Many people treat Condorcet's VSE edge as settled. It isn't. It rests on two assumptions that quietly do all the work: that voters perceive viability perfectly, and that they mostly vote honestly. Relax either and it collapses.

I took Quinn's VSE engine and made one change. His runs let strategic voters read the true standings exactly, so their strategy is always aimed at genuinely weak candidates and barely dents Condorcet. I replaced that with a viability signal that's 50% true, 50% random, since real electorates misjudge who can win all the time (a Democrat written off in a red state, etc.). Now strategy sometimes buries a candidate who'd actually have won, the favorite-betrayal failure cardinal methods are immune to. (STAR voting technically isn't, but voters strategize intuitively would have no idea how STAR actually works, in real life—so just assume it's score voting.)

Then I rigged the rest in Condorcet's favor: cardinal methods (0–5 score, approval, STAR) run at 90% strategic; Condorcet methods get a very favorable 10% and a favorable 40% strategic. Maximum strategy against the cardinal methods, little against the ordinal ones.

Spatial bimodal electorate (two ideological camps, candidates along the axis):

Polya / urn electorate (Quinn's default):

Even with the deck stacked this hard, fully strategic Score beats every Condorcet configuration on both models, including the most favorable. And once a realistic 40% of Condorcet voters strategize, strategic Approval passes Ranked Pairs on both and beats Schulze across the board.

The spatial model, if anything, favors Condorcet: it puts voters and candidates on a real ideological axis where head-to-head comparisons mean the most. Score wins there anyway.

The strategy model is just frontrunner polarization, what plurality voters already do reflexively: "I don't think X can win, so I'll bury them and rank the viable compromise first." Biden > Warren > Trump, Cornyn > Talarico > Paxton. The honesty-and-perfect-viability assumptions are the only thing holding Condorcet's edge up, and neither survives contact with how people actually vote.

(±95% CIs ≈ 0.005 spatial, 0.009 Polya. Built on Quinn's engine with a half-random viability signal; happy to share the patched code.)

@toby-pereira said in score voting is king, condorcet not so much:

It's been said that score suffers worse with one-sided strategy than other methods. That is - if the supporters of one candidate strategise more than the supporters of another, then the result will be more skewed (favour the strategisers) under score than some other methods. Would you agree with that, and do you see it as a problem?

I ran it on a spatial bimodal electorate with a realistic (half-random) viability signal, average strategy pinned at 90%, comparing symmetric (90% of each faction strategizes) against the most lopsided split possible at that level (100% of one faction, 80% of the other).

So at a fixed strategy level, adding maximum asymmetry costs Score about a point, and costs the Condorcet methods more, Ranked Pairs −0.016, Schulze −0.022. Approval is the most robust of all. The ordering of who's hurt is the reverse of the claim: one-sided strategy is harder on the ranked methods here than on the cardinal ones.

I'd agree the underlying property is real, score does respond to one-sided exaggeration. But two things keep it from being a problem in practice. The optimal score strategy is just threshold voting (approve/max everyone above your expected value of the winner), which is close to a sincere ballot, so the distortion per unit of asymmetry is small. And every method has an asymmetric-strategy weakness; the ranked methods' version is burial, which diverges much further from sincerity and does more damage, which is what the table shows.

my latest on this.

@cfrank well thank you very much for reading it. I know it is very long but I think anyone looking to seriously consider and criticize the system should read it so I'm glad you did.

there is no perfect information environment but I'm just saying it's better than any other solution that could possibly be created. and how sequestered you are is completely tunable. in principle you could sequester the jury 3 months ahead of the election or a year ahead of the election or two years agrad of the election if you think it's really that important. I'm not particularly taking any stance on that.

@toby-pereira ok toby, sure.

@toby-pereira

no, i'm saying that an argument about any topic, ethics or dinosaurs or infant formula, depends on the underlying meaning of whatever words you're using to discuss it. and since "ethics" is a meaningless word, similar to "free will", that makes it functionally useless.

@cfrank said in election by jury (www.electionbyjury.com/manifesto):

That does not imply aggregate utility maximization in cases of conflicting interests.

this is extremely simple. genes are "trying" to maximize their expected utility. they've been selected on that basis. therefore our preferences operate on a "maximize it in expectation" basis. therefore from behind the veil of ignorance we just want whatever regime maximizes net utility.

That does not imply aggregate utility maximization in cases of conflicting interests.

you seem to be implying NOT behind the veil of ignorance. but we can't choose a social welfare aggregation function in front of the veil of ignorance, because every person's proposed function will be "i am a dictator", duh.

however, even then, you can mathematically prove the social welfare function is cardinal via reductio ad absurdum. this is one of the most elementary facts of social choice theory.

https://clayshentrup.medium.com/a-simple-proof-that-majoritarianism-is-wrong-5ac15b195b66
https://rangevoting.net/XYvote

It applies only where all affected parties would freely choose an outcome while fully informed.

"fully informed" is confusion. we're talking about actual intrinsic utilities, not instrumental utilities. that's "omniscient" by construction. of course someone might make a bad investment unintentionally. that's irrelevant. the social choice function can only use the information they provide to it.

We are clearly not getting anywhere.

i mean, i'm stating incredibly trivial proofs of things and you're not understanding them.

@cfrank said in election by jury (www.electionbyjury.com/manifesto):

@clay you can plug your ears and say La La La all you like. Kantian ethics is a perfectly legitimate alternative to utilitarian ethics.

brutally false. i completely debunked kant indirectly at wonk.blog/ethics - and the massive gaping holes in his arguments are widely known.

The wonk.blog framework versus Kantian ethics

The axioms are directly contradictory

Preference sovereignty holds that no outcome is unethical if those subject to it would freely choose it fully informed. Kant's Formula of Humanity holds that one can act wrongly toward a person who has consented, and even toward oneself: suicide, selling an organ, consensual self-degradation are all forbidden because they treat humanity merely as a means. Kant needs the moral law to bind independently of inclination; that independence is the entire point of "categorical." The wonk.blog framework holds that any rule capable of overriding an informed preference is thereby refuted. There is no diplomatic middle. Kantian ethics is the purest available instance of the thing section viii declares fundamentally flawed.

Kant reified a heuristic and built a metaphysics under it

The rules-as-instruments reading in section viii applies to Kant more than to anyone. "Do not lie" is a cheap, fast, usually-correct approximation. Kant took the felt apodicticity of that rule and inferred that it must be a synthetic a priori truth of pure practical reason. On the wonk.blog account, the phenomenology of "this is simply forbidden, full stop" is exactly what an evolved fast rule feels like from the inside, because a rule that presented itself as revisable would be worse at its job. Kant mistook the introspective signature of a heuristic for evidence of its source.

The murderer-at-the-door case is the framework's section vi test run on Kant's own commitments and failing. Any rule appearing universal should be checked for whether informed agents would prefer its violation. The person hiding in the house would obviously prefer the lie. Kant knew this and bit the bullet in "On a Supposed Right to Lie," which is about as clean a demonstration as the literature offers that the rule was never intrinsic.

The trolley section is a Kantian argument being dismantled

The recoil at pulling the lever, the doctrine of double effect, the means/ends distinction, all of it is Kantian in structure. The essay's veil treatment holds that the ex ante vote is 5-to-1 and unanimous, so the recoil is a bias about agency and physical proximity rather than a moral discovery. Greene's dual-process work is the empirical companion piece and would be worth citing in any expansion of that section: deontological verdicts track the affective, fast system and correlate with personal-force features carrying no plausible moral weight.

Mackie's queerness argument lands hardest on Kant

Section ix deploys queerness against moral facts in general. Kant's version is the most exposed: a noumenal realm outside space and time, a will that legislates universally, a faculty of pure practical reason that a spatiotemporal brain somehow accesses. Whatever the epistemic access story is supposed to be, it is not one. The essay's amendment to Mackie applies as well: Kant was not asserting a falsehood so much as elaborating an opinion he mistook for a discovery, at extraordinary length.

The parts of Kant that work are veil arguments in disguise

This is the strongest move the framework has available and the one that should lead. Kant's own argument for imperfect duties of beneficence in the Groundwork is that one cannot will a world in which no one helps anyone, because there are circumstances in which one would need others' help. That is positional uncertainty. It is Harsanyi with worse bookkeeping and a metaphysical superstructure bolted on top. Where Kant's arguments reduce to something like the veil, they are plausible; where they do not, they produce the murderer at the door and the ban on suicide. A reduction that explains both the successes and the failures of a rival theory is the good kind of critique.

Autonomy inverted

Kant's central value is autonomy, but Kantian autonomy means conformity to a law given by pure reason, which routinely forbids what the actual agent, informed, wants for himself. By the framework's lights that is heteronomy toward an abstraction. Preference sovereignty is the position that gives actual sovereignty to actual persons over their own lives.

@cfrank said in election by jury (www.electionbyjury.com/manifesto):

@clay Kant does a fair job of that, in direct contest with utilitarian ethics, through his categorical imperative, as one example. If you engaged with ethics as deeply as you claim, you would be well aware of that.

there is no alternative to utilitarian ethics. preference sovereignty principle says you cannot separate "ethics" from "what people would prefer", and preferences are maximized in in terms of expected utility.

@cfrank said in election by jury (www.electionbyjury.com/manifesto):

@clay “real philosophers” say many different things about almost everything. I’ll repeat the main point that you did not address: What you have not explained is why an “ought” proposition cannot likewise describe a higher-order property of a broader configuration involving agents, actions, alternatives, and consequences.

this is word salad. give me a specific concrete definition of "ought".

@toby-pereira said in election by jury (www.electionbyjury.com/manifesto):

Ethics 101 does not determine what a word means.

i don't care if "ethics" does it, it's still obviously true that it does not "mean" anything. and it technically just communicates "i have a preference for this thing", and possibly also, "i wrongly believe that it's objectively true, because i do not realize it's subjective not objective."

A ghost (by some definition) is the soul of a dead person that still in some manner inhabits our physical realm.
It doesn't matter if ghosts are real - this is what it means to be a ghost under this definition

there you've clearly named a specific concrete concept, whether it's real or not. you can draw a ghost even.

but no such thing exists for the word "ought". it has no actual meaning. it's just expressing a preference.

i beg of you to actually spend a few minutes learning what real philosophers say about this. here's alex o'connor saying essentially this: it's just expressing an emotion. it's not an actual semantically meaningful word.

Youtube Video

this is just super elementary ethics and i cannot believe you're unfamiliar with it even after i've repeatedly pointed it out to you.

@cfrank said in election by jury (www.electionbyjury.com/manifesto):

@clay “aggregate utility maximization objectively gives everyone more of whatever it is they subjectively prefer.”
Unless there is a known feasible option that maximizes every individual’s utility simultaneously, with no interpersonal tradeoffs, that is simply false. Aggregate utility maximization maximizes the sum or expected sum of utilities, not each person’s utility.

they are SAME THING, in expectation.

@cfrank said in election by jury (www.electionbyjury.com/manifesto):

@clay, well then, at least we can be highly confident that you subjectively prefer aggregate utility maximization—and, on your account, that neither you nor anyone else has any substantive, non-preference-based reason to adopt that position.

this is nonsensical. aggregate utility maximization objectively gives everyone more of whatever it is they subjectively prefer. you cannot "prefer" something other than aggregate utility maximization as that would be like saying "i prefer what i don't prefer."

@cfrank said in election by jury (www.electionbyjury.com/manifesto):

that does not answer the question. You rejected “ought” because it does not refer to an objective physical law of nature. “Preference” does not refer to one either.

okay, deep breath...

i did not "reject" ought. i rejected the idea that it is anything other than subjective preference.

of course we already know that "preference" refers to "subjective preference" so there's nothing to argue about there.

Granting that preferences are descriptively real, it still does not follow that satisfying or aggregating them is good.

it is good by definition. "satisfying my preference" is what "good" means.

That is the normative premise you keep declining to examine.

there's no such thing as "normativity". there's just subjective preference.

it's almost as if i thoroughly explained this umpteen times and am having to repeat myself.

@cfrank said in election by jury (www.electionbyjury.com/manifesto):

@clay, can you show me any actual objective physical law of nature that the word ‘preference’ could possibly refer to?

i already said utilities are subjective, by definition.

no one is claiming obama is objectively good. we are saying people have subject preferences for different outcomes.

and you can elicit those via revealed preference. e.g. would you take a guarantee of a million dollars, or a coin flip for 3 million.

dear. god. this is not complicated.