   Clay Shentrup
   The Center for Election Science

   LikeLike
4. 

   Yoram Gat, on July 8, 2013 at 6:27 pm said:

   Mass elections, regardless of the fine details, are an oligarchical tool. By the time the list of candidates is small enough to fit on a ballot the important decisions have all been made.

   See here: https://equalitybylot.wordpress.com/2012/04/21/direct-democracy-and-mass-politics-part-1/ and https://equalitybylot.wordpress.com/2012/05/15/direct-democracy-and-mass-politics-part-2/.

   LikeLike
5. 

   vintermann, on July 8, 2013 at 6:50 pm said:

   Terry Bouricius has been posting here for a while, Clay Shentrup. You haven’t.

   However, you have a long history of pushing score voting to any blog that remotely concerns elections, and not been willing to learn anything. Even on its own terms, score voting falls apart under very modest assumptions of strategic voting (which you work around by assuming “strategic” voting in competing systems – strategies that are not very rational at all).

   Since Arrow, social choice theorists have accepted that even perfectly honest voters can’t accurately compare the strength of their preferences with one another. Thus, the extra information in a rating system contra a ranking system is junk; it can’t make for better decisions, only equally good at best. (Smith’s Bayesian regret experiments, which paint Score so favorably, implicitly assume that voters CAN compare strength of preferences and make lots of other dubious assumptions too.)

   But Terry is right that this is off-topic for this blog, unless you want to discuss the fairly narrow issue what voting systems members of parliament should use to vote among themselves. We are calling for reform of an entirely different sort.

   LikeLike
6. 

   Clay Shentrup, on July 9, 2013 at 2:30 am said:

   vintermann,

   > Terry Bouricius has been posting here for a while, Clay Shentrup. You haven’t.

   This has no relevance to the veracity of our respective arguments. 100% ad hominem fallacy.

   > However, you have a long history of pushing score voting to any blog that remotely concerns elections, and not been willing to learn anything.

   My voluminous writings on election theory quite clearly demonstrate that I’ve learned quite a great deal actually. Perhaps your your definition of “learning anything” means “agreeing with your views”.

   I rather think your criticism applies quite well to Terry and his FairVote colleagues, due to their continued insistence on promoting false and misleading claims like “IRV elects majority winners” or “your best strategy with IRV is a sincere vote”. We have documented quite a number of these types of claims, e.g.
   http://www.electology.org/fact-check

   > Even on its own terms, score voting falls apart under very modest assumptions of strategic voting (which you work around by assuming “strategic” voting in competing systems – strategies that are not very rational at all).

   We can debate what is “rational”, and I’m confidence I’ll win that debate. But aside from that, you seem to ignore two critical factors:

   1) Score Voting outperforms IRV even if we generously assume that 100% of IRV voters are sincere and most Score Voting users are tactical:
   ScoreVoting.net/BayRegsFig.html

   2) Many voters use what is called “naive exaggeration strategy” (e.g. polarize the presumed frontrunners), meaning that it DOES NOT MATTER what is actually “rational”. These voters aren’t even necessarily aware of the mechanics of e.g. IRV. It just seems intuitively obvious that an A>B>C>D voter should vote B 1st and C 4th if B and C are the apparent frontrunners.
   ScoreVoting.net/NESD.html

   > Since Arrow, social choice theorists have accepted that even perfectly honest voters can’t accurately compare the strength of their preferences with one another.

   A) It does not have to be 100% accurate. Bayesian regret calculations specifically account for several sources of noise:
   – normalization
   – tactical exaggeration
   – voter ignorance

   Score Voting still outperforms all commonly proposed alternatives. ESPECIALLY Instant Runoff Voting, which is generally the worst of the five commonly proposed alternative voting methods.

   B) If anything, Arrow actually showed the OPPOSITE. Arrow’s Theorem demonstrates that the correct social utility function cannot be ordinal.
   ScoreVoting.net/BayRegsFig.html
   * Incidentally, I’ve visited Arrow twice recently at his home in Palo Alto.

   > Thus, the extra information in a rating system contra a ranking system is junk; it can’t make for better decisions, only equally good at best.

   This is simply incorrect. There is SOME noise. It is not 100% noise. If it were, that would show up in the BR calculations.

   > Smith’s Bayesian regret experiments, which paint Score so favorably, implicitly assume that voters CAN compare strength of preferences and make lots of other dubious assumptions too.

   I’m always impressed at how consistently the critics of Bayesian regret evince their near total ignorance of how it actually works. Simply put, voters DO NOT “compare preferences”. They convert their preferences (along with ignorance, normalization, strategy) into ballots. Period.

   If you believe the other assumptions are dubious, feel free to cite them. In fact, go ahead and run your own Bayesian regret calculations with your own utility generators or strategy models that you believe are more realistic. Then, if those actually favor a system other than Score Voting, we can debate whose assumptions were more realistic. But I predict we won’t even get to that debate, because Score Voting will still win if you plug in any remotely plausible assumptions.

   > But Terry is right that this is off-topic for this blog, unless you want to discuss the fairly narrow issue what voting systems members of parliament should use to vote among themselves. We are calling for reform of an entirely different sort.

   I suggest Score Voting, or Approval Voting.

   LikeLike
7. 

   Clay Shentrup, on July 9, 2013 at 2:36 am said:

   > Even on its own terms, score voting falls apart under very modest assumptions of strategic voting

   This claim really stands out as being so monumentally unfounded as to warrant a separate response. There has been a tremendous amount of study into the tactical behavior of Score Voting.

   1) There is a theorem that, given plausible models of voter strategy (i.e. voters being rationally self-interested and voting in the ideal way from their own selfish point of view), Score Voting is more likely to elect a Condorcet winner (when one exists) than real Condorcet methods:
   ScoreVoting.net/AppCW.html

   2) Score Voting maximizes the “pleasant surprise” factor.
   ScoreVoting.net/PleasantSurprise.html

   The graph at the top of this page demonstrates the typical fallacy which you are probably making about what constitutes “vulnerability to tactical voting”.
   http://www.electology.org/tactical-voting

   At the bottom of that page are over 10 separate links to further analysis. Quite a lot of this analysis is independent of Bayesian regret.

   LikeLike
8. 

   Clay Shentrup, on July 9, 2013 at 2:50 am said:

   > Arrow’s Theorem demonstrates that the correct social utility function cannot be ordinal.

   The subsequent link should have been:
   ScoreVoting.net/ArrowThm.html

   LikeLike
9. 

   Clay Shentrup, on July 9, 2013 at 3:09 am said:

   > Mass elections, regardless of the fine details, are an oligarchical tool.

   Well, look. Elections have significant consequences. Maine currently has one of the nation’s least popular governors. The Republican governor won because the Democrat nominee and Democrat-leaning independent (Eliot Cutler) split the moderate-to-progressive vote.

   Virtually any reasonable voting system would have produced a different outcome. Even the horrible IRV system, and even top-two runoff.

   It is not helpful to let some hypothetical perfect system be the enemy of what could be a huge improvement. Especially when Approval Voting means an incredibly simple change to rules, and no major change in ballots or counting procedures.

   LikeLike
10. 

    Yoram Gat, on July 9, 2013 at 6:58 am said:

    Clay,

    Spending the energies of reform on yet another variant of elections is misguided. Both theory and experience indicate that electoralism itself is the problem, not this or that particular variant.

    The fact that meaningful reform seems like a long way off at this point is not a good reason to settle for quick fixes that, at best, would yield minor improvements. On the contrary – let’s get started as early as possible and as vigorously as possible on that long way toward democracy.

    LikeLike
11. 

    keithsutherland, on July 9, 2013 at 7:35 am said:

    DG: >I have been focusing on insuring legislative proportionality via use of sortition. But when it comes to the executive, I can only imagine continued use of elections. I wonder if other Kleroterians are of the same opinion?

    I agree that government executives might only be appointed by sortition in a small community where rotation is viable. There are no good reasons for using sortition to select executives in large political communities, as it would be neither democratic nor productive of competent officials. Very few Kleroterians argue for the appointment of government officers by lot.

    But this doesn’t mean we have to resort to election — this is just a consequence of the breakdown of the separation of powers. I imagine that when people vote for a president they are mindful of their likely influence on the legislative process, rather than their competence as commander in chief.

    Why not select government executives on competence, but subject to approval/dismissal by the allotted legislature? If you want to find the best flute players you would not choose them by bean and would be unwise to rely on Pop Idol. Much better to leave the shortlist to the Royal College of Music, but subject to final approval by a representative sample of the people who are going to have to listen to their warbling.

    If there is an ongoing role for election it is for political advocates to set the legislative agenda, although this role could equally well be filled by direct democracy initiative — a competitive process of public petition.

    LikeLike
12. 

    Harald Korneliussen, on July 9, 2013 at 10:50 am said:

    “There has been a tremendous amount of study into the tactical behavior of Score Voting.” There have been you two, Clay Shentrup and Warren D. Smith, who have posted endlessly about the topic on your own site and every blog that’s remotely connected. That’s the tremendous amounts of study.

    You use trivial attacks on IRV – which some voting theorists agree with – to push your own system, which is soundly rejected by just about anyone not called Clay Shentrup or Warren D. Smith. Condorcet systems – which voting theorists pretty universally recommend – you studiously avoid comparing your system with, unless you can cheat by deciding that voters will pursue dumb strategies under it, or deciding that Range voters will be supernaturally informed and selfless (able to compare their strengths of preferences accurately, and willing to limit their influence in favor of those who feel more strongly).

    Your inability to learn you attest nicely in that you have yet to say a word about sortition, which this blog is about.

    LikeLike
13. 

    Clay Shentrup, on July 9, 2013 at 3:18 pm said:

    > The fact that meaningful reform seems like a long way off at this point is not a good reason to settle for quick fixes that, at best, would yield minor improvements.

    Your naive view that these systems would yield “minor” improvements is wildly inaccurate. Bayesian regret figures show that an upgrade to e.g. Score Voting would improve humanity’s lot by as much as the original invention of democracy.

    If you actually look at the effect of democracy on welfare metrics like infant mortality, lifespan, GDP, etc., it turns out that the objective benefit is quite significant. E.g.
    ScoreVoting.net/LivesSaved.html

    Comparing this to other democratic reforms:
    ScoreVoting.net/RelImport.html

    Now, maybe you have some idea that you THINK would yield even bigger benefits. But:

    A) That’s highly speculative. (You don’t have Bayesian regret or any comporable metric to prove how big a benefit it offers, objectively.)

    B) You have no indication of how feasible it is, whereas there’s immense precedent for electoral reforms similar to the one I’m proposing.

    LikeLike
14. 

    Clay Shentrup, on July 9, 2013 at 3:31 pm said:

    Harald,

    Nice ad hominem fallacies. Looking at an example of Smith’s research, we see that he simply did an experiment and then reported what the results were. E.g.

    http://scorevoting.net/RVstrat3.html

    If you want to ignore that because “it was calculated by Warren Smith”, fine. But when you ignore science, you’re likely to get the wrong answers, and you certainly won’t win arguments.

    Smith has actually compared Score Voting to Condorcet methods in considerable detail. E.g.

    ScoreVoting.net/CondBurial.html

    > you can cheat by deciding that voters will pursue dumb strategies under it [Condorcet]

    You are confused. As the CondBurial link describes, we’re talking about negative results that occur when voters use GOOD strategies. That is, good from their individual selfish points of view. There is this basic principle called “the prisoner’s dilemma”, and that’s what you’ve got here. Most Condorcet systems react quite badly when voters do what is ACTUALLY RATIONAL. I.e. not “dumb strategies”.

    > or deciding that Range voters will be supernaturally informed and selfless

    This statement is just bizarre. The voters you say we’re assuming to be “supernaturally informed” are the perfectly tactical ones, specifically in the simplistic theorem about Score Voting producing Condorcet winners. The SELFLESS ones are those who simply vote honestly (i.e. just normalize their honest utilities).

    In our Bayesian regret calculations, you have a broad spectrum of varying amounts of ignorance, and with some voters being honest and some tactical.

    The bottom line is that your entire perspective seems to be based on gross misunderstandings of the material you’re criticizing.

    LikeLike
15. 

    vintermann, on July 10, 2013 at 6:10 pm said:

    No, burial is not a good strategy from an individual perspective in Condorcet, except in rare cases where you can confidently predict that there will be a preference cycle. Then it can work for SOME Condorcet methods. You’d have to see some pretty immense advances in polling for that to be a concern, even then it would only turn up if there was a preference cycle

    There are also some cases where dishonesty is a good strategy, if a losing group can agree on a strategy in secret without their winning opponents finding out. But how realistic is that even in a chamber of 200 members, let alone mass elections? If voters are that unequal in coordination power, you’re in trouble anyway.

    Score is notable in that you have a strategy that beats honesty even when you have NO information about other people’s preferences. Yet in half the results you use to justify Score, you pretend people won’t use that strategy.

    > In our Bayesian regret calculations, you have a broad spectrum of varying amounts of ignorance, and with some voters being honest and some tactical.

    Sure. It’s only that in your calculations, even the “ignorant” Score voters use information they simply don’t have, namely the utility they get from a choice on a scale comparable with other people’s scales. On a plot it’s easy – it’s just the distance, and distances are comparable from voter to voter! But in the real world, people can’t compare their preferences on an absolute scale unless they’re capable of going into other people’s heads.

    Even the “tactical” Score voters voluntarily reduce the impact of their vote for the common good if they use any but the extreme values – and they do in just about all of your simulations.

    With tactical voting, Score looks good because it models Score voters as selfless and Condorcet voters as stupid. With honest voting, Score looks good because score voters use information they can’t have in the real world – preferences on an absolute, comparable scale. Never mind that Score voters have vastly less reason to be honest in the real world.

    But you still haven’t said a word about Sortition. You don’t even appear to be interested in where we might be interested in using a voting system (hint: it isn’t in general elections!)

    LikeLike
16. 

    Common Lot Sortitionist, on July 10, 2013 at 6:58 pm said:

    I originally posted this entry because I — perhaps naively — was wondering if there is any voting system that approaches being an analog for sortition insofar as insuring something close to proportionality.

    As stated, I assume that merit-based positions will always require value-based election. As, for instance, the Greeks elected their generals and treasurers.

    Would the ‘answer’ to my question about proportionality-within-election be something close to what the Amish do? Namely: the community first chooses (informally nominates) a few potential leaders; and then selects from among them sortitionally.

    LikeLike
17. 

    Yoram Gat, on July 10, 2013 at 8:09 pm said:

    Clay,

    > Your naive view that these systems would yield “minor” improvements is wildly inaccurate.

    I find your belief that fiddling with the details of the voting mechanism would yield, or does yield, significant changes in outcome unjustified. But instead of accusing each other of moral or intellectual failures, I would be interested in evidence. If you could provide systematic empirical evidence (i.e., not a set of dubious anecdotes or analysis of artificial behavioral models – Bayesian or otherwise) tying electoral variants with policy outcomes, I would be very interested. As far as I am aware no significant relationship has been demonstrated, but I am interested to be educated on this matter.

    LikeLike
18. 

    Clay Shentrup, on July 10, 2013 at 8:14 pm said:

    > burial is not a good strategy from an individual perspective in Condorcet, except in rare cases where you can confidently predict that there will be a preference cycle.

    This is one of the most common fallacies made by people who are new to election theory. I made it myself for a week or two.

    This argument is analogous to saying that you can only vote tactically in Plurality Voting when e.g. you know that you will cause Gore to win instead of Bush if you insincerely vote for Gore instead of Nader. Well, no. You don’t have to confidently predict that this will happen. You vote for Gore JUST IN CASE. You know that your vote has a higher _expected value_ if you vote for Gore instead of Nader (given, of course that you consider Gore to be significantly better than Bush; expected value is of course probability _plus_ difference in utility).

    This is explained here:
    ScoreVoting.net/DH3.html

    Here’s a page where Warren Smith did trials to calculate the empirical efficacy of simplistic burial strategy with several Condorcet systems.
    ScoreVoting.net/CondBurial.html

    > Then it can work for SOME Condorcet methods.

    No, it can work for ALL Condorcet methods. Some are less vulnerable than others, as the CondBurial link demonstrates.

    > You’d have to see some pretty immense advances in polling for that to be a concern, even then it would only turn up if there was a preference cycle

    This is wrong, for reasons we’ve described in great detail. It would be helpful for you to make your assessment based on facts rather than naive intuition.

    You also ignore actual human psychology. Naive exaggeration appears to be common. E.g. here in San Francisco, we use Instant Runoff Voting. But the vast majority of people I poll (most of them being highly educated people in the software engineering “Silicon Valley” circles I work in) can’t even describe how IRV works. A study by Dana Chisnell, for instance, found that a lot of people assumed IRV worked like Borda. Ergo, they will use the naive frontrunner polarization strategy with ANY ranked system. It DOES NOT MATTER whether it’s actually rational. Warren Smith’s calculations show that it just happens to be rational in many models.

    > If voters are that unequal in coordination power, you’re in trouble anyway.

    We explore the effects of asymmetric strategy here. Score Voting still reigns supreme.
    ScoreVoting.net/StratHonMix.html

    > Score is notable in that you have a strategy that beats honesty even when you have NO information about other people’s preferences.

    True. Now what do you think that proves? Again, you are probably making the fallacy illustrated by this graph:
    http://www.electology.org/tactical-voting

    > Yet in half the results you use to justify Score, you pretend people won’t use that strategy.

    No. Smith’s Bayesian regret figures include the entire range of strategic-to-honest, from 100% tactical to 100% honest.

    I don’t know why you arriving at these strange misconceptions that would be easily remedied if you had actually read the material you’re criticizing.

    > in your calculations, even the “ignorant” Score voters use information they simply don’t have, namely the utility they get from a choice on a scale comparable with other people’s scales.

    Incorrect. Voters only know their normalized utilities. Another basic fact about Bayesian regret that you’d know if you had read even a basic introduction to it, like this one:
    ScoreVoting.net/BayRegDum.html

    > in the real world, people can’t compare their preferences on an absolute scale unless they’re capable of going into other people’s heads.

    We DO NOT assume that voters can compare their own utilities on an absolute scale in the real world.
    ScoreVoting.net/WhyNoHumans.html

    > Even the “tactical” Score voters voluntarily reduce the impact of their vote for the common good if they use any but the extreme values – and they do in just about all of your simulations.

    This is simply false. The commonly cited Bayesian regret figures use a model where a tactical voter is one who uses only the min and max score. At least, I’m 99.9% sure of that. I can verify it with Warren. But only if you can cite whatever indication you saw to the contrary.

    > With tactical voting, Score looks good because it models Score voters as selfless and Condorcet voters as stupid.

    No. There are selfless (i.e. honest) voters, and there are selfish (i.e. tactical voters). And Smith calculated figures for the entire range from 100% selfish to 100% selfish, in gradual increments.

    And, let me explain again, the Condorcet voters ARE NOT STUPID. They are employing a RATIONAL strategy that increases their INDIVIDUAL expected value. It seems like you may be confused about the prisoner’s dilemma. That is, that actors utilizing selfish strategies can collectively be worse off, because each actor’s strategy harms his “opponents” more than it helps himself.

    > With honest voting, Score looks good because score voters use information they can’t have in the real world – preferences on an absolute, comparable scale.

    Again, this argument just shows you have some significant misunderstandings about how BR is calculated.

    > Never mind that Score voters have vastly less reason to be honest in the real world.

    We are absolutely aware that many (arguably MOST) voters will be tactical in real world Score Voting elections.

    ScoreVoting.net/Honesty.html
    ScoreVoting.net/HonStrat.html

    The thing is, even in spite of that, Score Voting STILL outperforms competing systems.

    You have a lot to learn about voting methods.

    And I have a lot to learn about sortition. I’ll weigh in on that subject once I know more about it.

    LikeLike
19. 

    Clay Shentrup, on July 10, 2013 at 8:16 pm said:

    > I find your belief that fiddling with the details of the voting mechanism would yield, or does yield, significant changes in outcome unjustified.

    ScoreVoting.net/BayRegsFig.html

    There is the justification.

    LikeLike
20. 

    Terry Bouricius, on July 11, 2013 at 1:32 am said:

    I would ask that contributors refrain from discussions of the merits and defects of specific voting methods, and that such discussion go to either personal email or some other Blog devoted to voting method reforms.

    However, to respond concisely to Common Lot Sortitionist’s question…Most methods of proportional representation (PR) are party based, and while better than winner-take-all systems, still fail to reflect the full diversity of the population. There are some PR methods that are candidate-based, such as the single transferable vote (STV) method used in places like Cambridge, Massachussetts, Ireland and Australia, (and even a theoretical complicated variant of Score Voting), that are not automatically dependent on party structures (though still gravitate to party forms), which may do a better job of reflecting a communities diversity…but as Yoram points out, these systems require people to run as CANDIDATES, and thus we end up with a skewed legislature that leans towards ego-maniacs and power grabbers.

    There have also been proposals for a lottery voting method… voters cast a vote for a favorite candidate, and then the winners are selected by randomly drawing from this pool of ballots. If candidates of persuasion A got twice as many votes as candidates of persuasion B, there will be roughly twice as many A candidate “elected.” Also individual candidates with many votes have greater chance of having a ballot selected with their name, etc.

    Perhaps the closest “voting” method to sortition, in terms of descriptive representation (though it TOO will favor power-seekers) is Proxy Voting (also called Asset Voting), where those seeking to serve gather endorsers, and win a weighted vote in the legislature based on how many voters have lined up behind them. This system has no elections, or votes… citizens simply form their own constituencies behind a designated representative, and a representative with 10,000 supporters gets twice as much weight in the legislative roll call as a representative with 5,000 supporters.

    After decades of activist work for proportional representation, I reluctantly came to the conclusion that the reform wasn’t adequate, and that only sortition had the potential to make democracy real.

    LikeLike
21. 

    Yoram Gat, on July 11, 2013 at 4:27 am said:

    > ScoreVoting.net/BayRegsFig.html

    You seem to be putting too much faith in this kind of analysis – these results are no more than a reflection of the artificial modeling assumptions. What happens if the five candidates are all pursuing essentially the same policies? This is the real effect of the electoral method.

    I am looking for an empirical comparison of actual policy results in various countries and their correlation with the various parameters of the electoral systems in those countries.

    LikeLike
22. 

    Clay Shentrup, on July 11, 2013 at 2:04 pm said:

    > There are some PR methods that are candidate-based, such as the single transferable vote (STV) method used in places like Cambridge, Massachussetts, Ireland and Australia, (and even a theoretical complicated variant of Score Voting)

    You do understand that Proportional Score Voting (aka Reweighted Range Voting) is algorithmically _simpler_ than proportional STV, right? In fact, the simplest case is just Proportional Approval Voting, which you can conduct in a Google Drive spreadsheet, as I demonstrate here:
    

    Perhaps the simplest proportional method is Asset Voting. You just let the candidates redistribute their votes for some period of time before you call the final result.
    ScoreVoting.net/Asset.html

    LikeLike
23. 

    Clay Shentrup, on July 11, 2013 at 2:07 pm said:

    > these results are no more than a reflection of the artificial modeling assumptions.

    The impressive thing about the Bayesian regret results is that they are actually quite robust against major changes to the models. Warren Smith had knobs to tune the ratio of strategic-vs-honest voters, the number of candidates, the amount of voter ignorance, etc. He also tried different utility generators, including one based on real world ballot data. Score Voting still won, generally by a lot. The results seem sound.

    LikeLike
24. 

    Yoram Gat, on July 11, 2013 at 4:06 pm said:

    Clay,

    Again – you are optimizing over a space of false choices. Even if you showed that your voting scheme manages to choose your “magic best winner”, I would not be impressed.

    Again – the important property of the electoral method – of any electoral system – is that your candidates can all be expected to be representing elite interests and therefore pursuing similar policies, all of which are detrimental to the average citizen. It therefore matters very little which one of them gets elected. Once you limit the set of choices to those 5 candidates, the public has already lost.

    LikeLike
25. 

    Clay Shentrup, on July 12, 2013 at 7:34 pm said:

    Yoram,

    There are numerous metrics that demonstrate a massive benefit from ordinary electoral democracy. Consider infant mortality, life expectancy, and GDP to name but a few. Warren Smith cited some of those here:
    ScoreVoting.net/LivesSaved.html

    I absolutely recognize that there are big problems with the traditional system of “electing leaders”. Studies have shown people vote based on things like looks and personality, which have absolutely nothing to do with policy. And citizens often know very little the candidates they vote for. I read the voter guides here in San Francisco, and I _still_ often feel like I don’t really understand the ballot initiatives the way jurors understand the facts surrounding a court case.

    But the fact remains that voting methods can produce radically better outcomes, even within a fundamentally flawed construct. Many other voting systems would have produced an outcome in Egypt, for instance, which would not have led to all this death, civil unrest, and economic turmoil.

    And consider that Approval Voting requires an incredibly simple change to current procedures. We simply REMOVE the rule that says “your ballot gets discarded if you vote for more candidates than there are winners”. So much benefit for so little cost seems like a no brainer to me.

    Given that we have limited resources, we also have to focus on reforms that have the best EXPECTED value. Expected value entails both actual benefit AND probability of success. With little modern precedent for legislative juries (at least in the USA), it may actually be much more politically difficult/costly to enact this idea. If similar political energy could be spent to remove the vote-for-just-one rule, then it could actually be more optimal to pursue that course of action, despite the plausibly greater benefit of legislative juries.

    We’ve seen numerous organizations adopt alternative voting systems. E.g. the San Francisco Young Democrats just adopted Approval Voting for their endorsement process. Can you direct me to a list of examples of precedent for legislative juries, actually used to make decisions? If we’re going to promote this idea to cities, we have to start with precedent on a smaller scale.

    Regards,
    Clay

    LikeLike
26. 

    Yoram Gat, on July 15, 2013 at 5:34 pm said:

    > There are numerous metrics that demonstrate a massive benefit from ordinary electoral democracy. Consider infant mortality, life expectancy, and GDP to name but a few.

    Actually, I am not sure this is true, and the evidence provided by proponents of electoralism are pretty weak, as far as I am aware. (Certainly, the evidence you offer in the page you referred to is practically non-existent.) the A comparison of India with China, for example, does not put electoralism in a favorable light.

    Be that as it may, however, my query was not about comparing electoralism with other oligarchical systems but about correlating policy outcomes with parameters of electoralism.

    > But the fact remains that voting methods can produce radically better outcomes, even within a fundamentally flawed construct.

    Again – that is exactly what I asked for evidence for. As far as I know, there is no evidence supporting this statement.

    > Many other voting systems would have produced an outcome in Egypt, for instance, which would not have led to all this death, civil unrest, and economic turmoil.

    This is pure speculation, and in any case, an effective oppressive regime could also provide stability – does this imply that such a regime is a desirable one?

    > Can you direct me to a list of examples of precedent for legislative juries, actually used to make decisions?

    I don’t see what such a list of precedents, or a lack of it, would prove, but since you are interested, there are quite a few experiments with sortition in one way or another, compelling in varying degrees. See pages 9 to 11 of Reconstructing Democracy by Cronkright and Pek, the Australian newDemocracy foundation, the work of James Fishkin, among others.

    LikeLike
27. 

    Terry Bouricius, on July 16, 2013 at 3:56 pm said:

    Clay and Yoram,

    On the topic of electoralism, you might find this talk by a Chinese investor and political scientist about the Chinese alternative to electoralism.

    I am not endorsing all his points, but find it interesting.

    LikeLike
28. 

    Yoram Gat, on July 17, 2013 at 5:54 am said:

    Hi Terry,

    Thanks – very interesting talk offering a useful countering view to the electoral dogmatism common in the West. I’ll create a post linking to this talk.

    LikeLike
29. 

    Clay Shentrup, on July 21, 2013 at 4:36 am said:

    > that is exactly what I asked for evidence for. As far as I know, there is no evidence supporting this statement.

    I just cited statistical evidence based on various welfare metrics. Citing a single counter-example does not refute that.

    China, in particular, has some very interesting properties. Various things I’ve read about it portray it as a sort of intellectual meritocracy where the best and brightest are elevated to positions of greater power and influence. I don’t know how accurate that is, but I’ve learned of some interesting things in that vein recently.

    1) This interesting article on Chinese eugenics. Caveat: it was lambasted by some who said it was inaccurate or out of proportion.
    http://edge.org/response-detail/23838

    2) China appears to be gearing up to build a LFTR reactor. If this happens, China would plausibly have the most abundant, cheap, safe, and green supply of energy the world has ever known.
    http://www.telegraph.co.uk/finance/comment/ambroseevans_pritchard/9784044/China-blazes-trail-for-clean-nuclear-power-from-thorium.html

    That’s precisely the kind of complex policy decision in which a highly democratic country can make disastrously wrong decisions due to widespread ignorance of a highly technical field. In particular, the general public’s understanding of the relative safety of nuclear energy is pretty horrendous. So in some sense, you could say that China’s system might end up allowing them to out-compete democracies.

    That doesn’t change the fact that democracies are, on the whole, far better off than non-democracies.

    LikeLike
30. 

    Yoram Gat, on July 21, 2013 at 8:25 am said:

    >> that is exactly what I asked for evidence for. As far as I know, there is no evidence supporting this statement.

    > I just cited statistical evidence based on various welfare metrics. Citing a single counter-example does not refute that.

    I missed that – what statistical evidence did you cite? I am looking for evidence of correlations of policy results with parameters of the electoral system.

    (Again, let’s leave the electoral-vs-oligarchical-non-electoral systems issue aside – I think you are far too cocksure than the evidence would justify, but let’s assume that you are right and keep our focus on variations of different systems within the electoral range.)

    LikeLike
31. 

    Common Lot Sortitionist, on July 21, 2013 at 2:06 pm said:

    Regarding the TED talk on electoralism and China’s one-party system…
    Consider: “Let us never forget that government is ourselves and not an alien power over us.” – Franklin Delano Roosevelt

    When sortitional selection is implemented through legislative juries and Roosevelt’s statement thus finally becomes true, then the glaring absence in the TED talk of any mention of human rights (much less protection of minorities) will, it seems, be the great fear that (pace, FDR) we truly will need to fear.

    So then … In China, the tyranny of the majority, eugenically imposed … preferencing intelligence … leads, doesn’t it, to homogenization?

    Not, then, to defend Western electoral ‘democracy’ but rather to harken to Hitler’s preference for his ‘master race’ and his dismissal of the ‘mongrel’ USA.

    Good thing about the margins, about the edges, the place where things get ‘mixed up’ … that’s where ‘the light comes in’.
    An a-rational place.

    “Imagination is more important than knowledge. For knowledge is limited to all we now know and understand, while imagination embraces the entire world, and all there ever will be to know and understand.” — Einstein

    Seems to me that use of sortition is the only way to assure that the full range of humanity’s imagination gets to play.

    LikeLike
32. 

    keithsutherland, on July 22, 2013 at 9:11 am said:

    Electoral polyarchy and constitutionalism have the best record to date for human and minority rights and David is right to draw our attention to how these might be preserved in a sortitional democracy.

    >Seems to me that use of sortition is the only way to assure that the full range of humanity’s imagination gets to play.

    This is a puzzling claim as, by definition, the play of imagination is restricted to a tiny minority. Referring back to the TED talk, Chinese-style responsive authoritarianism would be more open to innovation, if only on account of the numbers involved in governance (900,000 + at any one time). Li’s charting of political and economic innovations between 1950 and the present day would suggest a system that is uniquely open to imaginative change. In the Chinese system, sortition would play an important democratic role in the responsive part, rather than policy innovation. I’m still to be convinced that sortition could pass the necessary epistemic and democratic tests for policy innovation.

    LikeLike

This site uses Akismet to reduce spam. Learn how your comment data is processed.