Hello Culi,

I love the website. I want to second Robert's suggestion of adding a few
Condorcet methods. Condorcet methods really really really are superior.
They are the only ones that reliably produce a candidate with strong
support that everyone can live with (the Condorcet winner) and a lot of
them are surprisingly robust to strategies or other pathologies that often
creep into election systems.

You mentioned you wanted to add Copeland. Why don't you go ahead? Copeland
is one of the simplest Condorcet methods, and it is even Smith efficient.
It's main downside is that it often produces ties, so you might want to
insert a tie-breaker method (e.g. do Borda or Plurality among the Copeland
finalists). Another very simple Condorcet method is Minimax (very popular,
very easy). If you are interested in putting a little bit more work to get
the strongest possible method, I suggest you have a look at River:

https://electowiki.org/wiki/River

I believe River has the strongest set of election criteria that it
satisfies.

Cheers,
Daniel

On Sat, Jan 22, 2022 at 1:24 PM culitif@tuta.io wrote:

--
Dr. Daniel Carrera
Postdoctoral Research Associate
Iowa State University

(I accidentally Replied only to Daniel and not the mailing list so I'm resending this)

Hey Daniel,

Haha yes, I am well aware of the importance of Condorcet methods and have studied them pretty well. I certainly plan on adding those eventually, but mostly I'm interested in this as an educational tool. It's very much a proof of concept and most of the work that went into it so far has just been on improving the presentation and optimizing the performance. 

I would like to add Ranked Pairs and similarly complex Condorcet methods at some point, but I might need to ask for help with the logic behind those as they can be notoriously difficult to optimize.

Thanks for your suggestion of the River method. I haven't actually given it much of a look before. I'd suggest we make a poll to see which methods people think would be most important to add to the site, but I have a feeling we'd end up stuck arguing about how to conduct that poll before ever actually doing the poll :P

I appreciate you showing interest in the work so far though. It makes me want to continue with the project. I just purchased the domain votevote.page http://votevote.page so hopefully it can have a proper home soon!

Best,
Culi

Jan 22, 2022, 12:19 PM by dcarrera@gmail.com:

On Sat, Jan 22, 2022 at 2:40 PM culitif@tuta.io wrote:

Ha ha. :-)

Yeah, RP is notoriously expensive. It was my favorite method until I
discovered River. RP is very strong and I find it a million times more
intuitive than Schulze, which is its best known competitor. I suspect that
you'll get a lot of buy-in for River in this list. Yes, it is less widely
known but I think everyone in this list would love to fix that. It has
nearly the same intuition as RP but it is vastly cheaper (computationally),
and if I recall correctly, it meets all the criteria of RP and Schulze plus
one more... I forget which one... Maybe independence of Pareto-dominated
alternatives.

So... if you like RP but you find it slow... River is its faster, better,
but obscure cousin.

Yeah, it looks great. Kudos to you. One little detail that I like is that
you say "Coomb's RCV" , "Culi's RCV", etc. As you might know, some of us
are frustrated that the IRV folk have tried to appropriate the "ranked
choice" language, as if theirs was the only election method with ranked
ballots. Honestly, if any method was to appropriate the name RCV it should
be Copeland because it's the older method. Anway, by adding language like
"Coomb's RCV" it sends the message "yeah, this is also RCV, it's a
different (better) one". So it helps chisel away at the "RCV = IRV" idea.

Would you consider replacing "Standard RCV" with something other than
"Standard"? Maybe "IRV RCV" or "Australia RCV" or whatever.

Cheers,

Dr. Daniel Carrera
Postdoctoral Research Associate
Iowa State University

Yeah that's a good point. I just figured most Americans would be more familiar with the term Ranked Choice Voting and I didn't wanna scare them off thinking they were gonna have to learn something new. Anyways, I made the change. Should be up already

Best,
Culi

Jan 22, 2022, 12:50 PM by dcarrera@gmail.com:

Hi Daniel,

Le samedi 22 janvier 2022, 14:50:54 UTC−6, Daniel Carrera dcarrera@gmail.com a écrit :

River in its default WV form is certainly the sort of method that I think is
good. WV gets you the Plurality criterion, and experimentally very low
compromise incentive and few FBC failures (i.e. where compression is an
inadequate strategy to be able to safely vote a favorite candidate at the top).

But I feel like I ought to say, that I am pretty sure that any analysis which
concludes that Condorcet//IRV (or similar) is a good method is not going to
agree that River is a good method. A surface-level analysis will probably show
River and similar methods have a lot of burial incentive.

To me at least the principle of Schulze is intuitive. It's sort of a de-cloning
of MinMax.

One property unique to Schulze(WV), which perhaps I alone find interesting, is
that it always elects from the CDTT, which is the Schwartz set defined using
only the full majorities.

For example, it could be that some candidate "X" has a majority-strength
beatpath to every other candidate, and there are no majority-strength paths back
to X. It could be that River tosses out X's wins because they defeat candidates
already locked as defeated (by an even larger majority). And then X is
subordinated to another candidate via a sub-majority win.

Now does this really matter, probably not, in this context, where the
alternatives are RP(WV) and River, which I think can be trusted not to mess
anything up too badly when electing from outside the CDTT.

Kevin

1. I don't consider Ranked Pairs to be particularly expensive.  How many candidates do you expect to have?  50?  I am expecting at most 5 or 6 in most elections and I support sufficiently strong ballot access laws to keep irrelevant candidates from wasting real estate on the ballot.

2. Culi, I have been involved in some activism to both educate and promote evolving from Hare RCV (what used to be called "IRV" until FairVote appropriated the more general term "RCV", which I consider to be both dishonest and entitled) to one of any Condorcet-compliant methods.  Please take a look at https://tinyurl.com/2tety9tj .

3. For a Condorcet method to make it into practice in a governmental election, the algorithm needs to be expressed in words and needs to have such a description self-contained (not to simply refer to "Tideman Ranked-Pairs" or "Schulze using margins").  It needs to be fully spelled out and a C program, by itself, is not good enough.  That said, simplicity is key, because we are already dancing on the edge even with Hare RCV.  To date, I have fiddled with Bottom-Two Runoff IRV and with a straight description of Condorcet with a contingency of FPTP if there is no Condorcet winner.

If we're gonna do Ranked-Choice Voting, then let's do it right.  Burlington 2009 proves that Hare RCV (or Ware, take your pick) is not RCV done right.

robert

--

r b-j . _ . _ . _ . _ rbj@audioimagination.com

"Imagination is more important than knowledge."

.
.
.

That's an excellent point. The theoretical complexity of RP is irrelevant
in any real world election. In fact, in real world elections Hare is the
one that has a problem because it is not summable.

Culi, are you familiar with Burlington 2009? They tried Hare / IRV and it
failed miserably. There were basically three candidates:

1. The plurality winner.
2. The obvious Condorcet winner that everyone could live with.
3. The dude that IRV picked.

After IRV gave a nonsensical result, there was a strong pushback from the
losers. Since the winner did not have a lot of backing (because...
obviously... IRV chose wrong) the IRV system was rejected and Burlington
went back to FPTP. But now Burlington is going to review the concept of RCV
in an upcoming cycle (I forget when; ask Robert), but this time Robert is
there to explain to the politicians that there is a better alternative. If
he is successful, we have a real chance that Burlington might choose a
Condorcet RCV in the near future.

On Sat, Jan 22, 2022 at 5:52 PM robert bristow-johnson <
rbj@audioimagination.com> wrote:

--
Dr. Daniel Carrera
Postdoctoral Research Associate
Iowa State University

On Sat, Jan 22, 2022 at 5:13 PM Kevin Venzke stepjak@yahoo.fr wrote:

The term "de-cloning of MinMax" doesn't clarify it for me; I honestly don't
understand what that would look like. Schulze might be intuitive if you are
a computer scientist; I know that its algorithm is something similar to
something that computer scientists learn. I am not a computer scientist;
I'm a physicist. I find Schulze simply bewildering. I couldn't tell you
right now how it works without thinking really hard about it, and I'm not
confident I would get it right. I tried to explain it once to my wife (who
is a microsystems engineer) and she looked like a deer on headlights. And
that's someone who is used to me explaining voting methods.

What is the CDTT?

Cheers,

Dr. Daniel Carrera
Postdoctoral Research Associate
Iowa State University

that's overstating it ...

strong enough to repeal with a thin majority.

Unfortunately, Daniel, the Progressive Party pushed the RCV measure through during covid times when we are not able to meet in person in our NPAs (Neighborhood Planning Associations) to really discuss it.  They put it (Hare) on the ballot and it passed 64% to 36% last March.  But in Vermont, any city charter change must be approved by the legislature and governor because a city charter has the status of Vermont law.  (Like Burlington in the past passed a charter change to prohibit people carrying firearms into establishments serving alcohol but Vermont, as liberal as it is, is also a hardcore 2nd-Amendment state. That charter change, while approved by the voters of Burlington, went nowhere in the Legislature.)

Anyway, it's in the legislature now.  I am advocating with the House Government Operations Committee and, providentially, one member of that committee is a friend of mine in my legislative district that I have helped get elected.  And he read my paper (cited below) and "gets it".  This "getting it" is a kinda hard nut to crack because this sequential rounds STV thing is hard to get people to think past.

I just say (repeatedly) "If a simple majority of voters agree and mark their ballots that Candidate A is a better choice than Candidate B, then Candidate B is not elected."  That is the simple ethic.  FPTP has a simple (but mistaken) ethic: "Vote for one candidate. The candidate with the most votes is elected."  Condorcet has a simple (and correct) ethic.  But Hare RCV has a procedure but no ethic other than the procedure.  But that is hard to get around because of all the promotion from FairVote.

I think the narrative what is going on in Burlington is detailed well in https://tinyurl.com/2tety9tj in the Conclusion.  My hope is that the Gov. Ops. committee might decline the RCV charter change because there are some other RCV bills for statewide elections in consideration.  I also have the ear of the Secretary of State and the Director of Elections in the SOS office.  They are very interested in the property #5 in my paper, which is Precinct Summability.  We're watching Maine and we know that they are having an absolute clusterfuck with statewide RCV which prevents them from announcing a winner for a couple days.  Precinct Summability will allow results to be provisionally announced on the night of election, in addition to the process transparency feature that became very interesting to the Director of Elections.

So I am hopeful.  If the Guv Ops declines the RCV charter change and says specifically because they don't want to undo the repeal of IRV in 2010 until we really understand what the fuck is going on, then maybe I can get the Progs in town to listen to me.  But they will not listen otherwise.

--

r b-j . _ . _ . _ . _ rbj@audioimagination.com

"Imagination is more important than knowledge."

.
.
.