Re: POLL RESULTS: What is the action that produces a Verifiable Presentation?

On Sat, Mar 4, 2023 at 10:30 AM Manu Sporny <msporny@digitalbazaar.com>
wrote:

> https://www.opavote.com/methods/recommended
>
> I'm happy to recount using a different algorithm as long as the person
> suggesting it is willing to explain/defend why the recount method is
> better and then interpret the results. There are a number of
> algorithms that we can choose from that would require hours of reading
> to understand why you'd use one over the other, so we'd need someone
> to boil that down for us commoners. :P


I studied, wrote and taught about these “collaborative choice” systems in
mid 00’s.
http://www.lifewithalacrity.com/2005/12/systems_for_col.html

I’m not going to recommend a recount as the information was helpful, and it
was a poll, not a vote. However, I find ranked choice STV variants are not
always ideal for collaborative choice, other than that, they are easier
than other alternatives for non-dedicated voters to do. (I actually prefer
approval voting over ranked choice as they are even easier for voters and
almost as good as ranked choice).

If you have a community that is knowledgeable and dedicated enough to spend
the time to vote, a Condorcet will often yield a better result.[1]

Condorcet basically presents every option as pairs, such as  “present vs
create”. The pairs are all evaluated. This addresses many properties that
the various STV choices can’t do.

It isn’t perfect, as a tie can still happen, which requires another
approach to break the tie. Another problem is too many choices make for an
exponential explosion of pairs (thus my recommendation that it only be used
by dedicated & knowledgeable voters as every pair is a choice, so it can
take a while). I particularly like Condorcet for fiduciary groups like
larger-sized boards (>10), or executive committees, as they have a legal
requirement to be dedicated.

For more details on Condorcet support see OPAVote:
https://www.opavote.com/methods/condorcet-voting

My personal wish is that someone could offer me an optional Condorcet pair
front end for STV that facilitates me to create a ranked-choice list. For
instance, in Manu's poll I only voted for a small number of choices, the
rest of which I ignored. I did not even evaluate “create” and if I had
would have ranked very low as it was too bland and generic.

[1]: Beyond Condorcet being IMHO of the best set of choices given Arrow’s
Impossibility Theorem (see https://ieeexplore.ieee.org/document/6483218), I
feel Condorcet is also particularly good for choices between people as our
socio-psychology often makes other forms of voting for people bi-modal (see
an example of what this is in my article on Amazon 5-star voting
http://www.lifewithalacrity.com/2006/08/using_5star_rat.html) and helps
address that voting about people triggers always voting behavior issues
(sometimes unconciously): https://en.wikipedia.org/wiki/Voting_behavior.
Unfortunately, Condorcet just requires too much work by the voter to be
practical.

— Christopher Allen