Best posts made by Toby Pereira

@sarawolk said in Allocated score (STAR-PR) centrist clones concern:

In any case, I think that Clones are a much bigger problem in hypothetical math scenarios than they ever will be in real life campaigns, and if a faction can really pull off running 2 or 3 clones that all break through and win over voters then that's frankly impressive. The reality is that if voter behavior doesn't do them in, limitations in campaign funding and volunteer power likely will.

I'm not sure I really see this as a problem of clones specifically. If parties exist, then it's fairly normal for parties to field several candidates in a multi-winner election.

But I do think the particular voting behaviour in the example election is a bit "edge case", although it's still best to avoid vulnerability to it.

Well, it partly depends on what you do with equal ranks or incomplete ballots. If an unranked candidate is scored as 0 then a 4-3-2-1 system would be different from 3-2-1-0. But if it's done in a more sensible way, they would be equivalent.

I suppose he's used that assumption because a hereditary monarch is essentially a leader arbitrarily picked, like in random winner (as opposed to random ballot). But this is obviously very simplistic. When you have an all-powerful monarch versus some other system, the entire political and cultural landscape is likely to be very different and that isn't modelled by this.

Yeah, I mean ChatGPT has some big holes in its ability (as I've complained about), but it's also a bit scary what it's capable of sometimes. I don't think anything we can do is off limits in principle to a ChatGPT type thing.

I've been looking at this and I don't think it is the best. One (minor) problem is that when you're summing the scores, for voters that haven't had any candidates elected and also gave a score of 0 to the candidate in question, you get 0/0. Obviously you just need to count it as 0 to get it to work, but it can make one suspicious that there are problems lurking beneath.

But the main problem is that it fails scale invariance. Well it passes in a multiplicative way as it is defined on the wiki, but not if you add to the scores.

For example, if everyone scores 1 to 10 instead of 0 to 9 (so just adds 1 to every score), you can get a different result. KP + SPAV (also known as Sequential Proportional Score Voting or SPSV) passes this. I know it might seem unsatisfactory to "split" the voter with KP, but in terms of passing criteria, it seems to do the job.

@sarawolk I've seen the initial video, but not the 3-hour follow-up! I thought in the first video, the criticism of score was weird.

This method heavily depends on turnout for more accurate scores. What if turnout is extremely low and only extremists turn up to the polls?

Every method depends on turnout for more accurate results. No reason has been given for why score should suffer any more than any other method in this regard.

@cfrank I've seen a few as well which I've deleted, but they're not overwhelming the board or anything, so I wouldn't want to make anything worse for any new users we might get, which isn't that many anyway! So I'd probably say leave it for now, but keep an eye on the situation.

@wolftune I'm not sure this is just an Allocated Score thing. I think that sequential methods that start by electing the candidate with the highest total score and go from there are always likely to have this sort of thing happen. Non-sequential methods may avoid it more easily, but that's obviously more expensive computationally. The other option is to not go by total score.

While people aren't likely to cast votes that are perfectly related to utility, I still see scores as more akin to utility than to something like money, where the increase in utility drops off the higher up the scale you go.

So what I'm saying is that I see a 5 and a 0 as in the same ballpark as a 3 and a 2, rather than the 3 and 2 being preferable for equity reasons.

How good score voting is generally is a separate debate obviously, but where it gives the same tie as a pairwise method, I don't see any reason in principle to prefer one result over another. But as a tie-break, it's probably fine to choose the result you might consider less divisive.

An interesting follow-up question would be whether you would consider divisiveness over score where there isn't an exact score tie (but is a pairwise tie still) or whether it's only useful as a tie-breaker. You could, for example, reduce C's score of 3 to 2. That way, the pairwise result is still a tie but on average scores, A and B are now marginally ahead of C despite being more divisive. Is there still an argument to elect C?

As said, voters can often face a dilemma of whether to approve someone or not. What counts as approval etc. If I approve my second favourite candidate, what if it turns out my favourite could have won after all?

Also under ranked voting, ranks have less of an obvious meaning so a voter doesn't have to feel they are explicitly endorsing a candidate when they rank them over someone else. Say my preference order is A>B>C and B and C are the frontrunners, but I hate both B and C while preferring B to C. I might happily rank A>B>C. But to explicitly approve B might be a step too far, even though it's the strategically optimal vote for me.

Also, it really invites people to say that it violates one person, one vote, and you have to explain why it doesn't.

@cfrank Can you explicitly explain this method, rather than making us infer it from a related Wikipedia article?

@cfrank I've had similar thoughts. There are lots of voting methods that lots of people think are better than First Past the Post, but approval voting is pretty much unanimously thought so among anyone who has any interest in voting methods.

It might not be most people's absolute favourite, but it's probably the most consistent performer. It also has basically no disadvantages relative to FPTP, being just as simple and not failing participation (or monotonicity).

Basically all other methods have something that the FPTP lobby can latch onto as a disadvantage of it.

@matija I'll put a link to the method for reference. I'm not sure either tiebreaker is ideal. Number of 3s doesn't seem to make sense anyway - surely it should be 3s, 4s, and 5s together in that case. I think this seems logical anyway because the approval threshold has moved down to 3, so it's just the most approved candidate at that point. Or maybe the total score from the ballots that count as approving it - so you add up the 3s, 4s and 5s.

I don't know really. It's not a method I've spent too much time thinking about. It's a quota method, and I think they're crude and outmoded.

@jack-waugh Right, I see. Though I think it's no better or worse than what I thought it was. Just arbitrarily favouring voters who approve a particular number of candidates. And it encourages putting up clones or non-entities accordingly.

I've said previously on this group that non-deterministic elections can be a good way to simplify proportional representation. It doesn't have to be as crude as simple random ballot where you have one representative per constituency selected by the drawing of a single ballot. As I said here just earlier this month:

Proportional methods tend to just be more complex by their nature. But if you allow them to be non-deterministic then that goes away. E.g. COWPEA Lottery which uses approval ballots. Or if you have a region that elects, say, 6 candidates, voters just rank their top 6 candidates. Then you consecutively pick six ballots at random and elect the unelected candidate that is highest ranked on that ballot.

This type of method, while it doesn't guarantee a very proportional result in each region, would actually give better proportionality nationally than deterministic methods that use these smallish regions (like STV), and they also keep the election candidate-based, which other nationally proportional methods tend not to.

Random ballot with just one representative per region guarantees that honest voting is the best strategy, but I tend to think that it becomes too lotteristic at that point. With e.g. five or six chances to be elected (as in the above methods), particularly popular candidates would not be on such a knife-edge of being elected.

I also think that non-deterministic methods send out a good message - that there are no "safe seats", and that representing the electorate is a privilege and not some guaranteed right.

So while non-deterministic methods might be a tough sell, I personally prefer them for national parliaments.

@jack-waugh I'm not sure I'm convinced by those specific scores being allowed but not others. Also, there is also Smith//Score, which I would consider to be superior. It elects the score winner of the Smith set. Under Condorcet//Score, the winner could be a candidate outside the top cycle.

It's from way back in 2012, but there was a poll among academics (using approval voting) and approval voting won.

@k98kurz said in Cumulative voting: more popular in corporations than in politics:

@cfrank the main issue with STV is that it is fairly complex, making it somewhat challenging to implement and also to follow the algorithmic logic with any real detail. I read through the ballot tallying report for an Australian Senate election a few years back, and it was awful and tedious -- iirc it was over 60 pages long. By comparison, a cumulative vote tallying report would just be one page of numbers.

It seems that MMP is a much simpler and easier method than STV that gives reasonable results. (Whether the official inclusion of parties is a problem or not is philosophical speculation considering that political parties exist in reality, but that is a separate matter.) Are there any other methods for proportional representation that are simple enough to be both practical and easily comprehensible to concerned citizens?

Proportional methods tend to just be more complex by their nature. But if you allow them to be non-deterministic then that goes away. E.g. COWPEA Lottery which uses approval ballots. Or if you have a region that elects, say, 6 candidates, voters just rank their top 6 candidates. Then you consecutively pick six ballots at random and elect the unelected candidate that is highest ranked on that ballot.

This type of method, while it doesn't guarantee a very proportional result in each region, would actually give better proportionality nationally than deterministic methods that use these smallish regions (like STV), and they also keep the election candidate-based, which other nationally proportional methods tend not to.

Random ballot with just one representative per region guarantees that honest voting is the best strategy, but I tend to think that it becomes too lotteristic at that point. With e.g. five or six chances to be elected (as in the above methods), particularly popular candidates would not be on such a knife-edge of being elected.

I also think that non-deterministic methods send out a good message - that there are no "safe seats", and that representing the electorate is a privilege and not some guaranteed right.

So while non-deterministic methods might be a tough sell, I personally prefer them for national parliaments.

@multi_system_fan It's quite interesting, although allowing more than one round of voting changes things quite a lot, and there are probably also other alternatives that might be as good or better.

I don't think you could apply it to the metric being considered in this thread though.

There are methods where voters rank candidates and also have an approval cut-off, so the winner isn't necessarily the most approved, but is also dependent on the ranks in some way.

But one thing I would say is that with approval voting, I don't see getting over 50% as anything particularly special. If it's OK for a candidate with 51% to beat one with 52%, then it's also OK for a candidate with 49% to beat one with 51%.

Edited to put the numbers the right way round.