@toby-pereira

being a better speaker is already an advantage in our present system. "I don't know what he said. I don't think he does either."

having hours upon hours and days upon days of adversarial interrogation by the other candidates Will take us as far as is physically possible toward minimizing The impact of that bias relative to the impact of the bare facts.

Will this change people? absolutely that's the whole point. we want them to be statistically representative in terms of their intrinsic preferences, not their instrumental preferences. we want the electorate to get what they would have gotten if they had all gone through the same process.

Will this make people tune out of politics? God willing. why should anybody have to care about how their computer operating system is working? they just want it to work. and an ideal world we barely even talk about political issues. we know we have a process that matches policy to our collective well-being as accurately as is humanly possible and we pity all those countries using democracy and getting the kind of chaos we have now. we don't have anti-vaxxers running HHS.

it's hard to say whether the public would react negatively. I haven't been able to find any negative public commentary about election by jury in Georgia. maybe we need to give people a few things they still vote on to address this. kind of like giving your kid a toy steering wheel dor the car.

my latest on this.

for what it's worth, i've decided to keep my name the americanized shentrup and not legally change it back to the original german spelling, for several reasons. so i'll be changing my online profiles over time LOL.

continued here.

The first two voters fail to use the endpoints of the range available to them. Do you expect such behavior in political elections that matter? I think comparisons of voting systems based on hypothetical examples should begin with the voters' valuations of the possibilities normalized. I don't know that it's worth studying examples where voters voluntarily give up power. We as students of these systems need all the time and attention we can put on cases where the purpose of the election is to resolve strong political disagreement among the voters. We need elections to work well when the voters are struggling for political power. Such voters would use the endpoints of the range unless they are not properly informed as to how the system works.

@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).

method 90/90 symmetric 100/80 asymmetric change Score (0–5) 0.943 0.933 −0.010 Approval 0.926 0.921 −0.005 STAR 0.914 0.905 −0.009 Ranked Pairs 0.874 0.859 −0.016 Schulze 0.863 0.840 −0.022

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.

2886b1d8-338c-4f1c-b37b-c3a7b9ff865c-image.png

@psephomancy sorry for tangent, I won't continue discussion here on this.

But quickly: you're first person to emphasize this. "Choose any" only makes sense in contrast to "choose one". I suppose indeed totally out-of-context it could be misinterpreted, and your point is valid, and it does make me less totally confident about the label. Thanks

"The original Borda count and partial voting" by Emerson, Peter.  Social Choice and Welfare; Heidelberg Vol. 40, Iss. 2, (Feb 2013): 353-358. DOI:10.1007/s00355-011-0603-9

explicitly says they are equivalent if everyone submits full rankings, but not equivalent if partial rankings are allowed:

In a Borda count, bc, M. de Borda suggested the last preference cast should receive 1 point, the voter’s penultimate ranking should get 2 points, and so on. Today, however, points are often awarded to (first, second,..., last) preferences cast as per (n, n−1, …, 1) or more frequently, (n−1, n−2,…, 0). If partial voting1 is allowed, and if a first preference is to be given n or n −1 points regardless of how many preferences the voter casts, he/she will be incentivised to rank only one option/candidate. If everyone acts in this way, the bc metamorphoses into a plurality vote… which de Borda criticized at length. If all the voters submit full ballots, the outcome—social choice or ranking—will be the same under any of the above three counting procedures. In the event of one or more persons submitting a partial vote, however, outcomes may vary considerably. This preliminary paper suggests research should consider partial voting. The author examines the consequences of the various rules so far advocated and then purports that M. de Borda, in using his formula, was perhaps more astute than the science has hitherto recognised.

@jan This sounds like it could get complicated. However, it doesn't need to be. If everyone is bringing 10 games each, I don't think there needs to be a vote on that at all. People just bring the games that they want to. Essentially everyone gets 10 nominations into the pool that you vote on.

It might be the case that one person's favourite game is owned by someone else, but they can easily ask to borrow it for their own 10.

As for what to play, if you are playing a lot of games, I would prefer some sort of proportional system rather than a simple single-winner type vote each time. E.g. If 3 out of 5 players generally prefer one type of game and 2 prefer another type, it makes sense for there to be a 3:2 split rather than the 3 choosing every time.

You could simply take it in turns to choose. Alternatively I would recommend a lottery method, so that you would get proportional representation over the long haul without having to keep track of what has already happened. I think simple random ballot, where you randomly pick who chooses each game, would be inferior to taking turns. Some people would get to choose more, just by luck.

But one alternative, which has a balance between randomness and consensus is COWPEA Lottery. Each person approves as many games as they want. Then you pick a ballot at random and if that ballot lists more than one game, you just pick other ballots at random as a tie-break. So if the next ballot approves all or none of the remaining games in contention, you ignore it. Otherwise you eliminate games not approved. Continue until one game is left. If there is still a tie after all ballots have been picked, just pick at random among the remaining games.

@jack-waugh I'd like to find an explicit example eventually if possible. I haven't fully formalized the criterion as I imagine it yet. In terms of the loose "mere existence" criterion, there are definitely examples, since the modification can be applied to any voting system, and if you consider any voting system non-additive, then we can use that one.
For the updated concept, I think both the concept itself and the concept of what it means for a system to be additive need improved formalization. Do you think you can suggest what it means for a system to be additive mathematically?

@robla The feedback I've heard from various stakeholders and reformers across the landscape is that many people oppose narrowing the field too much, which can prevent minor parties from having a meaningful voice in the general or being seen as viable. Being on the general ballot is an important part of the path to becoming viable for a third party.

For that reason, I think there's more consensus around advancing a set number of candidates to the general (instead of setting an approval threshold). Given that the voting method does fine with multiple viable candidates, I think advancing the top 4, 5 or 6 is totally reasonable. The upper limit is set by voter fatigue.

I think two stage Approval is generally the way to go for Approval elections. Approval Top-Two makes a lot of sense for jurisdictions that already have Top-Two, but all other things being equal I'd recommend Approval Top-5 or Top-4. I'm not sure if an even number is better or not, but that's something that should be studied and modeled for each system.

For STAR I generally recommend having a conditional primary, where the top 5 (or 4) advance. And, if less candidates than that register you just skip the primary all together, which saves a bunch of money.

I think it's fine to have it here, but being a US-specific thing, I probably don't have much to say on it!

@poppeacock a Condorcet method is defined in terms of the notion of a “Condorcet winner”, which is a candidate that beats every other candidate in a majoritarian head-to-head match up, also called a “beats all” winner. There can be at most one Condorcet winner in an election; however, there are pathological cases when a Condorcet winner does not exist at all, caused by what are known as Condorcet cycles.

The classic example is three voters using rank ballots over three candidates:

V1: A>B>C
V2: B>C>A
V3: C>A>B

You can see that A>B 2:1, B>C 2:1, but C>A 2:1. So A>B>C>A is a Condorcet cycle, which is a generalized “rock-paper-scissors” situation. Whichever candidate you choose as the winner, there is some majority of the voters who would have preferred a different candidate. That’s the unfortunate thing that happens when a Condorcet winner doesn’t exist…

Regardless, a Condorcet method is any method that guarantees electing the Condorcet winner when one exists. Condorcet methods differ in how they reconcile choosing a winner when the Condorcet winner does not exist, I.e. in effect how they determine which majority group(s) to jilt.

So for example, if Ranked Robin doesn’t specify how it resolves when there is no Condorcet winner, then it’s really a blanket term for Condorcet methods in general. Or maybe it’s a label for a particular curated subset of Condorcet methods.

There are many Condorcet methods, including Ranked Pairs, Schulze’s method, Copeland’s method, Minimax, and Bottom-Two-Runoff (by Tideman).

I like Bottom-Two-Runoff because it’s efficient and also equivalent to a seemingly (but not actually) more robust system: https://www.votingtheory.org/forum/topic/564/bottom-n-and-bottom-2-runoffs-are-equivalent

@jack-waugh I’m imagining the decisions could be automated by making assumptions about the group utility and computing the mixed strategy Nash equilibria of the game.

The deepest problem there is the utility function, since it would be an aggregate utility of some kind that we would have to presume emerged somehow from the individual utility functions of the voters that constitute the group.

All of that is definitely questionable, I don’t believe in utility functions, but I’m unsure how else to proceed in putting this kind of situation into an analytical framework.

@toby-pereira I see, it seems that the properties of maximal lotteries related to participation aren’t what I believed they were. Looking into this, I misunderstood the definition of an (x,y)-improvement, this is swapping the adjacent positions of x and y in a ballot that already exists in the election, not introducing a new ballot where x>y.

That’s unfortunate, but also makes sense.

@cse4129 no problem, that’s what forums are for! Your idea is good, it just happens it was already invented. It also does have some problems/failed properties, though (like any method).

For example, it isn’t Condorcet compliant or Condorcet loser compliant, and it fails participation. It doesn’t satisfy independence of clones either, and it fails various other binary criteria. That doesn’t mean it’s a bad method, but it means it can occasionally produce pathological results. It satisfies the majority winner and majority loser criterion, and later-no-help.

There’s a “mind map” I made some years ago of some of the best-known/characterized voting systems in terms of the binary criteria they satisfy and fail in that discussion above. Probably nowadays I would make a better one (I might at some point). I should have done something like PCA or a graph embedding, but I tried to make that map before I knew about those analysis methods.

@jack-waugh Oh, okay! Thanks for the shoutout! Here's a more stable link to use: https://electorama.com/discord .