You are viewing a single comment's thread from:

RE: Ideas for Future Rule Changes - Voting, Earnings, Maximum Social Benefits - a Discussion Document

in #steemit7 years ago

Therefore, there can be a rule that voting for the same user, whether oneself or another, will decrease the power of each subsequent vote, within a limited period.

Really interesting ideas and I think it's a huge contender for dealing with the self voting problem. It's also very comprehensive because it simultaneously changes the incentive for so-called "circle jerk" behavior, which is a similar issue to self voting via other accounts (sock puppets).

Looking forward to hearing more debate on this but at present I actually think this is the best idea going for a blockchain system level solution to self voting.

💯 😎

Sort:  

Brainstorming: I wonder what would happen if you just discard the highest value vote and lowest value vote and calculate the post's payout based on the remaining N-2 votes (but still give curation rewards to all voters). This is how they scored kata in my son's karate tournaments.

  1. A post would need 3 votes to get a reward.
  2. Self-votes would only pay off if they were "ratified" by a higher value voter and a lower value voter.
  3. It's still vulnerable to "gaming" by multiple accounts, but - unless ratified by others - at the cost of the sacrificed steem power in the highest value account doing the gaming.
  4. Not sure how this would incentivize/disincentivize whale voting.

I have used this technique in the past but it works when people are voting up and down, or giving stars out of 5. To stop gangs both upvoting and downvoting, I shaved off the top 10% and bottom 10% and calculated the average from the middle 80%. However, this algo was kept secret from users ;-)
In this case, I wonder if the loss of a genuine large upvote is too much of a price to pay.
Thanks for your thoughts.

Interesting. I'll think about this. Also really interesting that it has an investigable use already in Karate.

My first thought is that it would be effective in reducing the reward on gaming. That largest account in particular will lose its voting power on this, which would be the highest SP sock puppet account, and wouldn't seem attractive to someone trying to maximize their take, they might just get back into curation.

Second thought is that people, especially minnows, will probably bemoan the loss of that largest vote. It may make the idea seem like more elitist meddling.

Okay last one, third thought is that I wonder what the psychological effect on whales will be when they come in to gift a post with a high powered vote which now has no effect. The UIs will need to be updated, but if a new vote is the highest vote the estimated payout would not change!

Okay last one, third thought is that I wonder what the psychological effect on whales will be when they come in to gift a post with a high powered vote which now has no effect. The UIs will need to be updated, but if a new vote is the highest vote the estimated payout would not change!

Assuming the post had 2 or more other voters, the estimated payout would still change, but the new value would be based on the value from the previous highest voter, so yeah - it would be a very different experience.

Maybe there's not really much of a point in removing the lowest value vote in this context. If you remove just the highest, it might work out to be something like a second-price auction. I don't know much about them, except that they're used for digital advertising by Google and others. However, according to this link:

The theoretical nicety of second price auctions, first pointed out by William Vickrey, is that bidding one's true value is a dominant strategy.

This doesn't sound appealing to a minnow who's biggest vote is booster, or if they are really lucky, a
Dolphin/whale who is interested in what they read/see.

Yes, voting-for-self and voting-for-the-same-other are the same thing in this proposed algorithm.

It is also an important point that such behaviour cannot be totally stopped - there are also some legitimate reasons for it - and so the algo does not try to stop it, just to reduce its effect.

Sure. To me reducing its effect is completely enough. Right now it's out of control.