Steembay (a bot) under attack by a wannabe AI
Some of you may have heard about the sadkitten project. The idea is to countervote massive selfvoting and "educate" people to "give back" to the community. This is not entirely a bad idea. We at Steembay believe that everybody should do with his stack as he thinks is best. Sure there are good ways and bad ways to spend Voting power, but as this is strictly subjective, most of these projects centered around policing content and voting behaviour are more or less futile.
*@sadkitten and @zer0hedge ??
Identifying spam or plagiarism as example is a very important task.
Steembay is working and running now for 9 months, providing a service that allows you to auction away or buy goods, doing business, selling and buying services and lately it is a main way to figure out the value of the @steemmonsters.
Since we started we organized several thousand auctions and tens of thousands of SBD changed owners.
This means, Steembay is one of best ways to get involved with the Steem Blockchain and to build a Stack.
We know our system has flaws and we are working on it as often as it is possible. And since the beginning we are
100% selfvoting
We do this to finance the FREE Service, that is dependent on donations and upvotes and our selfvotes. Also it is used to provide a better overview to the auction process, which is also quite important.
So @sadkitten is true about the fact, that we are probably the worst and most evil selfvoters on the platform. Since also @zer0hedge started to downvote us, it REALLY hurts not only our rewards, but also our reputation. He has a reputation of 73 !
Unfortunately @sadkitten does not react to my attempt to contact him on steem.chat. So I am asking the community for support.
Do you believe that Steembay is a Service that should go on, or do you believe that our insane amounts of 3 $ a day on average are undeserved and has to be downvoted / diminished ?
Until now we recieved 45 flags in 3 days...
Another question would be: Is there ANY point in an AI trying to teach a botsystem "good voting manners"?
i believe in the Project and it's totally okay for me, if you upvote your own posts. Sounds jealously for me ;-)
no worries. the Project deserved it
Delegating to bidbots has the exact same effect as selfvoting and should be punished just the same, then. Just saying ...
From what I can see, @zer0hedge has not much SP.
How much did the flags cost you ?
It is not much... about 3 $ in 3 days. But the reputation-loss bothers me.
Also it's the "feeling" these downvotes cause.
It's not the exact same effect because of the curation reward.
Bidbots are about 85-90% as efficient as self-voting. I'd be curious of the current averages but it's variable and all over the place.
Or in other words, sadkitten ought to flag you with 85 - 90% flags then.
sadkitten attempts to vote to about 85% of the self voted amount anyway, so yes, that's consistent
you guys should be downvoting the people who delegate to bidbots with 90% of the strength you use on selfvoters. that would be consistent.
please read the context.
well said
One bot doesn't need to do everything, we are not the police, we're just one group expressing disagreement along certain lines. There is a perception among some people that if you do anything you should be doing everything. That does not make sense.
I responded to a similar comment before:
Along with @personz I'm the designer of @Sadkitten.
Unfortunately it's impossible to do anything about it without affecting the strength of the ideology.
There are easy way around it, maybe sharing away just one vote a week would be enough to have the kitten off your back for a while.
I don't think you realize the actual self-voting efficiency the bot it's currently countering.
Maybe one day @sadkitten will get to @haejin level and we'd have actual argument to present about having done something about people who are worst than him in terms of % without discriminating about his content.
Ps: @SadKitten is not an AI anymore, I've edited the profile.
Rather than downvoting, I would encourage you to support people, who stack up their SP and actually vote manually.
I have posted this many time before, but maybe you are not aware of it:
selfvoting would not be an issue like this, if rewards were not linear.
I don't think there is any better way to support people voting manually than what @sadkitten is doing.
Hit me up next Steemfest.
I can help you out with better ideas ;)
Post about it for everyone's benefit
You would not understand, anyways.
Try me
I am very glad, that reacted to this post. Thank you for that, also for the resteem.
It is very interesting that you call your approach an "ideology".
Or translated into understandable english: An ideology is more or less a pile of inconsistent BS.
Sorry for being so blunt about it. Imho most of todays problems result from "natural morals" turned into more and more complex "laws" without asking for common sense. The fact that @sadkitten isn't an AI anymore makes this even worse, because this means you made the decision to flag @steembay well knowing "morals" and able to consider what is helpful and what not.
There is no question, that massive selfvoting is damaging the overall value of STEEM. Providing a FREE service and financing it with a stack that is provided by the community is a completely different story.
If you don't see the difference and if you are not able to react accordingly, then you weakened your own "ideology" much more yourself than I ever would have been able to.
Fun Fact:
With a relevant Stack (or a nice delegation) we would PAY people to use steembay and upvote EVERY real auction. This is planned since we started the project. Unfortunately we earn less than a third world hourly with it, because the Service is widely ignored by those with power. You and your "Service" are slowing down reaching this goal even more.
I will not use the easy way to wiggle around the problem ( e.g. by voting for myself with steembay once a day ). Your approach reminds me in its mechanics too much the spain conquerers "missionizing" the Indians to Christianity.
I don't agree with the characterization that there is some stupid ideology behind the rationale for @sadkitten. The comparison to "missionizing" is poor. I don't think English is @transisto 's first language, but they can defend themselves for their own comments. I can only say that there is a consistent reasoning for @sadkitten if you care to find out.
Read this: https://steemit.com/steemit/@sadkitten/self-voter-return-on-investment-svroi-notoriety-flagging-bot
There is also plenty of discussion from people who disagree with it there.
@sadkitten could arguably still be an AI, I don't really know what @transisto meant. I am the technical lead for the project. What it does is a fully automated scan of the blockchain on a weekly basis to determine the most optimal self voter and then flags from highest to lowest the next week using whatever SP and accounts currently available.
The use of your SP for solely your own gain is up to you, as it is up to everyone. It's not however the vision of that platform that we agree with. Down votes AKA flags are a disagreement of rewards. We only ever down vote when there is a self vote, and only to 85%.
I never said, that it is a "stupid" ideology. The main point behind an ideology is that there are good ideas behind it, spoiled by a line that devides between "right" and "wrong" or "good" and "bad" without inducing common sense.
I think we both agree upon trying our best to make the platform "better". There is no doubt that in general self voting is harmful to the platform. When your approach is harming other projects, that without doubt make the platform better as well, you have to consider a better way (e.g. a whitelist ). The main point of an AI is to learn... If it is up to me to find or use a way to wiggle around your "ideology", then your ideology is weak and you are in fact "missionizing" for a good reason with the wrong means of doing it. I cannot imagine that this is what your ideology is aimed at.
PS: English is not my first language either. In fact it is my fourth...
You could also see our philosophy as community induced selfvotes, because if there is no auction started by a community member, there is no selfvote as well...
This way we can guarantee the service to stay free to use and to grow with each auction, even for those with no stack at all. An alternative would be to charge for each auction, which would ostracize everyone new to the platform from our service.
Anything done dogmatically is prone to massive error. Determining right and wrong however is not in itself problematic, only when done dogmatically, or without reference to reality and one's ever changing grasp of it.
We will not whitelist any account. You are contending that the means (self voting) are justifying the ends (supporting your project). Wouldn't everyone love to support their project this way? Why should yours in particular be viewed differently from any other account, any other project?
The main point of an AI is not to learn, it is to automate tasks. Perhaps you buy into the grander hype of AI and the coming robopocalypse, but sadkitten is what is technically termed narrow or weak AI. The problem domain is very small.
I only said that English is probably not @transisto 's first language to broaden consideration of the word "ideology" which is a trigger for you. Again, he can explain the usage himself, but again, I disagree with it.
How you fund your operations is your business. If you self vote optimally sadkitten will be interested in you. It is our view that Steem is not a piggy bank for entrepreneurs. If you do that we won't and can't stop you, no one can. But we can choose to vote against you.
As I personally am not dogmatically minded though I engage with people such as we are doing now and I'm genuinely open to the debate around it. The project has changed several times based on these debates, and I have gained a lot from that too. However you're not putting anything forward here which convinces me to change my mind.
I thought about this for months... and I prepared an answer that was VERY long. Probably noone would read it nor understand it.
So for the moment we probably really have to agree to disagree, as I strongly oppose the path you choose. And only the LAST reason I say this would be that I am actually a "victim" of it.
I would be happy to read and I am sure I would understand it. But that is your call.
Yes, I think your goals are counter to ours, that is just the way it is. It's clear it would be the case that a self voting operation would be at odds with a self voting countering operation.
I think on the positive side you try to do something for the community but you touch an area that is not the main problem where you go after and punish smaller players - huge whales (or larger stake holders - with or without delegation) that e.g. upvote their shit with only a 95% click and 4 posts a week while giving generous 1% votes to 5-6 people a week probably will not get into your target zone - just as an example - not knowing the tech specifics - that is it hitting the wrong accounts here. That is my opinion.
I know what you mean, but as I said in another comment we don't have to do everything, this is just one thing and I think the rationale is defensible. There is relatively low SP behind sadkitten accounts and friends currently.
For example @haejin did appear on our list last week in 51st place. When we improve the algorithm to not only take into consideration self voting but clique voting we will see the focus change a bit and I expect the bot will be more interested in even more problematic cases.
The bottom line is that there's nothing controversial going on here, @steembay are massively self voting. Even a little less self voting and the bot doesn't care about them.
@personz I hadn't really looked into @sadkitten till today but I like the concept. It would be great to see this bot growing to tackle some of the bigger fish who are probably getting around the self voting in other ways. I am sure you will get there.
I was curious as to the feedback your getting on the bot and if your tracking how it is changing peoples voting patterns. Has anyone done some analysis on this?
We're currently on v3 of the algorithm, and all versions so far have focused only on the self votes of the user. In this version we were able to pretty much perfect a behavioral model instead of a value based model, so with that working well the next step moving to v4 is to take all account votes into account. Self voting will no longer be the strict focus then, but variance in votees and behavior around that. So if an account only votes for one other account, it will be rightfully recognized as effectively the same as self voting. And that will scale up.
I get a lot of positive feedback in private, and we've always had some opposition from people who disagree with it, such is the kind of venture it is. It's hard to know but several large stakeholders, developers and people I respect and think highly of are backers, which encourages me.
I did attempt to reach out @paulag recently to ask her to do an analysis. I don't have the time right now to do one except my own internal metrics. I was intending to publish something soon and @transisto and I are still discussing the best way to do this.
What is it you'd like to see in particular?
This post is one big data point, where a person feels that the bot is taking away his right to rewards. It would be interesting to see if other users felt the same and wrote a post, or if they adjusted their voting behaviours as a result of the bot downvote.
It would be also good to track repeat offenders and and what level of SP Holding it took for people to take notice of the downvote. The whales surely don't care at the moment as the downvote is inconsequential to them.
If we could get out of the analysis what cohort of people are affected by this bot and what cohort take notice of it that would be very interesting.
I cant wait to V4 as I think that will greatly improve the effectiveness of the bot.
Oh? Clique Voting!?
What percentage of random votes do we need to stay out of the badkitten crosshair?
You have my full support, therefore I am voting @steembay everyday - these guys that drive the sad cat (did I mention I hate cats!) simply go after their % criteria and simply do not take a look behind the bigger picture so what steembay is about, which value it provides to the community for users - all for free.
They simply do not care - there is no need to explain or justify it - you could vote other sock puppet you create and then do not be a target anymore as they watch their self vote percentage - do not spend too much thoughts and heart into defending - not worth a second. I love Steembay!
On a technical note, sadkitten doesn't look at percentage self votes directly, it looks at voting patterns and determines the most optimally self voting accounts per week.
This is one of the moments when you realize that artificial and intelligence do not fit together. This is bulls*** - your self votes are more than 100% justified. The true rape of the reward pool is the bid bots, which turn upvotes into profits for the owner of the SP, which should have gone to content creators. But the system is as it is and I will not be able to change it.
@SadKitten is not an AI anymore, I've edited the profile.
Please clarify what you mean by this for people @transisto
Also ich finde die self votes zur finanzierung von solch einem Projekt komplett in Ordnung.
Everything the Steem allows is legal.
Conversely, this means that self-votes and flag campaigns must also be accepted.Voting in applause is a voluntary act. In contrast, voting as an instrument for generating income appears to be a desperate, strategic act.
To vote for oneself sometimes seems like loving oneself. It could be like narcissism and it is considered socially intolerable. This leads to these insane excesses, such as self-proclaimed policemen who proclaim their view of Steemit as a general rule and punish violations with sanctions.
These pathological downvote excesses have had an unfortunate tradition since the so-called "experiment" of some self-important, arrogant whales as in early 2017. As long as the flag exists, it will probably also be used by these dubious characters. We have to take the farts of these self-important people, like the weather.
But in the case of Steembay, one can consider whether it is actually a sustainable concept to base its financing on its own votes. That seems a little desperate to me. At this point there could be a whole range of much more effective measures, if the operators would just try their imagination a little more virtuously.
Translated from German with DeepL
The thought process behind the development of steembay was far away from desperate. We did not want to take a fee for auctioning, but make it free to use. The selfvoting serves two points: financing it ofc, while we never voted with the steembay stack for our own accounts but more important to influence the overview of the bidding process.
We are online now since Nov. 17 and NOT ONCE anybody complained about us "selfvoting". We had completely different plans (as already explained we wanted to upvote every auction), but the support we got was minimal also the SBD spike harmed our initial plans a lot.
Once more from a game theory perspective: I would invite a million businesses trying to provide services while storing in their account 1000 $ instead of trying to make it "feel" more "fair" for those who never invested a dime.
Selfvoting does not really harm the blockchain, but it slows down a lot healthy growth. "Policing it" or as personz is putting it "making a statement" can be a short term approach. But people will always find a way to act the most opportunistic possible. Thus any way they try to "teach manners" will lead into people finding ways around it followed by even stricter arbitrary lines or more complex algorithms. In the end there is a lot of power and influence centered around a small group able to change rules at a whim. I oppose this approach as I did with gilds and voting trails, as I do with bidbots and vote selling.
Don't get me wrong... I LOVE flags, even if I don't use them often.
Organizing votes OR flags is simply another way of giving away responsibility. We loose and miss a huge opportunity that could have been realised by blockchain technology. But probably that is just me fooling around with my thoughts and utopias of a new society.
This blockchain had an integrated mechanism to reward consensus. ( non linear posting reward algorithm )
Selfvoting has always been 'legal' but was no issue before the equality hardfork.
It absolutely was an issue, but only for whales (or people who would team up and effectively create a 'synthetic whale'). Indeed in those cases, it was an even bigger issue because such large voters were amplified by the effect of n^2.
@smooth, as you are here discussing about the topic, I really would like to know your opinion about a reward curve which starts non-linear but ends linear?
I had written about it here.
Later I saw that @clayop had made a similar approach to reach the same aim.
This is my main idea, however, I think also implementing diminishing returns when upvoting the same accounts (including own ones) again and again and reintroducing the restriction to four (or less) full paid posts per day (from some hard forks ago) could be interesting?
I think that's a pretty good idea. Another way to describe that is a penalty, tax, or some similar terms, on the low end which discourages milking of rewards via high volume spam. I've written in support of essentially the same idea and Steemit even proposed it once. Apparently there are some technical obstacles to implementing at least some versions of it though.
I don't support that because it encourages creating large stables of round robin accounts to evade the repeat voting rule and/or the four post rule. That doesn't help anything and in fact adds more costs onto the system. The four post rule (and maybe the repeat voting rule) also disadvantages many plausible blockchain usages that are different from long-form blogging. Finally, comments are rewarded just as posts are. Applied to comments, four is much too low, so your rule would either encourage reward milkers to use more comments or would discourage comments. Neither is desirable.
See my comments above. As much as we would like people to behave in the desired manner, we can't wish that into existence. All system rules must consider and be evaluated in light of the full range of possible human behavior including evasion or 'misuse'.
Thanks for your input/opinions, I really appreciate it.
I would say the issue was (is) the ridiculous imbalance in distribution of shares and that those shareholders act like they won their stake in the lottery.
You made a lot of sense back in the day, but now you are just picking on my wording.
So were large downvoters. You all had a chance to 'pull the other crabs down'.
I think it was too few people and you guys just ganged up / could not be bothered.
Glad you are still lurking, tho.
Same as now, and now, as before, we see that downvoting as currently envisioned is a mostly failed model because there are is too much to lose (pissing people off, arguments, wasted voting power, retaliation/flag wars) and too little (in fact effectively nothing) to gain.
The crab bucket is a nice analogy but also by now, nicely disproven in practice.
BTW, when you start talking about what people 'act like' as being the real problem, you are always going to be on the wrong track. People are people and will always behave in a variety of ways, not the way you think they should. If a system doesn't have mechanisms to function properly in the presence of the full range of human behaviors, particularly a system explicitly intended for humans to use, then such a system is broken.
I am proposing non linear posting rewards, because it is a mechanism, that rewards consensus.
Such a curation process also creates a valuable peer review mechanism yadda yadda yadda.
I am saying that the chosen handful of people, who got the initial stakes, had no integrity and are not a big enough group to call this a conclusive experiment.
Just because you 10 people did not get your shit together does not mean the idea was wrong.
It is not called 'ned and his 5 crab buddies and 2 bad guys in a bucket' -analogy.
Anyways, linear rewards make no sense at all.
There are still some idiots like me, who vote for posts, but sooner or later, nobody will use the vote function anymore. The system right now does not reward curation, at all.
It would be more efficient to offer a service, that gives you a direct cashback on your vote.
Perhaps I will set that up, just to take the piss.
I'll be an entrepeneur like bernie or ned !
Except that it doesn't actually do that (except in a fantasy world of wishful thinking). It encourages 'consensus' (in a perverse manner of speaking) not between people but between individual units of SP. The most efficient way to achieve that consensus is for one person to own a lot of SP, or for a small number of large SP holders to work together to vote the same posts and then split the rewards.
The number of non-downvoters is a lot larger than that. In fact it is almost anyone (and for the record I'm one of the biggest downvoters in the history of Steem, but still, that doesn't make my point about the lack of downvoting generally wrong.)
Besides, even if what you wrote is correct, it does show that the idea is wrong. Because as, I noted above, for something to work on a blockchain, it has to work for the full range of human behaviors, including a bunch of people (regardless of number) not behaving as you would like.
You're right. It is a good thing we don't have linear rewards, because we have downvoting. If A upvotes and B downvotes, the resulting reward is (assuming equal vote power, etc.) proportionate to SP_A - SP_B, which is not linear in SP. It is precisely a consensus rewarding mechanism as you described above.
Unfortunately, people don't downvote (for a number of reasons including this already stated above), which is the crux of the problem. Fix that and Steem's concept of voting on rewards might have a chance (though blockchain voting has many other problems still). Otherwise, it probably doesn't.
That said, this entire thread (sadkitten vs. steembay, an exception to the general rule of hardly any downvotes) illustrates my point nicely. Self-voting, particularly when practiced in a certain, highly focused, manner, is an example of voting outside of consensus in the sense that there are stakeholders (sadkitten supporters) who don't support it. Therefore steembay's manner of voting its stake is out of consensus and is rewarded proportionately less. Since the reward pool is zero sum relative to voting, that means that everyone else's (in-consensus) votes are rewarded more.
See ? you are picking on my wording. Proportional progression ?
Unfortunately, I did not have my math education in English language. That is why I am not as precise in my statements, as I usually am.
Above, I have told you the reason, why this experiment failed.
All the people who went to steemfest, all the people who went to any steem meetup, all the peope, who I ever talked to, together they own a fraction of the stakes in steem of one of the shadow accounts.
These accounts are not even trying to act like they would, if they had paid for their coins.
A solution would have been to spread votes as wide and small as possible (pharesims approach was ONE possible way of doing it). Still remains the problem of a VERY thin reference group. I think to call n² "failed" is simply not true. Even now with a mio registered accounts it seems there are not much more than 50k real people behind it. A linear voting will never have an effect of "consensus" and flags ... as much I like those... will probably never have the needed effect as they will always be seen and used as punishment => praxeology
I told Ned, when linear rewards started at the same time the whale-not-voting-project ended that there are no reliable and robust results to work with. When he then just smiled at me, telling me with a smirk that "We will know"... I, for the first time questioned the reasoning behind some decisions.
Well... I still have hope.
That's not 'a solution' it is a statement of a goal. What mechanism do you propose to accomplish that, considering the full range of possible human behavior?
Better late than never
Considering the "full range" is not that difficult. Most people will just act opportunistic, possibly inhibited more or less by their morals or an understanding of the mechanics of creating value.
In the early days I wrote a lot about my ideas/thoughts but unfortunately I was so "unimportant" that my considerations have been left unheard or ignored. I think a lot of much better people than myself have left the ship for that same reason.
I am not a fan of statements like "if only everybody had done xyz", so I just will be silent about that.
In the meantime I am not sure that STINC even knows where they are heading to. All I hear is "buzzwords" like SMTs, communities, hivemind, GUI.
I think it was paulag who made an interesting approach for an accurate estimation of "real" users. 50k if I remember well. This opposed to perhaps 100 whales of which at least half have no clue about long term market mechanics is not delivering any reliable numbers.
An image that I see quite often in my head thinking about STEEM is:
Children playing with a (very expensive) petri dish.
The main problem (even before voting) is how to manage mass adoption and account creation fees. Until that is managed, every other discussion is futile and also every thought about "failing reward curves" will have no reliable numbers and thus remain tea cup readings. ( I think and hope this is what the giant stack is for, but as there is near no communication about it, I can totally be wrong).
Well said (even through translation)
It a welcome development. And it selfish act for someone to upvote him self and refused voting others. Thanks for this.
you did not get the point...
Seems like one of the only ones who did 😂
Congratulations @steembay! You have completed the following achievement on Steemit and have been rewarded with new badge(s) :
Click on the badge to view your Board of Honor.
If you no longer want to receive notifications, reply to this comment with the word
STOPTo support your work, I also upvoted your post!
Do not miss the last post from @steemitboard:
SteemitBoard World Cup Contest - The results, the winners and the prizes
Rumor has it that self voting has been removed (again) in fork 0.20. Resistence is futile. 😎
that's not the point here. the point is that this service isn't really rewarded, although so many use it! big buzz goes over the table, and nobody cares about the table itself. steembay may have to change something. never mind. but the worst way of support is certainly to chop up the table!
From a "game theory" perspective this would be useless. There are so many ways around that it would not be worth the effort of even thinking about finding a solution against selfvoting.
From what I read, you'll still be able to self vote, but the setup is changing so that people can't get more than 100% return on it.
Self voting should be removed, though people will just create fake accounts and upvote them
That's why @sadkitten only remove 80% of the reward for a small amount of time.
85% actually, with some few percentage error possible