The progression of this network is staggering and saddening to look at.
The points identified on how to find (through code) the related accounts are reasonable. When you can publish the script, it would be great to see it. We are going to generate a library of anti-abuse tools to increase the efficiency of our abuse fighters.
Indeed, everyone can help tackle abuse. Downvoting abuse and reporting them are the most common ones. Raising awareness on this 3rd option of removing Steemit Inc delegations is another.
What a great start on your first post for the anti-abuse category. I'm sure we'll see more excellent posts from you in this category.
Your contribution has been evaluated according to Utopian policies and guidelines, as well as a predefined set of questions pertaining to the category.
To view those questions and the relevant answers related to your post, click here.
What a great post @crokkon. You did an awesome job and what a great contribution to the antiabuse Initiative. I want to personally thank for you making the pull request to remove the faucet delegations from this vote farm.
My eyes watered up when I got to that part of the post. It must have suddenly got dusty in my office? LOL
I have worked with @themarkymark on some vote farms and the SDL. Its very rewarding when you see those delegations removed! He does more than most on the platform to combat abuse.
More of us need to learn how to make pull requests.
Thanks for fighting abuse thanks for your help with the SFR bot.
I hope to see more of this kind of contribution in utopian antiabuse. With all the data transparent and available from the blockchain we should be able to make good use of pattern recognition tools to uncover persistent abuse.
I think a big data approach. Throw all the blockchain data into a pool. See what patterns past abuse cases throw up. Use these to find and predict new abuse at early stages.
Thanks @miniature-tiger! In this case, it was indeed very clear. However, not all are that clear. Having the blockchain data helps a lot, though. Pattern recognition is an interesting approach, and also ML as mentioned in @lovenfreedom's comment. I tried once to analyze only a few days in the context of "who votes mostly together with whom" in a brute-force approach - This gave a few interesting results, but the processing time requirements would not have scaled to larger time ranges. Applying ML/AI techniques could indeed help there to build models...
Nothing against steemcleaners, the steemflagrewards folks, or anyone else who uses their stake to genuinely flag bad actors (as opposed to those who seem to use them for retaliatory purposes, or just because), but this—removing Steemit delegated SP in the nascent stages of accounts—to me, is the most effective way to fight spam.
While I know there is plenty more stake that could get involved in the flagging of rewards if they so chose, as it stands, there is still a goodly amount of resources and man hours that go into knocking down these accounts who only seem to find ways to adapt and continue more than they fade away. Or the same bad actors open more accounts and start all over again in the blockchain version of cat and mouse.
So, thank you for this post. I'm glad you're able to identify these folks and Steemit is willing and able to take pretty swift action against them.
It definitely does. In my mind, that speaks to the inefficiency of the flagging tool as a means of actually stopping these guys, as opposed to having their accounts essentially powered down. No SP, No RCs. No SP, no VP. And it doesn't require any resources or man hours be spent at all after that, both of which could actually be turned around and used for something else.
STINC’s sign-up process has been one of the most exploitable processes in this ecosystem. They routinely allow exploiters to create hundred- or thousand-account bot-nets. One was able to grow to over 20,000 accounts.
They need to get out of the account creation game until they have at least some kind of adequate measures to prevent obvious exploits. It’s sad that this continues. And it’s even sadder that they wanted the ability to create these accounts for free...so that they can assume zero responsibility AND pay no price for it.
good point about the free account creation mechanism. While it costed at least 0.1 STEEM before, it is fully free now. I guess the whole signup process is a tricky balance between convenience for genuine users and public perception one one hand, and a burden for abusers on the user hand. But seeing wait times of several weeks isn't the best advertisement anyway. Apart from abuse prevention, the new account creations are currently at an year to date low while still taking most of @steem's RC - I wonder to which extend the account creation process could scale at all?
Nice work @crokkon :) As a follow up analysis I would be curious to see (and I think it would add a lot of context to this) the earnings from this farm over time as well as the number of accounts. How much Steem/SBD was withdrawn from network? And if you do look at that, it would also be interesting to see if there was a noticeable reduction in income potential after HF20 and the dust vote shift.
Thanks, Carl! These are some interesting ideas for a follow-up :) The number of accounts and the growth over time is shown in a graph above. This network had author rewards of around 8 SBD + 10 SP in its last week before undelegation, plus a bit from involved accounts that have a few hundred own SP. The 160 SP after undelegation give a rough picture of the "lifetime earnings". Good point about the dust vote shift - living off hundreds of 15 SP accounts which suddenly have a vote "worth" only 13.8 SP per vote certainly makes a difference. However, the rewards in general changed quite considerably over the HF, so the plain numbers would probably give a biased picture there.
That's a great result and I appreciate a lot of work is involved. Steemit should really be doing this themselves as it's their delegation that's being abused. I realise it's almost impossible to stop fans accounts being created, but bad behaviour like this can be detected automatically and dealt with. U until they do it we need people like you. Thanks
Thanks @steevc! It's true that this should actually be Steemit's job to find those or avoid the situation in the first place. Having the delegations removed only a few hours after reporting the accounts suggests to me that helping out is at least appreciated.
Thanks for the resteem!
Thank you for this awesome contribution @crokkon.
The progression of this network is staggering and saddening to look at.
The points identified on how to find (through code) the related accounts are reasonable. When you can publish the script, it would be great to see it. We are going to generate a library of anti-abuse tools to increase the efficiency of our abuse fighters.
Indeed, everyone can help tackle abuse. Downvoting abuse and reporting them are the most common ones. Raising awareness on this 3rd option of removing Steemit Inc delegations is another.
What a great start on your first post for the anti-abuse category. I'm sure we'll see more excellent posts from you in this category.
Your contribution has been evaluated according to Utopian policies and guidelines, as well as a predefined set of questions pertaining to the category.
To view those questions and the relevant answers related to your post, click here.
Need help? Write a ticket on https://support.utopian.io/.
Chat with us on Discord.
[utopian-moderator]
Thanks a lot, @lovenfreedom! I heard about your ideas to build a library of anti-abuse tools and I'd be glad to contribute a few bits and pieces! :)
You are simply an amazing person. Thank you for all you do for us!
Thank you for your review, @lovenfreedom! Keep up the good work!
What a great post @crokkon. You did an awesome job and what a great contribution to the antiabuse Initiative. I want to personally thank for you making the pull request to remove the faucet delegations from this vote farm.
My eyes watered up when I got to that part of the post. It must have suddenly got dusty in my office? LOL
I have worked with @themarkymark on some vote farms and the SDL. Its very rewarding when you see those delegations removed! He does more than most on the platform to combat abuse.
More of us need to learn how to make pull requests.
Thanks for fighting abuse thanks for your help with the SFR bot.
Thanks a lot, @iamstan, highly appreciated!
Great work @crokkon! A very clear case.
I hope to see more of this kind of contribution in utopian antiabuse. With all the data transparent and available from the blockchain we should be able to make good use of pattern recognition tools to uncover persistent abuse.
Totally agree.
I was wondering if this is something ML (Tensor Flow and the like) can be applied?
I think a big data approach. Throw all the blockchain data into a pool. See what patterns past abuse cases throw up. Use these to find and predict new abuse at early stages.
Thanks @miniature-tiger! In this case, it was indeed very clear. However, not all are that clear. Having the blockchain data helps a lot, though. Pattern recognition is an interesting approach, and also ML as mentioned in @lovenfreedom's comment. I tried once to analyze only a few days in the context of "who votes mostly together with whom" in a brute-force approach - This gave a few interesting results, but the processing time requirements would not have scaled to larger time ranges. Applying ML/AI techniques could indeed help there to build models...
I'm hopeful that this is exactly what the anti-abuse team are looking for in regards to research and presentation of contribution. Nice!
100% - the graphs were very helpful.
I was really happy to see his contribution. This is another flavor of antiabuse contribution.
We need to look at how @pibarabot creates his life-feed of flag wars and see how we can apply that to faucet upvoting networks.
Thanks for your support to the anti-abuse community.
@pibarabot is not a bad place to look for drama, and sometimes hidden abusive behaviors.
thanks Asher!
Hey, @crokkon.
Nothing against steemcleaners, the steemflagrewards folks, or anyone else who uses their stake to genuinely flag bad actors (as opposed to those who seem to use them for retaliatory purposes, or just because), but this—removing Steemit delegated SP in the nascent stages of accounts—to me, is the most effective way to fight spam.
While I know there is plenty more stake that could get involved in the flagging of rewards if they so chose, as it stands, there is still a goodly amount of resources and man hours that go into knocking down these accounts who only seem to find ways to adapt and continue more than they fade away. Or the same bad actors open more accounts and start all over again in the blockchain version of cat and mouse.
So, thank you for this post. I'm glad you're able to identify these folks and Steemit is willing and able to take pretty swift action against them.
They have gotten better, but it does feel like a bit too little given how long some of these guys have been abusing.
Hey, @enforcer48.
It definitely does. In my mind, that speaks to the inefficiency of the flagging tool as a means of actually stopping these guys, as opposed to having their accounts essentially powered down. No SP, No RCs. No SP, no VP. And it doesn't require any resources or man hours be spent at all after that, both of which could actually be turned around and used for something else.
STINC’s sign-up process has been one of the most exploitable processes in this ecosystem. They routinely allow exploiters to create hundred- or thousand-account bot-nets. One was able to grow to over 20,000 accounts.
They need to get out of the account creation game until they have at least some kind of adequate measures to prevent obvious exploits. It’s sad that this continues. And it’s even sadder that they wanted the ability to create these accounts for free...so that they can assume zero responsibility AND pay no price for it.
good point about the free account creation mechanism. While it costed at least 0.1 STEEM before, it is fully free now. I guess the whole signup process is a tricky balance between convenience for genuine users and public perception one one hand, and a burden for abusers on the user hand. But seeing wait times of several weeks isn't the best advertisement anyway. Apart from abuse prevention, the new account creations are currently at an year to date low while still taking most of @steem's RC - I wonder to which extend the account creation process could scale at all?
Nice work @crokkon :) As a follow up analysis I would be curious to see (and I think it would add a lot of context to this) the earnings from this farm over time as well as the number of accounts. How much Steem/SBD was withdrawn from network? And if you do look at that, it would also be interesting to see if there was a noticeable reduction in income potential after HF20 and the dust vote shift.
If you look at the post by @supposer here: https://steemit.com/utopian-io/@supposer/how-steemit-is-infiltrated-by-abusive-account-networks
Some networks do get big enough that they become dolphins and use their ill-gotten SP to spam farm. By that point, it's near impossible to stop them.
Thanks, Carl! These are some interesting ideas for a follow-up :) The number of accounts and the growth over time is shown in a graph above. This network had author rewards of around 8 SBD + 10 SP in its last week before undelegation, plus a bit from involved accounts that have a few hundred own SP. The 160 SP after undelegation give a rough picture of the "lifetime earnings". Good point about the dust vote shift - living off hundreds of 15 SP accounts which suddenly have a vote "worth" only 13.8 SP per vote certainly makes a difference. However, the rewards in general changed quite considerably over the HF, so the plain numbers would probably give a biased picture there.
That's a great result and I appreciate a lot of work is involved. Steemit should really be doing this themselves as it's their delegation that's being abused. I realise it's almost impossible to stop fans accounts being created, but bad behaviour like this can be detected automatically and dealt with. U until they do it we need people like you. Thanks
Thanks @steevc! It's true that this should actually be Steemit's job to find those or avoid the situation in the first place. Having the delegations removed only a few hours after reporting the accounts suggests to me that helping out is at least appreciated.
Thanks for the resteem!
A few hours is a pretty decent turnaround I think, and so they must be fairly happy with the lists of accounts coming in :)
Thank you for your work. tip! 1
Posted using Steeve
This post was shared in the Curation Collective Discord community for curators, and upvoted and resteemed by the @c-squared community account after manual review.