5 ms·

I don't understand how quadatic voting can deal with collusion. Say I want to give 4 credits to A, and a friend wants to give 4 credits to B. So each of us give

I don't understand how quadatic voting can deal with collusion. Say I want to give 4 credits to A, and a friend wants to give 4 credits to B. So each of us gives 2 votes to their preference. But if we join together, we each give 2 credits to each option, that is sqrt(2) votes. So each option gets 2*sqrt(2) ~= 2.8 votes instead. So we are strongly incentived to found a party and to pool our votes. I'm not sure this is what people intended. It also seems to punish caring strongly for a certain issue, whereas I wonder if that isn't maybe an indicator that you are informed about a topic, and thus your vote should count more rather than less?

ekelsen 5y ago
Collusion is prevented in the same way that vote buying is prevented now -- you can't verify how other people have voted. This means that you could tell your friend you'll vote for B, but actually still vote for A and A will get more votes. Since you can't verify that collusion has worked, the best strategy is to not do it.
- samatman 5y ago
  I wouldn't be so quick to dismiss here. Collusion in 'linear' voting loses the voter their preference in exchange for some reward, double-dipping is a simple defection. GP is talking about personal force multiplication through collaboration, it's win-win so you have a genuine prisoner's dilemma. Being unable to verify cooperation doesn't mean it won't happen in other ways. I don't know that this is actually bad just that a simple comparison to linear votes doesn't hold up.
  - ekelsen 5y ago
    Let's say we agree to cooperate because it's win-win. Then I go into the voting booth, do I have any incentive _at all_ to not vote for my preferences as opposed to yours? No. Why would I not defect and gain more votes for my preferred causes?
    - breuleux 5y ago
      Some people don't like lying about what they are doing, even if they can't get caught. Small scale cooperation between honest people would probably work for that reason. On a large scale, it probably wouldn't be very efficient, but insofar that both sides would defect or cooperate at roughly the same rate, you would still expect it to be somewhat beneficial if the cooperation rate is nonzero (but maybe not enough to be worth the trouble).
      - dane-pgp 5y ago
        > Some people don't like lying about what they are doing, even if they can't get caught. Unfortunately, some people do like lying, or at least they don't mind it as much, and if you create a system which rewards liars more than honest people, you shouldn't be surprised if the liars end up winning and you create more liars.
    - samatman 5y ago
      My "defects reliably when protected by incomplete information" T-shirt is raising a lot of questions already answered by my T-shirt?
    - comex 5y ago
      This is the same situation as prisoner’s dilemma and has the same possible answers. One is, as others have said, that some people will choose honesty for personal moral reasons. Another is superrationality: https://en.wikipedia.org/wiki/Superrationality#Prisoner's_dilemma https://en.wikipedia.org/wiki/Superrationality#Prisoner's_di...
    - notahacker 5y ago
      In certain scenarios (bloc votes) the defection is fairly obvious afterwards, so the standard conclusion that always defecting is a bad strategy if the prisoner's dilemma is iterated and [probable] defection can be established over time applies. The quadratic maths tends towards making the gains from collusion very large, which means there is incentive not to defect even if detection is uncertain. And bloc votes and ballots which aren't necessarily secret do represent some of the more plausible use cases for quadratic voting (corporate governance)
  - lukifer 5y ago
    Preference collusion in linear voting isn't completely unheard of. I've seen projects where voters in safe states swap with voters in a swing state [0]. (Obviously these are unverified, so they still contain a prisoner's dilemma.) [0] https://en.wikipedia.org/wiki/Vote_pairing#United_States_presidential_election,_2016 https://en.wikipedia.org/wiki/Vote_pairing#United_States_pre...
Ar-Curunir 5y ago
It's a good point; this article goes into detail about this issue in existing deployments, and potential solutions: https://vitalik.ca/general/2021/04/02/round9.html https://vitalik.ca/general/2021/04/02/round9.html
csense 5y ago
I came here to say this. The voting scheme described isn't Pareto efficient; it's possible for pairs of participants to make positive-sum private deals. In the base scenario, Alice votes for A with 4 credits and Bob votes for B with 4 credits. A, B each gets 2 votes. In the private deal scenario, Alice/Bob agree to each vote for A, B with 2 credits each. Then A, B each gets 2.818 votes.

ekelsen 5y ago

Collusion is prevented in the same way that vote buying is prevented now -- you can't verify how other people have voted. This means that you could tell your friend you'll vote for B, but actually still vote for A and A will get more votes. Since you can't verify that collusion has worked, the best strategy is to not do it.

samatman 5y ago
I wouldn't be so quick to dismiss here. Collusion in 'linear' voting loses the voter their preference in exchange for some reward, double-dipping is a simple defection. GP is talking about personal force multiplication through collaboration, it's win-win so you have a genuine prisoner's dilemma. Being unable to verify cooperation doesn't mean it won't happen in other ways. I don't know that this is actually bad just that a simple comparison to linear votes doesn't hold up.
- ekelsen 5y ago
  Let's say we agree to cooperate because it's win-win. Then I go into the voting booth, do I have any incentive _at all_ to not vote for my preferences as opposed to yours? No. Why would I not defect and gain more votes for my preferred causes?
  - breuleux 5y ago
    Some people don't like lying about what they are doing, even if they can't get caught. Small scale cooperation between honest people would probably work for that reason. On a large scale, it probably wouldn't be very efficient, but insofar that both sides would defect or cooperate at roughly the same rate, you would still expect it to be somewhat beneficial if the cooperation rate is nonzero (but maybe not enough to be worth the trouble).
    - dane-pgp 5y ago
      > Some people don't like lying about what they are doing, even if they can't get caught. Unfortunately, some people do like lying, or at least they don't mind it as much, and if you create a system which rewards liars more than honest people, you shouldn't be surprised if the liars end up winning and you create more liars.
  - samatman 5y ago
    My "defects reliably when protected by incomplete information" T-shirt is raising a lot of questions already answered by my T-shirt?
  - comex 5y ago
    This is the same situation as prisoner’s dilemma and has the same possible answers. One is, as others have said, that some people will choose honesty for personal moral reasons. Another is superrationality: https://en.wikipedia.org/wiki/Superrationality#Prisoner's_dilemma https://en.wikipedia.org/wiki/Superrationality#Prisoner's_di...
  - notahacker 5y ago
    In certain scenarios (bloc votes) the defection is fairly obvious afterwards, so the standard conclusion that always defecting is a bad strategy if the prisoner's dilemma is iterated and [probable] defection can be established over time applies. The quadratic maths tends towards making the gains from collusion very large, which means there is incentive not to defect even if detection is uncertain. And bloc votes and ballots which aren't necessarily secret do represent some of the more plausible use cases for quadratic voting (corporate governance)
- lukifer 5y ago
  Preference collusion in linear voting isn't completely unheard of. I've seen projects where voters in safe states swap with voters in a swing state [0]. (Obviously these are unverified, so they still contain a prisoner's dilemma.) [0] https://en.wikipedia.org/wiki/Vote_pairing#United_States_presidential_election,_2016 https://en.wikipedia.org/wiki/Vote_pairing#United_States_pre...

Ar-Curunir 5y ago

It's a good point; this article goes into detail about this issue in existing deployments, and potential solutions: https://vitalik.ca/general/2021/04/02/round9.html https://vitalik.ca/general/2021/04/02/round9.html

csense 5y ago

I came here to say this. The voting scheme described isn't Pareto efficient; it's possible for pairs of participants to make positive-sum private deals. In the base scenario, Alice votes for A with 4 credits and Bob votes for B with 4 credits. A, B each gets 2 votes. In the private deal scenario, Alice/Bob agree to each vote for A, B with 2 credits each. Then A, B each gets 2.818 votes.