RE: 100 DAYS OF STEEM : Day 33 - Tackling Abuse on Steem - Part I - What is Abuse?

You are viewing a single comment's thread from:

100 DAYS OF STEEM : Day 33 - Tackling Abuse on Steem - Part I - What is Abuse?

in the100daysofsteem •  5 years ago 

Limiting the number of posts per account will hit very active users and community bots as well.

Instead of limiting the number of posts one could introduce 'diminishing returns' which means one still could write as many posts as one wishes, but starting from a certain number of posts upvotes on every further posts would have a weaker effect than upvotes on previous posts. That means a very active author would earn less per post with an increasing number of posts (this effect could for example start after his second post per day) but still could earn anything.

Similarly one could try to prevent the effectivity of upvoting the same users again and again (circle voting). How about if after each vote on a specific account (including one's own account) each further vote on the same account would lead to significantly less curation reward for the voter and less profit for the upvoted account? Thus, when upvoting an account which I had already upvoted before, my voting power would be smaller than in case I upvote an account which I didn't upvote before.
If I upvote for example my wife, my daughter or my cat more than twice per day every following upvote for the same account would get weaker an weaker (like voting power is exhausting anyway already, just faster).

Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!
Sort Order:  

Instead of limiting the number of posts one could introduce 'diminishing returns' which means one still could write as many posts as one wishes, but starting from a certain number of posts upvotes on every further posts would have a weaker effect than upvotes on previous posts.

This would work, but it would come with a side effect. If I would create a few shit posts before posting a real masterpiece of a post, the good post might receive less rewards than the bad ones :)

Similarly one could try to prevent the effectivity of upvoting the same users again and again (circle voting). How about if after each vote on a specific account (including one's own account) each further vote on the same account would lead to significantly less curation reward for the voter and less profit for the upvoted account?

As we discussed many times over the years, in my eyes this would be the ultimately perfect solution. Using something similar to the calculated CSI on SteemWorld as a factor, so that the voted rshares would be multiplied by it prior to subtracting them from the pool and adding them to the active votes.

This would not eliminate all cases of abuse (for example, if someone owns many accounts with much SP and votes each day with a different one), but it would work very well for most common cases.

This would work, but it would come with a side effect. If I would create a few shit posts before posting a real masterpiece of a post, the good post might receive less rewards than the bad ones :)

Then just publish the "shit posts" after the masterpiece? ;-) (Of course also the ability to receive full votes for posts would recover again after some time.)

Or maybe not post any "shit posts" at all? Seriously, if you yourself think a post is a shit post, then, it is a shit post...

You won't find any "shit post" from me (but sometimes kinds of humorous meant replies).

"Using something similar to the calculated CSI on SteemWorld as a factor, so that the voted rshares would be multiplied by it prior to subtracting them from the pool and adding them to the active votes."

Sorry but I think that it's not a good idea to use CSI on SteemWorld as in my opinion CSI isn't a reliable factor.

If I see that an account with 100% self-voting (no vote to anybody else just vote for the own account) has an CSI from 0.0 but other accounts has negative CSI, f.e. my account in the moment, than Steemworld consider it as better to vote 100% only your own account instead of voting for others.

How this can be a reliable factor ?

I think as long as you can claim for every approximately 5000 Steempower one account a week I don't think that this solution will work.
If a whale has 500.000 Steempower he can just claim approximately 100 new accounts at once and than post in every account only one post every week.
So in this case he could always get full rewards for selfvoting.

He wouldn't be able to handle all these accounts in a way that every of them could create more or less reasonable comments/posts. However, pure automated comments, created for farming puposes only, could easily be detected and flagged by members of an implemented anti abuse committee, which I suggested, as well.

OK I wrote about 100 acconts.
You are right this is really hard to handle.
But if you allow one vote in full strengh every day it would be also ok to use only 10 accounts.
Every day one post in every of the 10 accounts and the other 9 vote this post.

I think it's not much difference in time to post 10 times a day in one account or to post in 10 accounts only once a day, isn't it ?

It's rather easy to spot (even by automated algorithms) if these ten accounts were only interested in upvoting each other instead of upvoting other users as well ...

For example also Voting CSI in SteemWorld would be very low.

haha, so I am the worsest guy at all cause my Voting CSI is negativ (-0,6)

Indeed, just another evidence of how well these kinds of algorithms are working. ;-)

  ·  5 years ago (edited)

Yes with 100% selfvoting you have 0.0 (like your friend hae...) so I am worse cause I have 0.0% selfvoting and - 1,1.

OK so your opinion is that selfvoting is better than votings others.

  ·  4 years ago (edited)

My guess is that the algorithm doesn't differenciate between 'self-voting' and upvoting 'other' accounts. It matters how often and strong certain accounts are getting upvoted (and also how many different accounts are receiving upvotes).
The algorithm cannot know anyway if the upvoted accounts belong to the same user or not.
I didn't write the algorithm of Voting CSI, so I can only guess how it works. It could well be that it also checks how often these upvoted accounts upvote each other (if they are building a closed group which is mutually upvoting each other) and how diverse the upvotes are which they receive from other users (in this aspect @haejin beats you).
All in all it does a great job to detect selfish voting behaviour. YOU and @haejin fully deserve the low rating. :)

Haha - I even get worse than I was before, now -1,1 :-)
Such I bad boy I am.

Instead of limiting the number of posts one could introduce 'diminishing returns' which means one still could write as many posts as one wishes, but starting from a certain number of posts upvotes on every further posts would have a weaker effect than upvotes on previous posts. That means a very active author would earn less per post with an increasing number of posts (this effect could for example start after his second post per day) but still could earn anything.

Completely useless. Alts can be used to circumvent that and frequently are by abusers.

Nothing wrong with plain old PoB. But it takes an active anti-abuse community to do curb abuse. Not even Steemit, Inc with its mightly voting power can curb anything if there is no anti-abuse community to constantly bring abuse to its attention.

I'm on Hive but I want Steem to succeed also. In fact, forks are good in the world of DPoS because that's how things decentralise . Each project takes a different direction and attracts different people.

Completely useless. Alts can be used to circumvent that and frequently are by abusers.

NOT completely useless!

I am pretty sure that most current abusers wouldn't take the effort to create that many alt accounts which are necessary to cirumvent this hurdle.

In addition it would be very easy to spot these accounts it they weren't active themselves and only received upvotes from one single abuser account (or circle upvote each other).

Concerning the committee, that's my idea since a long time.

NOT completely useless!

I am pretty sure that most current abusers wouldn't take the effort to create that many alt accounts which are necessary to cirumvent this hurdle.

It would only take creating one alt to double the number of daily posts one can make from ten to twenty.

In addition it would be very easy to spot these accounts it they weren't active themselves and only received upvotes from one single abuser account (or circle upvote each other).

That can easily be circumvented, too. Add a bit of randomization in the process and the detection becomes much harder. Add a few legit posts in the mix to attract votes from others.

Concerning the committee, that's my idea since a long time.

Nothing wrong with a committee.

It would only take creating one alt to double the number of daily posts one can make from ten to twenty.

I also suggested to reduce the number of fully rewarded posts per day. In the early STEEM days this number was actually four per day.
The combination of these two suggestions would be rather effective in my opinion.

That can easily be circumvented, too. Add a bit of randomization in the process and the detection becomes much harder. Add a few legit posts in the mix to attract votes from others.

I disagree. That's all rather effortful and most abusers wouldn't do that. In the past it was very easy to spot the majority of abusers just by checking their 'Voting CSI' in SteemWorld.
For the remaining rest the 'committee' could be responsible.

Sure it can be tried. Time will tell if it works.

Another interesting related idea is a tax on too many rshares spent on the same accounts within a time period. That would force users to either create a lot of alts, which would cost them money in the form of account creation, or actually motivate them to look for more users to curate, which would be an excellent thing for user retention.