Grant Round Improvements Discussion

Secured
#61
2. With the 0-100 scoring, with the above issue removed, the problem is the subjectivity, ESPECIALLY when you have as many votes as you want. This is part of the reason I suggested giving each standing party X votes and letting them carry forward to the next round. People liked that conceptually but they were worried about how it might complicate things which is fair. So let's be more conservative and take one step forward instead of two. We create a formula where each Standing Party gets X votes just for that grant round. For example, let's say there are 10 grants. Each Standing Party would get 40 points for each grant they have a right to vote in. So if I had no conflicts of interest, Factomize would get 400 points to award. We don't have to award them all, but they don't carry forward.
This solution opens a chance for pre-approved grants (as Guide Pay) not to fit into grant pool size :)
E.g. De Facto has 40 points, how many votes should be placed for Guide Pay? 3? 5? 10?
In strong grant round good projects may receive many more votes, so pre-approved grant like Guide Pay or something else won't get into awarded list.
 
Secured
#62
And I don't see an easy fix of this issue.
So maybe it's really better to move to binary votes, cause it's the least gaming and most clear voting system that have already been suggested.

In binary voting the position of grant in awarded list really depends only from total supporters count.
So grants, that supported by most of parties, will be higher than others.

Decide the cut-off level is not a big deal, really.
It should ONLY defence grant pool from bag grants being funded because off lack of good projects in the round.
 
Secured
#63
This solution opens a chance for pre-approved grants (as Guide Pay) not to fit into grant pool size :)
E.g. De Facto has 40 points, how many votes should be placed for Guide Pay? 3? 5? 10?
In strong grant round good projects may receive many more votes, so pre-approved grant like Guide Pay or something else won't get into awarded list.
Grants need 60 points or higher to pass. So you'd want to vote at least 60 for grants like Guide Pay. That would leave you 340 points to vote on other grants so you better think carefully how you want to do it.

Again, the 40 points per grant was arbitrary. It could be 50 or 60. But having a system like this forces more thought which is a good thing AND it fits within the current governance. And yes, before someone says it, an ANO or two may not put any thought into it but that'll be the case for any voting system.

I don't see what binary voting solves. I do see the complications it creates such as having to rewrite governance doc 001 with the new rules you're going to have to come up with.
 
Secured
#64
Grants need 60 points or higher to pass. So you'd want to vote at least 60 for grants like Guide Pay. That would leave you 340 points to vote on other grants so you better think carefully how you want to do it.

Again, the 40 points per grant was arbitrary. It could be 50 or 60. But having a system like this forces more thought which is a good thing AND it fits within the current governance. And yes, before someone says it, an ANO or two may not put any thought into it but that'll be the case for any voting system.

I don't see what binary voting solves. I do see the complications it creates such as having to rewrite governance doc 001 with the new rules you're going to have to come up with.
Even 400 (40 x 10 projects) or more votes keep gaming. Why I should vote 60 to abstain? What about 0-60 — what the sense of this scale if lower than 60 is not approved?

What the difference between 10 and 20 votes? In general and in specific project?

This system is still complicated and unclear. Dofferent scales among different ANOs.

Binary votes fixes it all. We have time to amend Doc 001
 
Secured
#65
Let me rephrase.
@David Chapman in your variant:
0-59 — don’t support
60 — neutral/abstain
61-100 — support

It’s all the same binary, but with 2 disadvantages:

1) Different scale. Look at previous round results. Luciap voted 0 for project he support, but thought it’s not a time fot it. Others voted 0 if didn’t support
1.1) Is 10/20/30 votes all the same? (not support). If not, what’s the difference, if yes, why we need this unclear untransparent scale?
1.2) All the same for 61-100. 65/75/85 what’s the difference?

2) Gaming kept. I can vote 0 for grants I don’t support, 60 for abstain, 61 for grants I support, so there lowering all grants I voted 61 in the grant list.

If both 61 & 100 are shown support, how to determine what grants should get 61, what 80, what 100? How to have ANOs use the same scale? Impossible. Gaming available. Results more unpredictable.

And limiting votes to any amount doesn’t solve this problems at all.

So, I see that your variant will keep both issues we declared few messages above.

So, I advocate to simplify it to -1/0/1.
 
Secured
#67
@David Chapman - Quick clarification question:

Could a given Standing Party vote 70 (say out of 400, from your example) on a given Grant? Or is 40 the theoretical maximum you can award to any one Grant?

If the former, couldn't a Standing Party vote the maximum on 5 Grants they love and the bare minimum on 5 Grants they don't like? Does a system like this inadvertently promote maximized and minimized voting (i.e. strong yes or strong no)? If so, maybe this would not be a problem based on the averaging of all scores from all Standing Parties? Just trying to wrap my head around the implications here.

If the latter, couldn't a Standing Party simply vote the 40 point maximum, per Grant, in an effort to hopefully hedge between all Grants and have as many funded as possible? Does this promote the potential for lack of researched voting?

My apologies if I am totally butchering my analysis of your proposed voting system.

Edit: Please disregard - just saw the answer to this in Post #63.
 
Secured
#68
From what I can tell, if I am a Standing Party and administered (for example) 400 votes on the 10 non-affiliated Grants I can vote on, is there a maximum I can delegate to a particular Grant; is it a score of 100? If I wanted to maximize my voting power over a few specific Grants, could I do so with 100 points on 4 Grants that I love? I am not saying there's necessarily an inherent problem with this, but just clarifying the extreme ends of what is possible within this system.

I think what Anton is getting at is: If 60 is the theoretical level of support needed for approval, and some ANO's are balanced in their voting (something more like voting 61 when in support, and 30-50 when in somewhat support, and 0 in no support ... so that they can distribute votes for many grants), do these "more balanced" votes end up fully washed out or over-powered by maximum-type votes (say, 100 over 4 grants)?

Basically I am trying to figure out how max-based/min-based voting may affect our results when some parties max/min vote and some balance vote.
 
Secured
#69
Anton, when you say -1/0/1 is there a numerical total and if I vote -1 does it take away from it? Or this this just no / abstain / yes ?
Yeap. When someone votes against grant, the total decreased by 1. If someone votes for the grant, the total increased by 1.

It’s the same as current system, but without gaming with large scale.

Additionally, everyone see who votes for and who against, without double senses like “here is my 0 is for support, but it’s not time for this” & “here is my 30 for non-support, but I want to show that this grant may be supported if will be modified”.

As many ANOs, as many scales, so I advocate for avoiding different scales and gaming with binary voting and transperent clear rules.
 
Secured
#72
From what I can tell, if I am a Standing Party and administered (for example) 400 votes on the 10 non-affiliated Grants I can vote on, is there a maximum I can delegate to a particular Grant; is it a score of 100? If I wanted to maximize my voting power over a few specific Grants, could I do so with 100 points on 4 Grants that I love? I am not saying there's necessarily an inherent problem with this, but just clarifying the extreme ends of what is possible within this system.

I think what Anton is getting at is: If 60 is the theoretical level of support needed for approval, and some ANO's are balanced in their voting (something more like voting 61 when in support, and 30-50 when in somewhat support, and 0 in no support ... so that they can distribute votes for many grants), do these "more balanced" votes end up fully washed out or over-powered by maximum-type votes (say, 100 over 4 grants)?

Basically I am trying to figure out how max-based/min-based voting may affect our results when some parties max/min vote and some balance vote.
Yes, you could max 4 grants at 100 each. And if there's four you absolutely love and don't at all like the rest, that's what you should do.

I should state that one of my reasons in devising this system is that I am of the opinion that it is important that a minority be able to push through a grant that the majority does not like. That will never happen in a binary system and sooner or later you ARE going to end up with a majority of like-minded parties dictating the direction of this ecosystem. One thing I've found in life is the majority are often wrong or there's no good answer but multiple avenues should be explored.
 
Last edited:
Secured
#74
@David Chapman
1. Do you have any ideas how to fix different scales (among ANOs) issue in system you suggested?
2. How to prevent gaming? Or it’s a gaming system and gaming should be?
3. I understand your idea of large positive scale (60-100), but I’m totally sure that large negative scale (0-60) with no clear rules how to use it, supported and agreed by all parties, allows the manipulation of results. Just a few 0 votes against good project with a lot of 60-100 votes will lower it significantly.
 
Secured
#76
What if instead of binary voting, you could vote -1, 0, 1, 2, 3. -1 means you don't support the grant, but you can still have different degrees of support for a grant if you like multiple grants but like some more than others.
 
Secured
#77
To get a more accurate view of grant support, maybe you could even add decimals in the positive support up to 3. So a grant could receive a 1.8 and another could receive a 1.9. This would only apply for positive support though and only a -1 can be given when going under 0
 
Secured
#78
@Anton Ilzheev - Maybe I'm not thinking of this the right way, but in my understanding of the binary system, we have possible scores, per Grant, of -1, 0, and 1, and in addition, when scoring these votes, the "total score" is calculated by taking all "FOR" votes (1) minus "AGAINST" votes (-1).

So you could have 31 Standing Parties voting something like 21 FOR and 10 AGAINST = 11 as the vote TOTAL.

What if a variety of Grants have a total score of 11? Is there a threshold here that Grants need to hit in order for them to be funded? And if so, what is that threshold?
 
Secured
#79
@Anton Ilzheev - Maybe I'm not thinking of this the right way, but in my understanding of the binary system, we have possible scores, per Grant, of -1, 0, and 1, and in addition, when scoring these votes, the "total score" is calculated by taking all "FOR" votes (1) minus "AGAINST" votes (-1).

So you could have 31 Standing Parties voting something like 21 FOR and 10 AGAINST = 11 as the vote TOTAL.

What if a variety of Grants have a total score of 11? Is there a threshold here that Grants need to hit in order for them to be funded? And if so, what is that threshold?
We may analyze previous grant round projects scores, especially grants with 50-60 points. And simulate -1/0/1 instead of 0-100, then analyze results and calculate threshold.

I can do it on 1/1/19, I believe, then can post result of this analyze here :)
 
Secured
#81
I want to suggest an alternative voting system: STV (single transferable vote) or ranked voting. I made this suggestion during round one, but there weren't enough time to implement. Ranked voting is designed to maximize participants' happiness with the result, which is exactly what we need.
(When watching the second video, think of the candidates in the video as placeholder for grant proposals during grant rounds, guides during guide elections, and ANOs during ANOs elections. You can skip to 2:25 and 6:00 mark respectively to get the gist of what it tries to achieve. Though I do encourage you to watch the whole thing because I think Ranked Voting is the way to go with our upcoming Guides and ANOs elections.)
 
Secured
#82
I want to suggest an alternative voting system: STV (single transferable vote) or ranked voting. I made this suggestion during round one, but there weren't enough time to implement. Ranked voting is designed to maximize participants' happiness with the result, which is exactly what we need.
(When watching the second video, think of the candidates in the video as placeholder for grant proposals during grant rounds, guides during guide elections, and ANOs during ANOs elections. You can skip to 2:25 and 6:00 mark respectively to get the gist of what it tries to achieve. Though I do encourage you to watch the whole thing because I think Ranked Voting is the way to go with our upcoming Guides and ANOs elections.)
Is ranking voting suitable with automatic-abstaining from voting on own grants?
 
Secured
#84
@Anton Ilzheev, the simplified YouTube video only uses color. Check out this version that shows percentages: SF Ranked Voting Explained

Each time the lowest vote getter gets eliminated, voters’ second choice gets distributed, increasing the total percentages. Based on those percentages, you can figured out who gets the grants similar to our last grant round. (There is a small chance of tie, but it can be resolved with a second vote to break the tie)

1. We don’t know in advance how many grants will be awarded, cause each grant has different budget
Since you have the percentages, you can allocate the grants based on who ranked higher. Same as last round.

2. There is no mechanism cut off unqualified grants proposals, even if they fit into grant pool size
As a community we can’t even determine what it means to be “unqualified”. Case in point, the last grant round’s argument about backpay and Inc’s lack of breakdown. I feel that this problem is a separate issue from how the voting system should be structured. Once the community determines what it means to be “unqualified”, voters can rank accordingly.

Is ranking voting suitable with automatic-abstaining from voting on own grants?
As long as ANO(s) who are abstaining don’t vote, their non-involvement shouldn’t impact the end result (based on my understanding). If out of 10 ANOs, 2 abstained and 8 votes. Then the percentages will be based on the 8 who votes. That’s very definition of abstaining without worrying about using abstention to game the system.
 
Secured
#85
This is a bit off topic from the Ranked Voting proposal, but my gut feeling is telling me we should be careful excluding ANOs from voting just because they are part of a grant. I tend to listen to my intuition but I am also conflicted because I completely agree with the sentiment about full disclosure. In fact, when I made the post about improving Factom Governance, I left out the last proposal. Suggestion #3 is to have all ANOs disclosure their relationship with the community (both financial and non financial). However, I got pushback privately from some ANOs and decided to sit on it a bit before posting publicly.

All of this to say that I agree we need full disclosure for the long term health of the community. However, attaching a vote exclusion clause to that requirement can set very bad precedents that is as dangerous as lack of disclosure. Just off the top of my head, it does two things:

1. Huge barrier of entry for ANOs to collaborate on projects because to do so means you cut yourself off from community votes. Maybe some ANOs don’t care about grant pools and don’t mind losing their voting power there. But the rationale behind this proposal can easily be carried over to non-grant related voting situations. (e.g. guides election, ANO election, governance document votes, etc) That is a dangerous path to head down because we are asking Factom stakeholders (beyond ANOs) to choose between collaboration and voting rights.

2. The vote exclusion implies voting rights of sponsoring stakeholders do not matter compare to the rest of the community. It also implies the rest of the voting parties will act in community’s best interest. I disagree with both assumptions. Automatically exclude sponsoring ANOs can be used as an attack vector to shoot down proposals for purely political reasons. It might not be apparent right now, but I think people voting for political reasons will become common place in the future. (All the more reason to get away from binary voting system and adopt Ranked Voting)

My reading of the situation is that people are getting behind vote exclusion in order to prevent multiple ANOs from sponsoring grants together to guarantee voting bloc. I see that as very valid concern. However, can’t this problem be resolve by limiting sponsoring ANOs to 3 or 4? (In the context of 70+ voting stakeholders, that’s not unreasonable). If grant proposal has more than 4 sponsors, it probably begs the question of whether the grant should be split up. Using the existing solution of splitting up grant is a lot safer path than the bad precedents I outlined above.

I still support full disclosure by ANOs regardless how the voting-exclusion clause gets decided. Last couple months I have already seen signs of ANOs non-disclosured collaboration affecting how they act publicly. I think this problem will get worse. However, that is a much larger discussion because we need to deal with enforceability, timing, business strategies, etc.
 
Secured
#86
Yeah, I think the better long term solution would be to let ANOs who sponsor a grant, vote on that grant the same as the other non-sponsoring ANOs. Hopefully in the future, the addition of other standing party groups (stakers, ec users, etc) will make it so ANOs don't have so much power in deciding which grants gets funded.
 
Secured
#88
I don't have an issue with sponsors voting on a grant as theoretically, they should be representing the community.

My interest in exclusions is having ANOs with financial connections not vote on each others grants and those who are part of the grant not vote on it. However, they can all vote on other grants, so their votes aren't be taken away completely. They just can't vote on their own grant.
 
Secured
#89
@David Chapman Don't sponsors also have "financial connections" if they will be paid by the grant recipient?

Just to be clear on my stance, the problem we are trying to avoid is a potential grant recipient stacking their grant with 3-4 other ANOs doing DEV work, 2-3 sponsors who will be paid by the grant recipient, and potentially other standing parties (equity holder, partnership, etc.) that would benefit from the grant recipient receiving the grant.

I'm totally fine with disallowing financially connected standing parties/sponsors from voting entirely, my suggestion of only allowing 2-3 parties to vote was simply an attempt at finding a happy medium, as I think @Xavier Chen brought up some good points.
 
Secured
#90
I share the same concerns with you Matt but sponsors are at the bottom of that list of concern as theoretically:

1. They will be the lowest paid people on the grant.
2. They are supposed to represent the community.
3. They aren't paid by the grant recipient but by the grant pool and thus the community.

I'm also not so naive as to think no potential sponsor out there will ever vote a little higher because there's some FCT in it for them if the grant passes. IF the consensus is that's it's a problem we should deal with at this juncture, then I'd suggest that the grant recipient not be able to choose the sponsor, they simply budget for it and the community chooses the sponsor after the fact.

From a development and administrative points of view, allowing 2 or 3 parties to vote but excluding the others would be a nightmare.