BitShares Forum
Main => General Discussion => Topic started by: bytemaster on July 11, 2014, 02:42:10 am
-
After doing much thinking and learning from experience with recent dry runs I have concluded that we need to have negative votes.
A negative vote is morally equivalent to a vote of approval for "everyone but X" like a positive vote is the moral equivalent as a vote of disapproval for everyone but Y.
In a world where the squeaky wheal gets the grease, it is far more likely that the average user will know more about who they are against than who they are for. When there is a bad actor it is far more difficult to get people to up-vote 101 other candidates than down vote the guy causing the problem. For automated voting it is also far easier to down vote delegates that are not producing blocks than to decide who to up vote.
The average user is lazy and unlikely to evaluate 101 different individuals. We must factor in this laziness in our designs.
Lastly in a world with 1000 candidates, it takes a lot of work to concentrate the approval consensus. Suppose Bad Actor has 25% approval, everyone else has to find someone they can agree to give 26% approval to bump the Bad Actor.... there are 2 solid candidates and they each split the vote 13% / 13% and thus neither is able to bump the Bad Actor. Everyone can agree the Bad Actor should go, so they simply vote against him.
Negative voting is effectively an easy way to compress a "vote for all, but 1" which in a way helps simulate the ability to approve more than 101 delegates.
-
:-\
-
It may be too late for this. But I was thinking that you could have this as a perpetual auction. And have it so becoming a delegate is only for a limited term.
That way it will be easy for positions to be filled at any point based on demand. And it takes care of the constant inflation when new delegates want to come in.
Let's say it's a quarterly thing (people in finance would appreciate that).
So every 3 months positions are auctioned. And people have the ability to look at stats from prior 3 months to decide who to vote for.
Also wanted to ask - if I wanted to be a super reliable delegate - is it possible to run the same client on multiple nodes with same user ID? that way if one of my machines goes down, I am still up and running and providing service. Or would that create unnecessary forks?
Sorry for the run on sentences and going all over the place, about to leave, and wanted to get all those in before I go.
-
Down votes should only be used in the short term because of the "wack-a-mole" problem.... but with approval voting it becomes very difficult to get the required votes to be in the top 101 in the first place and the long term solution is for people to find 101 delegates they can actively support.
-
It may be too late for this. But I was thinking that you could have this as a perpetual auction. And have it so becoming a delegate is only for a limited term.
That way it will be easy for positions to be filled at any point based on demand. And it takes care of the constant inflation when new delegates want to come in.
Let's say it's a quarterly thing (people in finance would appreciate that).
So every 3 months positions are auctioned. And people have the ability to look at stats from prior 3 months to decide who to vote for.
Also wanted to ask - if I wanted to be a super reliable delegate - is it possible to run the same client on multiple nodes with same user ID? that way if one of my machines goes down, I am still up and running and providing service. Or would that create unnecessary forks?
Sorry for the run on sentences and going all over the place, about to leave, and wanted to get all those in before I go.
You can run on multiple nodes, so long as you don't produce 2 blocks at once with the same key.
-
super frustrated... Down votes are not the answer. There should be no down voting. There is also no need for "random" voting. All people need are network statistics and no random voting and people will happily remove support for under-performing delegates. Down voting destroys the integrity of the election process. You are thinking about "users" all wrong. These are shares with shareholders. Shareholders are not lazy. Shareholders care. Look how active this community is and try to help. A new user with a tiny stake won't have an impact anyway. Now if I want my votes to have a bigger impact I have to waste time voting down everyone else. There also is no need for automated voting, it is not a good thing.
-
super frustrated... Down votes are not the answer. There should be no down voting. There is also no need for "random" voting. All people need are network statistics and no random voting and people will happily remove support for under-performing delegates. Down voting destroys the integrity of the election process. You are thinking about "users" all wrong. These are shares with shareholders. Shareholders are not lazy. Shareholders care. Look how active this community is and try to help. A new user with a tiny stake won't have an impact anyway. Now if I want my votes to have a bigger impact I have to waste time voting down everyone else. There also is no need for automated voting, it is not a good thing.
+5%
-
Approval voting made a lot more sense when there was talk of delegates being able to give back to those who vote for them. If delegates can see who voted for them and can utilize kickbacks then that is when approval voting would become useful.
-
Approval voting made a lot more sense when there was talk of delegates being able to give back to those who vote for them. If delegates can see who voted for them and can utilize kickbacks then that is when approval voting would become useful.
Delegates can still see what addresses are voting for them. Do you think something has changed?
-
Now if I want my votes to have a bigger impact I have to waste time voting down everyone else. There also is no need for automated voting, it is not a good thing.
My understanding was previously you had 1 vote. The above implies something like approval/disapproval voting. Where every candidate gets a yay/nay.
?
Delegates can still see what addresses are voting for them. Do you think something has changed?
No, I just never knew completely how it works. I tried participating last dry run but after several hours of not hearing back on my request for XTS I moved on to my other interests in Bitshares.
Honestly, they need to create a real simple faucet for these dry runs so people can just get involved without having to bother devs etc.
-
Now if I want my votes to have a bigger impact I have to waste time voting down everyone else. There also is no need for automated voting, it is not a good thing.
My understanding was previously you had 1 vote. The above implies something like approval/disapproval voting. Where every candidate gets a yay/nay.
I'm not sure I understand your position or what you are advocating for. Originally a share could only vote for 1 delegate at a time up/down. Now it seems he is keeping ability to vote for multiple delegates but can vote for or against. Either way is wrong. Pure approval voting with only up votes and no random votes or automated voting is perfect (just provide network statistics for warnings/recommendations).
-
Now if I want my votes to have a bigger impact I have to waste time voting down everyone else. There also is no need for automated voting, it is not a good thing.
My understanding was previously you had 1 vote. The above implies something like approval/disapproval voting. Where every candidate gets a yay/nay.
I'm not sure I understand your position or what you are advocating for. Originally a share could only vote for 1 delegate at a time up/down. Now it seems he is keeping ability to vote for multiple delegates but can vote for or against. Either way is wrong. Pure approval voting with only up votes and no random votes or automated voting is perfect (just provide network statistics for warnings/recommendations).
I was neutral about it until I heard your explanation about the problems with the original system. That being an opportunity cost to downvote when delegates give kickbacks. I'm back to being neutral until I understand your issues with the new voting system.
Without automated voting, how will a bad delegate be removed quick enough from the system ? You're also thinking of this in terms of just one chain. There will be multiple chains, all which require people to be involved. Automated voting needs to be implemented, but should always be optional.
Automated voting is also something people can implement in new wallets , regardless of what is decided on the official release.
-
super frustrated... Down votes are not the answer. There should be no down voting. There is also no need for "random" voting. All people need are network statistics and no random voting and people will happily remove support for under-performing delegates. Down voting destroys the integrity of the election process. You are thinking about "users" all wrong. These are shares with shareholders. Shareholders are not lazy. Shareholders care. Look how active this community is and try to help. A new user with a tiny stake won't have an impact anyway. Now if I want my votes to have a bigger impact I have to waste time voting down everyone else. There also is no need for automated voting, it is not a good thing.
How about that?
you can't vote both "Up" and "Down" with the same shares.
When block chain security state is green, the "Up" vote is active
otherwise, the "Down" vote is active.
If you are a delegate, you only need to vote yourself,if you vote against someone, you will lose your own vote.
for others not as delegate, normally they choose delegate they believe,
but when things go to bad, they can get rid of the bad delegate as soon as possible.
-
What about to let the aproval method as it is right now in general...
BUT have the oportunity to downvote ONLY active delegates (first 101)...
As soon they are not active (stand-by) you can not downvote them.
OR better make a hybrid aproval system !!!
Let the downvotes matures in time...
For example when someone get a downvote he should "keep" the downvote for x time (1 month?) (but recorded in his history for ever....)
Something like penalty. Only upvotes and penaltys. When you see after 3 years someone has 6 penaltys you will be more carefull with upvoting him for example
-
I will suggest something a little different:
1 Keep approval voting with no downvotes
2 Apply X% vote weight decrease after Y blocks. ( I suggest X to be approx 1%, Y ~101)
Using this a delegate need to have continuous support in order to stay in top 101.
Any vote will not matter once enough rotations have passed.
Those who use the system (making transaction) will dictate who will be delegate IN SHORT TERM.
Ofcourse one could just transfer stake between own wallets and routing votes through his own delegates so he is not paying taxes. (and this is one more reason for restrictions for delegates to burn at least some % of the transactions)
-
It may be too late for this. But I was thinking that you could have this as a perpetual auction. And have it so becoming a delegate is only for a limited term.
That way it will be easy for positions to be filled at any point based on demand. And it takes care of the constant inflation when new delegates want to come in.
Let's say it's a quarterly thing (people in finance would appreciate that).
So every 3 months positions are auctioned. And people have the ability to look at stats from prior 3 months to decide who to vote for.
Also wanted to ask - if I wanted to be a super reliable delegate - is it possible to run the same client on multiple nodes with same user ID? that way if one of my machines goes down, I am still up and running and providing service. Or would that create unnecessary forks?
Sorry for the run on sentences and going all over the place, about to leave, and wanted to get all those in before I go.
You can run on multiple nodes, so long as you don't produce 2 blocks at once with the same key.
Any easy way to do that on Windows with the GUI? I have two machines (its crashing on one, btw), and would like to use one of them as backup.
-
I will suggest something a little different:
1 Keep approval voting with no downvotes
2 Apply X% vote weight decrease after Y blocks. ( I suggest X to be approx 1%, Y ~101)
Using this a delegate need to have continuous support in order to stay in top 101.
Any vote will not matter once enough rotations have passed.
Those who use the system (making transaction) will dictate who will be delegate IN SHORT TERM.
Of course one could just transfer stake between own wallets and routing votes through his own delegates so he is not paying taxes. (and this is one more reason for restrictions for delegates to burn at least some % of the transactions)
interesting...
-
I will suggest something a little different:
1 Keep approval voting with no downvotes
2 Apply X% vote weight decrease after Y blocks. ( I suggest X to be approx 1%, Y ~101)
Using this a delegate need to have continuous support in order to stay in top 101.
Any vote will not matter once enough rotations have passed.
Those who use the system (making transaction) will dictate who will be delegate IN SHORT TERM.
Ofcourse one could just transfer stake between own wallets and routing votes through his own delegates so he is not paying taxes. (and this is one more reason for restrictions for delegates to burn at least some % of the transactions)
Wouldn't this just destroy security as the people who don't use their shares will have their vote diluted then a party that is a large share holder could grossly inflate their influence by making a transaction? They could also automate the process.
I know NEM has something like Proof of Interest. I should read more about it.
-
Considere something else:
Maybe deppended on % delegate participation to remove the last active delegates on every round... (NOT RANDOM)
if 80% participation, remove the last 20% of active delegates and replace them with the first delegates in the standby list...
if 95% participation, remove the last 5% of active delegates and replace them with the first delegates in the standby list...
if 60% participation, remove the last 40% of active delegates and replace them with the first delegates in the standby list...
etc. etc.
normaly the participation % will go up until it reach a equillibrium point that is the highest possible depended on delegates average accuracy...
and you give oportunitys for potential better delegates to get active !!!
-
Why not just have review sites outsource the process of rating delegates so users can remain lazy?
I don't see how this is a big deal.
-
Isn't the risk of downvotes that a powerful group of shareholders, either large/active, will have both the power to vote in a large portion of policies (via delegates) they want and effectively vote out the type of policies (via delegates) that they don't. - Like the Democrats would not only be able to vote themselves into a majority position but also vote out all of the Republicans from the remaining positions too?
I can see the very important need for downvotes to rapidly remove bad actors though.
-
Isn't the risk of downvotes that a powerful group of shareholders, either large/active, will have both the power to vote in a large portion of policies (via delegates) they want and effectively vote out the type of policies (via delegates) that they don't. - Like the Democrats would not only be able to vote themselves into a majority position but also vote out all of the Republicans from the remaining positions too?
I can see the very important need for downvotes to rapidly remove bad actors though.
If only the system can downvote delegates that not produce blocks (or acting bad) and replace them automatically with stand-by delegates?
-
Disclaimer: I just read the OP...
Maybe something like this would help reach the goal of effectively voting against bad actors: Down votes have twice the effect as positive votes + there are only 3 (or something similar) down votes you can make. The latter ensures that the bad actors doesn't use the "double down votes" to vote everyone down except him which others would be to lazy to do...
This The average user is lazy and unlikely to evaluate 101 different individuals. We must factor in this laziness in our designs.
I think is very important and probably in the end only those with huge stakes and the fanboys will vote at all....
-
Isn't the risk of downvotes that a powerful group of shareholders, either large/active, will have both the power to vote in a large portion of policies (via delegates) they want and effectively vote out the type of policies (via delegates) that they don't. - Like the Democrats would not only be able to vote themselves into a majority position but also vote out all of the Republicans from the remaining positions too?
I can see the very important need for downvotes to rapidly remove bad actors though.
If only the system can downvote delegates that not produce blocks (or acting bad) and replace them automatically with stand-by delegates?
But if we had downvotes we could pre-empt that bad behaviour. A group of delegates might be behaving well but observant people might be able to see that are setting up to collude and disrupt the system. For example if you could see these guys were among the 101 delegates, wouldn't you want the power to downvote?
(http://wac.450f.edgecastcdn.net/80450F/screencrush.com/files/2012/10/warped-movie-villains.jpg)
-
Isn't the risk of downvotes that a powerful group of shareholders, either large/active, will have both the power to vote in a large portion of policies (via delegates) they want and effectively vote out the type of policies (via delegates) that they don't. - Like the Democrats would not only be able to vote themselves into a majority position but also vote out all of the Republicans from the remaining positions too?
I can see the very important need for downvotes to rapidly remove bad actors though.
If only the system can downvote delegates that not produce blocks (or acting bad) and replace them automatically with stand-by delegates?
No such thing as something "only the system" can do.
-
I will suggest something a little different:
1 Keep approval voting with no downvotes
2 Apply X% vote weight decrease after Y blocks. ( I suggest X to be approx 1%, Y ~101)
Using this a delegate need to have continuous support in order to stay in top 101.
Any vote will not matter once enough rotations have passed.
Those who use the system (making transaction) will dictate who will be delegate IN SHORT TERM.
Of course one could just transfer stake between own wallets and routing votes through his own delegates so he is not paying taxes. (and this is one more reason for restrictions for delegates to burn at least some % of the transactions)
interesting...
Delegates share the fees in their block with all other delegates. You cannot "route transactions through your own delegate" to get free transactions.
-
Isn't the risk of downvotes that a powerful group of shareholders, either large/active, will have both the power to vote in a large portion of policies (via delegates) they want and effectively vote out the type of policies (via delegates) that they don't. - Like the Democrats would not only be able to vote themselves into a majority position but also vote out all of the Republicans from the remaining positions too?
I can see the very important need for downvotes to rapidly remove bad actors though.
If only the system can downvote delegates that not produce blocks (or acting bad) and replace them automatically with stand-by delegates?
No such thing as something "only the system" can do.
What about that?
Considere something else:
Maybe deppended on % delegate participation to remove the last active delegates on every round... (NOT RANDOM)
if 80% participation, remove the last 20% of active delegates and replace them with the first delegates in the standby list...
if 95% participation, remove the last 5% of active delegates and replace them with the first delegates in the standby list...
if 60% participation, remove the last 40% of active delegates and replace them with the first delegates in the standby list...
etc. etc.
normaly the participation % will go up until it reach a equillibrium point that is the highest possible depended on delegates average accuracy...
and you give oportunitys for potential better delegates to get active !!!
-
Agent86 is very persistent and talked with me on Skype last night to argue against negative votes.
The primary argument he made was that abstaining is the responsible thing to do for uninterested users and that large stake holders will be very pro-active.
Small stake holders collectively do not have much influence.
I have decided to make it an option that is easy to turn on in the future with a hard-fork, but off by default. If it becomes a problem then the delegates can vote to support it.
-
Agent86 is very persistent and talked with me on Skype last night to argue against negative votes.
The primary argument he made was that abstaining is the responsible thing to do for uninterested users and that large stake holders will be very pro-active.
Small stake holders collectively do not have much influence.
I have decided to make it an option that is easy to turn on in the future with a hard-fork, but off by default. If it becomes a problem then the delegates can vote to support it.
Agent86 is bytemaster's alter ego
-
Agent86 is very persistent and talked with me on Skype last night to argue against negative votes.
The primary argument he made was that abstaining is the responsible thing to do for uninterested users and that large stake holders will be very pro-active.
Small stake holders collectively do not have much influence.
I have decided to make it an option that is easy to turn on in the future with a hard-fork, but off by default. If it becomes a problem then the delegates can vote to support it.
Agent86 is bytemaster's alter ego
C'mon, don't play games toast
Sent from my iPhone using Tapatalk
-
So is there 3 voting states for each candidate delegate ? If there are 2 states it seems functionally equivalent to approval voting, so I assume there are 3 states.
I still haven't seen a good reason why this is better or worse. I suspect a rational person would negative vote everywhere they do not positive vote. This would be done for no other reason than to make the their positive votes more effective. So therefore a rational voter will only use positive/negative and it becomes the same as approval voting. ie yes/no.
Does anyone see a flaw in this line of thought? I haven't really thought about it much ... was hoping for Agent86 to state his case clearly.
I basically am still neutral. I don't see it as worse than approval voting. Even though it may be rational in a game theory sense to solely negative/positive vote, people will not likely do this will and save their negative votes for the bad actors because it seems more intuitive. So in that way it might be an improvement.
-
Suppose Bad Actor has 25% approval, everyone else has to find someone they can agree to give 26% approval to bump the Bad Actor.... there are 2 solid candidates and they each split the vote 13% / 13% and thus neither is able to bump the Bad Actor. Everyone can agree the Bad Actor should go, so they simply vote against him.
Should this point be overlooked? How important is it in real world terms?
It is possible for there to be a candidate who would receive 40% support for a role as delegate, but 60% strong opposition. Should it be possible for this candidate to become a delegate?
If people have the option to vote against, the candidate would be unsuccessful. Without "vote against" the candidate would probably become a delegate, and I propose this to be undesirable. In order for the candidate to be partially unsuccessful, the opposing 60% would likely have to agree to collaborate. They would have organize their "votes for" in a way which they may find unsatisfactory, thus compromising their REAL preferences.
Importantly, even if the 60% managed to use their votes to get another candidate in front of the one they oppose, BOTH candidates would likely still end up being delegates. Now we would still have a delegate who 60% STRONGLY OPPOSE!
Bear in mind that this needn't be a polarized case of a candidate being either a good or bad actor, or processing transactions well or poorly. What if the candidate is proposing to use transaction fees for a certain controversial purpose? It is easy to imagine a scenario whereby the candidate incentivizes 40% to vote for him/her by some means, and this 40% are not opposed to the way the fees will be used. But what if the other 60% are fundamentally against the proposed use of the fees? Should they not be given the chance to voice their concerns?
Could not "vote against" potentially be a useful, if not necessary, tool? If the majority opposes something in principle, shouldn't there be a proper way for them to oppose it in practice?
I admit I haven't thought much about this, and also haven't followed closely enough the way voting works with DPOS, so I may be mistaken.
However, I thought I should post this in case it is useful, or in case it may act as a catalyst in some way.
-
Suppose Bad Actor has 25% approval, everyone else has to find someone they can agree to give 26% approval to bump the Bad Actor.... there are 2 solid candidates and they each split the vote 13% / 13% and thus neither is able to bump the Bad Actor. Everyone can agree the Bad Actor should go, so they simply vote against him.
Importantly, even if the 60% managed to use their votes to get another candidate in front of the one they oppose, BOTH candidates would likely still end up being delegates. Now we would still have a delegate who 60% STRONGLY OPPOSE!
I have some bad news for you:
http://en.wikipedia.org/wiki/Voting_paradox
http://en.wikipedia.org/wiki/Arrow%27s_impossibility_theorem
-
I have some bad news for you:
http://en.wikipedia.org/wiki/Voting_paradox
http://en.wikipedia.org/wiki/Arrow%27s_impossibility_theorem
Neither of these limitations relate to our application. These describe single time point elections; we vote in real time. Approval voting works great for our application.
-
What's true at each single point in time must also be true at all times, literally.
-
What if we build into the software automated voting based on rules. That way majority will just auto vote and do the right thing even if someone is trying to abuse the system with custom made software.
-
What's true at each single point in time must also be true at all times, literally.
::) There are things that can be accomplished over time that can't be accomplished at a single time point. The links are essentially mathematical proofs that describe a very specific situation of "single winner" "single time-point" "instant-runoff (ranked)" style of election.
This is not our system. Our system doesn't use and doesn't require an "instant runoff" style ballot because there is no need for it to be "instant." In our system everyone can clearly see who everyone else is voting for and vote accordingly, allowing equilibria to be reached. Another way to say it is having 1000s of elections that compare preferences in various states provides more information about voter preferences than one election.
Also, neither "voting paradox" nor "Arrow's impossibility theorem" is in any way a proof that voting or elections have no utility or value. It is simply saying that in a very specific situation there are specific provable limitations that might seem unexpected to someone expecting some kind of perfect integration of voter preferences.
-
Maybe time should be involved. Kind of like in Reddit's up voting/down voting algorithm. I believe it works something like this: if a post starts to get many up votes within a few seconds or minutes it becomes weighted and starts rising to the front faster. So a post that gets 100 up votes in 2 minutes will be rising faster and surpass a post that has had 2000 up votes over the last 10 hours.
You just do this in reverse. That way if a delegate is truly a good delegate but gets down votes overtime just because someone wants their position, it doesn't boot them out. Only the truly bad delegates will get booted when something goes wrong and there is an influx of some amount of down votes really close in succession.
-
Maybe time should be involved. Kind of like in Reddit's up voting/down voting algorithm. I believe it works something like this: if a post starts to get many up votes within a few seconds or minutes it becomes weighted and starts rising to the front faster. So a post that gets 100 up votes in 2 minutes will be rising faster and surpass a post that has had 2000 up votes over the last 10 hours.
You just do this in reverse. That way if a delegate is truly a good delegate but gets down votes overtime just because someone wants their position, it doesn't boot them out. Only the truly bad delegates will get booted when something goes wrong and there is an influx of some amount of down votes really close in succession.
+1 +1 +1
This is the best idea I've heard yet.
Sent from my iPhone using Tapatalk
-
Maybe time should be involved. Kind of like in Reddit's up voting/down voting algorithm. I believe it works something like this: if a post starts to get many up votes within a few seconds or minutes it becomes weighted and starts rising to the front faster. So a post that gets 100 up votes in 2 minutes will be rising faster and surpass a post that has had 2000 up votes over the last 10 hours.
You just do this in reverse. That way if a delegate is truly a good delegate but gets down votes overtime just because someone wants their position, it doesn't boot them out. Only the truly bad delegates will get booted when something goes wrong and there is an influx of some amount of down votes really close in succession.
+1 +1 +1
This is the best idea I've heard yet.
Sent from my iPhone using Tapatalk
Incredibly difficult to implement idea.
-
What about limiting the total amount of votes to max 101 different delegates?
Both positive and negative votes limited like this will prevent 51% monopolising all the delegates.
I think negative votes are good addition to approval voting as it gives more flexibility.
I still believe that owners of 51% stake should not get all the delegates!
-
What about limiting the total amount of votes to max 101 different delegates?
Both positive and negative votes limited like this will prevent 51% monopolising all the delegates.
I think negative votes are good addition to approval voting as it gives more flexibility.
I still believe that owners of 51% stake should not get all the delegates!
limiting to 101 ok but really different is practical not possible...
-
Maybe time should be involved. Kind of like in Reddit's up voting/down voting algorithm. I believe it works something like this: if a post starts to get many up votes within a few seconds or minutes it becomes weighted and starts rising to the front faster. So a post that gets 100 up votes in 2 minutes will be rising faster and surpass a post that has had 2000 up votes over the last 10 hours.
You just do this in reverse. That way if a delegate is truly a good delegate but gets down votes overtime just because someone wants their position, it doesn't boot them out. Only the truly bad delegates will get booted when something goes wrong and there is an influx of some amount of down votes really close in succession.
+1 +1 +1
This is the best idea I've heard yet.
Sent from my iPhone using Tapatalk
Incredibly difficult to implement idea.
But it will stay for ever and it possibly makes a huge positive difference...
Can you implement this idea in the near future so you "waste" not time right now?
It's seems worth to think about it seriously...
-
Maybe time should be involved. Kind of like in Reddit's up voting/down voting algorithm. I believe it works something like this: if a post starts to get many up votes within a few seconds or minutes it becomes weighted and starts rising to the front faster. So a post that gets 100 up votes in 2 minutes will be rising faster and surpass a post that has had 2000 up votes over the last 10 hours.
You just do this in reverse. That way if a delegate is truly a good delegate but gets down votes overtime just because someone wants their position, it doesn't boot them out. Only the truly bad delegates will get booted when something goes wrong and there is an influx of some amount of down votes really close in succession.
+1 +1 +1
This is the best idea I've heard yet.
Sent from my iPhone using Tapatalk
Incredibly difficult to implement idea.
But it will stay for ever and it possibly makes a huge positive difference...
Can you implement this idea in the near future so you "waste" not time right now?
It's seems worth to think about it seriously...
and using the same system as reddit is a nice marketing argument
-
What about limiting the total amount of votes to max 101 different delegates?
Both positive and negative votes limited like this will prevent 51% monopolising all the delegates.
I think negative votes are good addition to approval voting as it gives more flexibility.
I still believe that owners of 51% stake should not get all the delegates!
limiting to 101 ok but really different is practical not possible...
We already limit it to 101 votes.
Sent from my iPhone using Tapatalk (http://tapatalk.com/m?id=1)
-
Maybe time should be involved. Kind of like in Reddit's up voting/down voting algorithm. I believe it works something like this: if a post starts to get many up votes within a few seconds or minutes it becomes weighted and starts rising to the front faster. So a post that gets 100 up votes in 2 minutes will be rising faster and surpass a post that has had 2000 up votes over the last 10 hours.
You just do this in reverse. That way if a delegate is truly a good delegate but gets down votes overtime just because someone wants their position, it doesn't boot them out. Only the truly bad delegates will get booted when something goes wrong and there is an influx of some amount of down votes really close in succession.
+1 +1 +1
This is the best idea I've heard yet.
Sent from my iPhone using Tapatalk
Incredibly difficult to implement idea.
I am sure I was over simplifying.
Here is a post about how reddits algo works: http://amix.dk/blog/post/19588
It probably doesn't have to be as elaborate as that.
Maybe just implement a "stale factor" where down votes become meaningless after a certain amount of time. That way the downvote is only relevant when larger community actions are taking place.
-
What about limiting the total amount of votes to max 101 different delegates?
Both positive and negative votes limited like this will prevent 51% monopolising all the delegates.
I think negative votes are good addition to approval voting as it gives more flexibility.
I still believe that owners of 51% stake should not get all the delegates!
I was thinking this too. But I think most companies with shareholders work on 'majority rule' & considering we especially need to keep out minorities that would try to bring 'bad actors' into the system. My thinking now is that majority rule (The majority dictating the majority of the delegates) is probably best.
-
What about limiting the total amount of votes to max 101 different delegates?
Both positive and negative votes limited like this will prevent 51% monopolising all the delegates.
I think negative votes are good addition to approval voting as it gives more flexibility.
I still believe that owners of 51% stake should not get all the delegates!
I was thinking this too. But I think most companies with shareholders work on 'majority rule' & considering we especially need to keep out minorities that would try to bring 'bad actors' into the system. My thinking now is that majority rule (The majority dictating the majority of the delegates) is probably best.
1 delegate cant do much harm
101 could
giving single group (even if that group owns 51% of the stake) is against "decentralisation".
In the current implementation (at least the way I understand it) the majority will dictate the majority of the delegates.
-
What about limiting the total amount of votes to max 101 different delegates?
Both positive and negative votes limited like this will prevent 51% monopolising all the delegates.
I think negative votes are good addition to approval voting as it gives more flexibility.
I still believe that owners of 51% stake should not get all the delegates!
I was thinking this too. But I think most companies with shareholders work on 'majority rule' & considering we especially need to keep out minorities that would try to bring 'bad actors' into the system. My thinking now is that majority rule (The majority dictating the majority of the delegates) is probably best.
1 delegate cant do much harm
101 could
giving single group (even if that group owns 51% of the stake) is against "decentralisation".
In the current implementation (at least the way I understand it) the majority will dictate the majority of the delegates.
I don't think so, the 51% stake could be owned by a mixture of 5 Million people, while the remaining 49% stake could be owned by 1 person, so I don't think majority rule affects the decentralisation argument too much. I could be wrong, it's just what popped into my head now.
-
What about limiting the total amount of votes to max 101 different delegates?
Both positive and negative votes limited like this will prevent 51% monopolising all the delegates.
I think negative votes are good addition to approval voting as it gives more flexibility.
I still believe that owners of 51% stake should not get all the delegates!
limiting to 101 ok but really different is practical not possible...
We already limit it to 101 votes.
Sent from my iPhone using Tapatalk (http://tapatalk.com/m?id=1)
Great!
So I'll recap once again to make sure I understand:
If I make a transaction I'll be able to do ONE of the following:
1 vote positive for up to 101 different delegates
2 vote negative for up to 101 different delegates
3 vote positive for x delegates and vote negative for up to 101 - x delegates (x<101 obviously)
Each 101 blocks votes will be counted like this:
For each delegate candidate the result "vote" will be positive - negative votes for that delegate.
Order candidates by vote and appoint top 101 as next round delegates.
Is this correct?
-
What about negative votes but that are only equal to 0.4 of a positive vote?
While I have no idea how this plays out with the ability to vote for multiple candidates with multiple votes ??? but...
If you are given only 3 candidates & one positive vote. With one candidate clearly being a bad actor - Darth Vader
Yes Vote
20% Darth Vader
20% Darth Vader
20% Mila Kunis
20% Eva Mendez
20% Mila Kunis
Then a minority who controlled 41% of the shares could win an election against an undecided majority.
(This could also be a minority of 20.5% in a company where only 50% were actively voting etc.)
However the majority can agree that Darth Vader should not be a delegate. So if you gave everybody one negative vote too...
Yes Vote No Vote
20% Darth Vader Mila Kunis
20% Darth Vader Mila Kunis
20% Mila Kunis Darth Vader
20% Eva Mendez Darth Vader
20% Mila Kunis Darth Vader
With negative votes that are equally weighted, the bad actor, Darth Vader doesn't win the election :)
The new problem it creates is that Eva Mendez wins the election (Eva=1, Mila=0, Vader=-1) :( (When really Mila Kunis is the favourite of the good majority)
But if you made Negative votes worth only 0.4 of a Yes vote then the results above would now be Mila=1.2; Eva=1; Vader=0.8
So Mila Kunis wins the election. (This is fair she is the most popular choice of the good majority) Eva Mendez would come second and the Bad Actor would come last. The best result.
However the Bad Actor group would then change their No vote to Eva Mendez, so that their Bad Actor, Darth Vader at least comes second Mila=2; Vader=0.8, Eva=0.2 :(
Still including negative votes that are lower weighted than positive votes
A) Makes it very hard for bad actors supported by a minority to win delegate positions if the majority are strongly opposed.
B) While at the same time also make it very hard for a minority to stop the most popular choice of the majority from winning the election.
-
What about limiting the total amount of votes to max 101 different delegates?
Both positive and negative votes limited like this will prevent 51% monopolising all the delegates.
I think negative votes are good addition to approval voting as it gives more flexibility.
I still believe that owners of 51% stake should not get all the delegates!
limiting to 101 ok but really different is practical not possible...
We already limit it to 101 votes.
Sent from my iPhone using Tapatalk (http://tapatalk.com/m?id=1)
Great!
So I'll recap once again to make sure I understand:
If I make a transaction I'll be able to do ONE of the following:
1 vote positive for up to 101 different delegates
2 vote negative for up to 101 different delegates
3 vote positive for x delegates and vote negative for up to 101 - x delegates (x<101 obviously)
Each 101 blocks votes will be counted like this:
For each delegate candidate the result "vote" will be positive - negative votes for that delegate.
Order candidates by vote and appoint top 101 as next round delegates.
Is this correct?
That was the proposal.
-
Ok, so we are limited to 101 votes either positive or negative. So I retract my previous post about this regressing into the same thing as approval voting by rational voters because blanket negative voting is not allowed.
I still think the rational move is to initially utilize all your negative votes to reinforce your positive votes.
I suppose the expectation here is that the negative votes will be shifted around to actual poor performing delegates at some point, and not negative votes used to reinforce your positive votes. SO in that way it seems like an improvement, but the additional options allow more gaming of the system which will introduce things I can't immediately comprehend.
It would be interesting to write little voting sim scripts and try to figure out how this actually changes the balance of power, because I think it is beyond my immediate intellectual capability. :(
-
Great!
So I'll recap once again to make sure I understand:
If I make a transaction I'll be able to do ONE of the following:
1 vote positive for up to 101 different delegates
2 vote negative for up to 101 different delegates
3 vote positive for x delegates and vote negative for up to 101 - x delegates (x<101 obviously)
Each 101 blocks votes will be counted like this:
For each delegate candidate the result "vote" will be positive - negative votes for that delegate.
Order candidates by vote and appoint top 101 as next round delegates.
Is this correct?
That was the proposal.
Seems like this is solving most of the issues I see.
I think this is the best system so far for this case.
-
Suppose Bad Actor has 25% approval, everyone else has to find someone they can agree to give 26% approval to bump the Bad Actor.... there are 2 solid candidates and they each split the vote 13% / 13% and thus neither is able to bump the Bad Actor. Everyone can agree the Bad Actor should go, so they simply vote against him.
Should this point be overlooked? How important is it in real world terms?
It is possible for there to be a candidate who would receive 40% support for a role as delegate, but 60% strong opposition. Should it be possible for this candidate to become a delegate?
If people have the option to vote against, the candidate would be unsuccessful. Without "vote against" the candidate would probably become a delegate, and I propose this to be undesirable. In order for the candidate to be partially unsuccessful, the opposing 60% would likely have to agree to collaborate. They would have organize their "votes for" in a way which they may find unsatisfactory, thus compromising their REAL preferences.
Importantly, even if the 60% managed to use their votes to get another candidate in front of the one they oppose, BOTH candidates would likely still end up being delegates. Now we would still have a delegate who 60% STRONGLY OPPOSE!
Bear in mind that this needn't be a polarized case of a candidate being either a good or bad actor, or processing transactions well or poorly. What if the candidate is proposing to use transaction fees for a certain controversial purpose? It is easy to imagine a scenario whereby the candidate incentivizes 40% to vote for him/her by some means, and this 40% are not opposed to the way the fees will be used. But what if the other 60% are fundamentally against the proposed use of the fees? Should they not be given the chance to voice their concerns?
Could not "vote against" potentially be a useful, if not necessary, tool? If the majority opposes something in principle, shouldn't there be a proper way for them to oppose it in practice?
I admit I haven't thought much about this, and also haven't followed closely enough the way voting works with DPOS, so I may be mistaken.
However, I thought I should post this in case it is useful, or in case it may act as a catalyst in some way.
My previous post was attempting to argue that we should be allowed to vote against a candidate. This is to avoid a situation where a candidate strongly opposed by the majority can be more highly ranked than one who is not. I haven't seen a clear argument against the case which I made above. Was my reasoning wrong? Are there proposed benefits of NOT allowing votes "against"?
Editing to add:
This is the reason I'm making these posts:
Agent86 is very persistent and talked with me on Skype last night to argue against negative votes.
The primary argument he made was that abstaining is the responsible thing to do for uninterested users and that large stake holders will be very pro-active.
Small stake holders collectively do not have much influence.
I have decided to make it an option that is easy to turn on in the future with a hard-fork, but off by default. If it becomes a problem then the delegates can vote to support it.
-
Suppose Bad Actor has 25% approval, everyone else has to find someone they can agree to give 26% approval to bump the Bad Actor.... there are 2 solid candidates and they each split the vote 13% / 13% and thus neither is able to bump the Bad Actor.
With approval voting this only happens if everyone has already maxed out their 101 votes, no?
-
Agent86 is very persistent and talked with me on Skype last night to argue against negative votes.
The primary argument he made was that abstaining is the responsible thing to do for uninterested users and that large stake holders will be very pro-active.
Small stake holders collectively do not have much influence.
I have decided to make it an option that is easy to turn on in the future with a hard-fork, but off by default. If it becomes a problem then the delegates can vote to support it.
I do not get the reasoning behind turning it off by default.
All the above-listed arguments look like speculations for me.
Perhaps there was something more meaningful in your private chat ?
-
Agent86 is very persistent and talked with me on Skype last night to argue against negative votes.
The primary argument he made was that abstaining is the responsible thing to do for uninterested users and that large stake holders will be very pro-active.
Small stake holders collectively do not have much influence.
I have decided to make it an option that is easy to turn on in the future with a hard-fork, but off by default. If it becomes a problem then the delegates can vote to support it.
I do not get the reasoning behind turning it off by default.
All the above-listed arguments look like speculations for me.
Perhaps there was something more meaningful in your private chat ?
I'm with Emski on this.
I also think that automatic voting based upon performance/shady behavior should be implemented in the client. If someone wants to manually manage their votes, they should be free to turn this off.
-
Considere something else:
Maybe deppended on % delegate participation to remove the last active delegates on every round... (NOT RANDOM)
if 80% participation, remove the last 20% of active delegates and replace them with the first delegates in the standby list...
if 95% participation, remove the last 5% of active delegates and replace them with the first delegates in the standby list...
if 60% participation, remove the last 40% of active delegates and replace them with the first delegates in the standby list...
etc. etc.
normaly the participation % will go up until it reach a equillibrium point that is the highest possible depended on delegates average accuracy...
and you give oportunitys for potential better delegates to get active !!!
Can you define % participation? Is it for the last round of 101 blocks, is it per candidate? Do you mean reliability for each delegate?
I agree a system which gives a delegate who fails to produce block a time out for awhile would be best.
e.g. if there are 5 consecutive missed blocks - that could be anything, could be that their network is just down, even if they have a bad ass equipment, why should they continue to get bad statistics, when they could be removed temporarily, say for an hour or two.
-
After doing much thinking and learning from experience with recent dry runs I have concluded that we need to have negative votes.
A negative vote is morally equivalent to a vote of approval for "everyone but X" like a positive vote is the moral equivalent as a vote of disapproval for everyone but Y.
In a world where the squeaky wheal gets the grease, it is far more likely that the average user will know more about who they are against than who they are for. When there is a bad actor it is far more difficult to get people to up-vote 101 other candidates than down vote the guy causing the problem. For automated voting it is also far easier to down vote delegates that are not producing blocks than to decide who to up vote.
The average user is lazy and unlikely to evaluate 101 different individuals. We must factor in this laziness in our designs.
Lastly in a world with 1000 candidates, it takes a lot of work to concentrate the approval consensus. Suppose Bad Actor has 25% approval, everyone else has to find someone they can agree to give 26% approval to bump the Bad Actor.... there are 2 solid candidates and they each split the vote 13% / 13% and thus neither is able to bump the Bad Actor. Everyone can agree the Bad Actor should go, so they simply vote against him.
Negative voting is effectively an easy way to compress a "vote for all, but 1" which in a way helps simulate the ability to approve more than 101 delegates.
+5%
-
Considere something else:
Maybe deppended on % delegate participation to remove the last active delegates on every round... (NOT RANDOM)
if 80% participation, remove the last 20% of active delegates and replace them with the first delegates in the standby list...
if 95% participation, remove the last 5% of active delegates and replace them with the first delegates in the standby list...
if 60% participation, remove the last 40% of active delegates and replace them with the first delegates in the standby list...
etc. etc.
normaly the participation % will go up until it reach a equillibrium point that is the highest possible depended on delegates average accuracy...
and you give oportunitys for potential better delegates to get active !!!
Can you define % participation? Is it for the last round of 101 blocks, is it per candidate? Do you mean reliability for each delegate?
I agree a system which gives a delegate who fails to produce block a time out for awhile would be best.
e.g. if there are 5 consecutive missed blocks - that could be anything, could be that their network is just down, even if they have a bad ass equipment, why should they continue to get bad statistics, when they could be removed temporarily, say for an hour or two.
I mean % participation for the last round of 101 blocks...
but you are right it makes better sense that the system removes temporary only delegates that missed blocks in a row
or even better if it removes delegates which reliability % for the last x blocks they produced, (compared with their average reliability) drops to much (for example 30%)
-
Any system that allows delegates to be removed by any method other than shareholder action opens the network up for attack because the "attack chain" could simply use that same logic to boot all the good guys. A large part of our security comes from the fact that shareholders must vote to change out the delegate pool. This forces the longest chain to always be the chain with the most shareholder participation.
-
I agree. I think negative votes will open another can of worms - people will pay for negative votes. There are times where they seem much needed though.