🚨 MGRC Voting Mechanism - Important Inquiry 🚨

[jmsjsph]

CC: @zooko @secparam @nathan-at-least @acityinohio

Dear @amiller or whoever this concerns at the Zcash Foundation,

After reviewing all publicly available information (see timeline), it is not clear why the precedent of 5 of N approval voting, set by past Executive Director of the Zcash Foundation, Josh Cincinnati, was not implemented for the inaugural MGRC election.

---

You were asked:

@mistfpga

Why was the switch made to approval voting? it was mean to be 1 cap member 5 votes. Maybe the zfnd can let us know.

To which you incorrectly replied:

@amiller

Its too unwieldy, the instructions were very confusing since you’d have to enter 5x of the same question

After debunking the response by easily implementing the precedent voting mechanism, myself, you followed-up to @mistfpga’s original question:

@amiller

This question of choose max N vs choose max 5, yes both are possible, but choose max N is most consistent with Approval Voting.
(Also imo this entire discussion should be moved to the separate thread as it does not pertain to mgrc diversity)

---

As we know, Josh confirmed the 5 of N mechanism on April 15, 2020, then again on May 14, 2020.

The precedent set by ex-Dir Josh Cincinnati was the voting mechanism that
the Zcash Community was told to expect for the vast majority of the inaugural MGRC election season.

In Josh’s Exit Letter on July 31, 2020, he warned: “consider what is and isn’t being said publicly”.

18 days later on a recorded community call, just 30 days before voting began, the mechanism was changed and considered off-of-the-agenda.

The change went unchallenged by the small handful of viewers on the call, but it is certainly a cause for concern (according to Josh’s, Nate’s and Ian’s language pictured below… plus my very own governance concerns of non-consent and election-engineering.

---

So I must ask,

Why didn’t we implement the voting mechanism set forth by precedent as we were expected to?

Thank you in advance,

For the zpl,

James

---

https://forum.zcashcommunity.com/t/mgrc-update-10-8-2020-and-meeting-minutes/37537/14?u=jmsjsph

[dontbeevil]

I don’t have anything specific to say but I applaud your passion to have the exact answer you’re looking for.

[amiller]

@jmsjsph

it is not clear why the precedent of 5 of N approval voting , set by past Executive Director of the Zcash Foundation, Josh Cincinnati, was not implemented for the inaugural MGRC election.

Thank you for this focused question thread, there’s no problem with digging into question details like this and this is a better way than through comments in other topic threads. Since I volunteered to take over administering the Helios instance for this election, I’m happy to take responsibility for the process and answer any Qs asked.

The vote counting process we carried out is indeed as close to the description of “5 of N approval voting” as we could get. I get that you are specifically asking a very narrow question about the details of the ballot: why were panel members allowed to click up to N checkboxes in the ballot rather than limited to only 5?
That is,
MGRC election “approval, select between 0 and unlimited answers, result type absolute.” (and then we took the top 5 winners, and had defined what we’d do in a tie-break)
as compared with
Zfnd board election “approval, select between 0 and 3 answers, result type absolute.” (and then we took the top 2 winners)

Since you asked the question in terms of precedent, I’ll respond in terms of three principles that apply to this process:

• Quality choice of voting method. I’m repeating my answer you quoted as in your post: the main reason to prefer unlimited choices is that it is closer to ordinary approval voting with multiple winners than limited 5 choices. It reduces the need to strategically rank all candidates, instead you can just vote honestly for all the candidates you approve of.
  I’ll point out here that you already carried out a twitter thread to get external input from approval voting experts, and the only comment you got agreed that this choice is better. You don’t seem satisfied with this answer, but I’m not sure what else to offer.
  In fact, the average number of votes was 578 / 91 = 6.35, quite close to 5, so that’s a quantitative argument that it didn’t make much difference.
• Public comment on process. It’s important that we presented our process in advance, via live stream, blog post, and even by preconfiguring Helios and sharing the link so that everyone could see the questions and settings in advance. One additional reason this is important is to give us a chance to correct any ambiguities in the question wording. So yes that would have been a good time to discuss it, and it’s still a good time to discuss options for future ballots.
• Precedent. It’s good guideline to follow precedent wherever possible, mainly because it makes us more predictable, which in turn helps everyone else participating to follow along with our process and make their own informed responses. Our process was indeed very closely inline with what we’ve done previously, which I think accounts in part for the strong turnout. That said, we’re not obligated to follow precedent 100%, and in fact ZIP-1014 even says we should be continually improving our process. Another improvement in the MGRC structure was that we committed to a tie-break policy at the outset.

Just for the sake of this thread, here’s an informal poll for those reading about how we could carry on:

• I’m glad we we used “select between 0 and unlimited answers”
• I would have preferred “select between 0 and 5 answers”
• I don’t think the distinction between the above two options mattered much for the MGRC election
• We should discuss further about the distinction between the above two options

[kek]

was going to stop posting here, but this has me curious… to the people that voted like this - where in the world is approval voting used? who uses approval voting, and why isn’t approved voting more popular. why isn’t approval voting used in political primaries, etc? why do you believe zcash should use a voting mechanism main stream politics rejects. do we know more about voting mechanisms than the pros?

[BrunchTime]

This is (rightly) apples and oranges. I sure don’t think our community needs to figure out a system of superdelegates.

[kek]

okay cool, but why is it zcash is the only place i’ve ever seen this done? if approval voting is superior to up/down voting - why isn’t approval voting more popular? hardly anybody uses approval voting (including zcash community until recently), and i think there must be reasons behind that. basically, the entire world rejects approval voting for important issues.

[mistfpga]

I would have quite liked to have voted yes on “we should discuss further” but because of the context I didn’t. I really don’t think it mattered that much. I am applying hanlons razor

We have 11 months to work out if this is what we want to do next time and ask the zfnd to change the style if thats what we want

Here is some stuff I dug up on approval voting.

• Three Alternative Voting Methods: Pros and Cons - Giving Compass

Approval Voting

Studied since the 1970s and often cited as one of the best alternative voting methods, approval voting simply asks voters to select all the candidates they approve of and the candidate with the most votes wins. It’s been shown to easily elect a “good” candidate and encourage competition. However, this simplicity comes at a cost that frustrates some, as voters can’t differentiate between how much they like or dislike candidates.

• Ten Critiques of Approval Voting

This is a pretty reasonable argument against plurality voting.

• https://hal.archives-ouvertes.fr/hal-00609810/document

Is this the thread for this though?

[kek]

@mistfpga

Is this the thread for this though?

sure hope so. this is an extremely dry subject to me, and really don’t feel like doing any deep research on this subject. i just want someone to “sell me” on approval voting. hasn’t happened, but haven’t read your links yet.

[sarahjamielewis]

@kek

i just want someone to “sell me” on approval voting.

While not 100% applicable I think this (very) short CGP Grey vid does a pretty good job at selling on the idea of approval voting:

Quick and Easy Voting for Normal People

(Not entirely applicable because we have multi-winner approval voting which means we have more dimensions to play with - namely scoring criteria which I covered and linked some sources for in another thread - we can tune the scoring system to maximize the kind of committee we would like (without running multiple elections).

[nathan-at-least]

Hi there, just saw I was tagged here. Not sure why. Is it for some kind of oversight?

I don’t see myself, nor ECC, as having an explicit oversight responsibility across the two independent organizations, since our only explicit organization is in the TM agreement which is specifically focused on what protocol specification gets to be called “Zcash”.

As discussed on the August 18 Zcash Protocol Hangout video that @jmsjsph links to in the spreadsheet, it sounds pretty clear that everyone there believes an oversight escalation / alarm bell involves convincing the CAP to convene and agree on some resolution.

I’ve seen @jmsjsph’s posts here and elsewhere about the voting mechanism, and it simply doesn’t seem like it rises to a substantial threshold of concern. I don’t personally believe any substantial number of CAP members would find the issue concerning. I haven’t seen almost anyone else express concern (although @kek seems curious to follow up, which seems like a good habit). The poll that @amiller held in this thread seems to substantiate that.

With that being said, even if this isn’t a “fire alarm” level of issue, I personally find it mildly concerning that there was confusion about the voting mechanism. I personally consider the change from precedent (5-of-N to n-of-N) to be very minor and nonconsequential. Also, I personally think announcing that change 20 days before voting is plenty of time, because I don’t believe it affected candidate applications, candidate campaigning plans, voter plans, and only very inconsequentially the vote itself.

For myself personally, I just can’t see how n of N versus 5 of N would make much difference at all, but I haven’t studied voting mechanisms much, so I don’t have any solid argument.

On top of that, I definitely don’t understand how that difference would favor insiders. If anything it seems like the opposite: approval voting would lead to stronger results for less known new-comers, because a voter can decide in isolation if a new-comer candidate is good or not rather than whether or not to spend a limited vote on a new-comer instead of placing it towards a “likely” (ie insider) candidate.

---

PS: since we’re discussing that video, I’d like to point out that I believe the spreadsheet @jmsjsph links to takes quotes out of context in a misleading manner. I recommend listening to the larger back-and-forth between myself and @secparam about the “fire alarm” starting at 49:00. It’s clear that @secparam isn’t talking about the announced voting scheme as “undermining … democratic ideals” but rather a more drastic post-hoc appeal to diverging polls.