RE: Update on Simplicity: Cutting Complexity with Steem 0.17.0
Changes to Reward Curves: Not Yet
I am disappointed by this bad news. Here's why.
We considered proposing a move from the n^2 curve (to either modified superlinear n or a linear n), which would mean more influence on rewards for smaller holders of Steem Power than is the case today.
The way you recognize the new curve is inadequate. It's not about increasing small holder's power, but about give fair influence proportional to all holder's stake. Concentration on concentration is a big social problem in Steem IMO.
A modification could be good, but there are game theoretic challenges that haven’t been fully modeled.
Please stop using game scheme, unless you want to create gambling site. Creativity and positiveness come from encouraging and fair social base, not from win-or-lose arena.
Rest assured, we’ve completed the implementation. We just need to see how it functions. If it performs better than the current model, we may include it in the 0.18.0 release.
You cannot know how it exactly works until you implement it. Different rule makes different behaviors and outcomes.
These changes are about rewarding the best content, so we’re handling any voting curve changes with care.
n^2 has no relationship of rewarding best contents. It's all about concentration of power on few whales. And it's obvious that the current system are generating many serious problems now.
What Ned told me is that n will encourage self-vote abuse and hence decreases amount of reward to other authors. But this is not a fatal problem, and can be addressed with community efforts IMO. We are already not rewarding "best contents" and many people are looking forward to a flattened curve. I hope that we have the new linear curve ASAP.
Actually, what seems to be happening now is that whales are doing their damnedest to drive new users from the system. It's disgusting.
New users do not get autovoted by whales so they are not impacted by this.
Actually they are benefiting since more users have influence in the platform so they are more likely to get rewards for their posts.
The people whining about this experiment are a minority of people that gets autovoted by whales and think of steemit as their main job. Most of them are shortsighted and are not interested in growing the value of the platform as a whole.
No this is not true. I know of several instances of new users getting hit with a whale auto downvote. And their posts were less than 1SBD in total.
I know of auto upvote but not auto downvote. Why would whales auto downvote minnows posts with 1SBD? Isn't that SP abuse to target someone like that? I've experienced Whales downvotes but not on a consistant basis.
Because other whale(s) upvoted it. The point is be no whale influence for a period.
Good point.
We still intend to do so. It's coming.
IMO, the "We" should be this community together, not only the devs.
The dev team makes the releases; without working and reliable software, no amount of community decisions result in any changes. The bottleneck isn't the community, but software development.
That said, what I said is still true: We (the community together) still intend to do so. It's coming.
Discussion or active comminication in the comminity can set the goal and direction that we are aiming together. Several dev team members can write the code better, buy thousands of comminity members sometimes provide insightful inputs to be coded.
What if the community don't accept changes you made? The code may not be adopted, or if the devs enforce the implementation community members will quit. Any of them is fatal to Steem; much more serious than some potential abuses by new code.
Theres no discussion or asking opinions to the comminity with HF17, while the new curvr code was being changed three of four times and ended up with reverting code. Is it really efficient?
Please stop only focusing on code and come to the community. They are valuable customer of your product, and the best marketer too. The reason why I am pushing the linerity is that while there are many dissonant opinons on other issues, it is the only thing the community members strongly agree for a change.
I am happy to be able to say that I don't think we're going to have issues like that in the future.
A downvote was applied to partially counter earlier whale votes as an experiment to reduce whale domination of voting influence. Not intended to express an opinion on the content nor result in a net reduction of rewards or reputation (automated notice)
You have the point.
They are referring to Game Theory. It is not about playing games or gambling. It is a study of how as you create rules there are ways to use them exploit(aka game) them etc. This is why they use the term GAME. It is more about studying systems and rules and how people using those rules can do things. It is intentionally trying to think of ways to exploit or take advantage of those rules and if possible try to find a way to eliminate or at least have a way to counteract the exploits.
It's not perfect. Yet when they use the term that is WHAT they are talking about, and why they use it. It has nothing to do with playing games, gambling, etc though all of those things ALSO are subject to game theory.
In a way LIFE itself is subject to game theory.
So when you see that word in the future hopefully that helps you know why it is used. It is a very suitable word and is a very accurate word as long as you know what they are talking about. If you don't then it likely is pretty confusing. https://en.wikipedia.org/wiki/Game_theory
@dwinblood
You are correct about game theory. The thing is that no one really explained how n^2 prevents people from gaming the system. Apparently this curve was meant to discourage self voting so the assumption is that self voting is a problem. Here I explained briefly why self voting is not a problem https://steemit.com/steem/@snowflake/reward-curve-doesn-t-discourage-self-voting
Self voting is like trolling, someone who were to do it repeatedly and excessively would get downvoted, put on the cheetah list and lose reputation in the process.
Also anyone who actually chose to upvote themselves will earn very little curation rewards as they would be wasting their power on post/comments with no community support.
Like I already said, the curve is similar to government banning encryption because a few terrorists used it to commit their crime. You basically penalize everyone because a few bad apples. ( which I havn't even seen yet on steemit)
Self voting should be dealt with downvotes not some curve that maximize the effect of stake disparity and remove the incentives for minnows/dolphins to buy steem power.
I'd be fine with removing self voting, I considered that, then I applied game theory type thinking and realized all I'd have to do is create another account and have them vote on each other. Suddenly I bypassed the restriction on self voting. That is an example of game theory thinking.
Part of the problem is that there are some BIG PROBLEMS that we haven't found a good code solution for. So we do need to experiment, but we need to do it at the code level and set a time period for how long before reviewing the results.
That would remove guessing, and speculating about who is right and who is wrong which just leads to decision paralysis.
I think we need to experiment with some things other than n^2 and while that is not in 0.17.0 there is a good chance it will be in 0.18.0.
Self voting is not a problem, selfish voting is. Votes which are at the expense of community engagement and satisfaction, such as auto votes and circle jerk votes that result in the same users getting to the trending page every day. Those things cause disengagement and do not even benefit the short sighted selfish voter.
As for a code solution, I suggest incentives to seek out new users, by varying the ratio of curation rewards to post rewards according to how many followers a user has and how many vests those followers have.
https://steemit.com/curation/@beanz/the-problem-with-the-current-curation-system
Yes, there are ways this can be gamed, however I don't see the results of this to be bad for the community.
It would reduce the benefits of current system of automated voting if curation rewards for popular authors were lower, and whales could also prevent "pile ons" for curation rewards by using their SP to follow authors that are being voted for out of speculation of the curation payout.
Whales and dolphins would be incentivised to seek out new users for higher curation rewards, giving all new steemians a welcome and proper chance of achieving recognition.
Doesn't if feel good to take part of this "Experiment/Beta"?
I only wish I could of found it sooner and could of possibly contributed more input with no upfront cost!
In the meantime, this will have to do.
Let's ALL continue to STEEM on...
Frank
Man, this experiment with the whole Steemit Platform is exciting and addicting. I am very Honored to have this opportunity to witness, take part, vote, comment, & continually learn and grow with the company. A lot of GREAT, POWERFUL, and WISE users on this platform, if you are keen to spot them...(notice I didn't have to actually say anything about my pay or a salary)
Just my 3cents worth again...
Yep. :) My discussions actually have little to do with price. I don't truly care too much about that myself. I do care about perceptions, and how actions can impact the community. We can get a bit too focused on the mathematics of it all at time and forget that humans are not all Vulcans based purely on logic and mathematics, so only paying attention to the stats is a recipe for disaster that any large group of humans can explode into pieces due to all the differing opinions, perceptions, etc.
Yes why not post the actual reward, the reward under the new formula and maybe one more. Payouts are based on the first but people could see what they would get under other foRmulas for reference
(nesting limit)
The supply of those store is already miss products whom demands are already nullified by 'big shareholders'.
For e.g., I cannot go to a store and buy heroin, though, I'm pretty sure there is a demand for it. However, I would like to mark it as bad if I were walking next to it in a store. That isn't nullifying the demand. Who want to buy it, will buy it.
@baah (nesting limit)
Legal / Illegal <== Marked
It's a real world example for marking products. The difference here is you are the gov (part of it) and you mark it directly.
You would mark heroin bad because you perceive it to offer no redeeming qualities, but if it was legal, in a store, to buy, marking it as bad would be ridiculous because you have basically damaged the product and therefore create more demand for it and second because nobody does that in the store, in real life, nobody goes to the gun store and marks guns with x because they kill people, nobody goes to the kitchen section and marks knives with a black marker, x, or any other product, people are sensible like dwinblood has pointed out and would not do something like that. Take drain cleaner, sure you can use it for it's intended purpose but if you were bombarded with numerous accounts of people using it to get high (hypothetically) and then dying because of that, then you will also go and mark drain cleaner as bad, therefore damaging the product, creating more demand for it etc? It's pure nonsense therefor, regardless of the fact that heroin does have redeeming qualities and has been used in the past in treating numerous things, as everything has value in various aspects, where as marking x on products you don't want doesn't have value.
It says:
As much as I like playing with Game Theory, I'm unconvinced game theory will make a correct prediction for choosing a reward curve. It's a candidate for an experiment, and the sooner the better.
Such as? ive never seen any details (or any community discussion) on these 'big problems'. just n^2 staying because INC decided to put it back in.
Well a BIG PROBLEM example... and "problem" could be subjective. If you prefer use the word CHALLENGE.
I was a proponent for UP VOTES only, but that leaves this place open for way more exploitation. So that is an example of a BIG Challenge. How could you do an up votes only (except for spam, plagiarism, abuse) without opening the door for it to be completely gamed?
I stopped calling for an up votes only system simply as I haven't come up with a solution to that Problem/Challenge. Dan said he was interested in an up votes only system, but again he does not do it for the same reason I stopped calling for it.
I would like to see the platform more like an economy than a share holders meeting. If I go into a store I am not pulling out a marker and putting big Xs on products I don't like. I walk past them and buy the things I am interested in. Other people walk in and might buy things I am not interested in.
Being able to put Xs on things you are NOT interested in is nothing like a market or economy. It actually allows you to nullify the interest of others as though it did not matter. It gives a false impression of demand due to the voting power being tied to a share like approach.
The demand can be there, but nullified by one person.
Yet solving that... Big challenge/Problem.
In a market demand = WHO IS GOING TO BUY THIS
It has nothing to do with WHO IS NOT GOING TO BUY THIS
that is more something a marketing department would focus on to get more of the NOTs into the WILL column.
Yet ultimately the NOTs do not get to nullify the demands of the WILLs.
If I have 10 people that want to buy something that is a demand of 10.
If there are 100 people that will not buy the thing that does not make the demand -90. There are still 10 people that will buy it.
Right now with the way steem/it works we get that -90 effect.
nesting
youre misunderstanding my point. The contention is that n^2 is staying because of a "big problem" in the GT of removing it. But no one has detailed precisely what that "big problem" is. In this case (as you mentioned elsewhere) the "game theory" thing is just an appeal to authority.
Yeah I wasn't talking about n^2 as the big problem. I only brought up Game Theory because of the response that they need to stop talking about Gaming.
The rest of what was said I had no problems with. The big problems came out of the responses discussion to my reply, and had nothing to do with n^2 or the original article.
All I was talking about in my initial reply was in the statement that they need to stop talking about Gaming unless they are going to become a gambling site or something of that nature.
Thus, my response was to explain game theory and WHY the use of gaming was relevant to the discussions. That was the only purpose of my original response. I don't believe I've referenced n^2 once here unless it was responding when someone else mentioned it in a reply to me.
The big problems I am referring to are the ones that are unfamiliar and difficult concepts that we as humans have not had to solve before. I don't consider n^2 a big problem. That is an easy problem. Implement a different curve, test it. If it doesn't work try something else.
I wouldn't call that a big problem. :) when I call something a big problem I am talking about those really big mind bending problems that we don't know how to solve yet, we suspect there is a solution out there, but it takes a lot of effort, and trying new things to find it. IF we ever do.
I knew they are using the term game theory to make people cooperate. It requires people to expect other peoples bevaviors and decide based on the estimation.
But is it really good for contents system? I don't think so. The game theory enforces people to choose posts strategically and somewhat discouraging them to freely choose what they really like.
This is not why they are using it. Read the wiki link I supplied you.
Game Theory is not used to shut down conversation. It is to put into context of understanding that WANTING something means you need to think of possible ways that can be EXPLOITED if it is given to you.
Everything has rules. How do these rules balance each other out? How can they cancel...
I friggin' hate the flag as it exists... for a long time I was an advocate of an up votes only system and I wrote asking for such for a good 5 months.
It was a game theory explanation that made me stop advocating for removing the down vote.
That doesn't mean I don't still want an up votes only system. It simply means I haven't thought of a way to stop an UP VOTES only system from being extremely exploitable. Until I can solve that I cannot advocate for removal of the down vote as it is currently the only thing that CAN negate such systems where people game/exploit the system.
This is not making you cooperate. This is life. Actions have consequences. When building things from code we do them, but we also need to think about not just technical hackers, but social hackers, who will exploit weaknesses.
In an up votes only system someone could create infinite accounts up voting their own single account and over time drain the pool and be very powerful without ANY interaction from other users. Down Votes from other users can stop this. This is but one example. Yet it shows how it is exploitable.
Personally I'd be willing to experiment with up votes only and having something like being able to flag stuff as spam, plagiarism, abuse and if witnesses agree they can do the equivalent of a flag or some system like that, but I don't know how feasible it will be.
Yet using the term Game Theory has absolutely nothing to do with wanting you to cooperate. It is two words that define the situation. They don't solve it. They put it into context. Every action has consequences. Wanting something doesn't mean there may not be bad consequences so game theory is about trying to determine the positives and negatives and ideally it will be a balanced system where every positive and negative has a counter balance.
I do think the flag as it exists in steemit/busy now has more negative impact than it does positive, so I do not think it is equally weighted. Numerically it is equivalent. Psychologically and system impact it does not seem to be equivalent. This is a problem, and Game Theory could actually indicate that if it were pursued with other factors such as PR, Social, Psychological, Communal perception impact. Just on raw money, votes, and reputation though it is equivalent.
Not in this case. Game theory is mathematical. Saying "we decided to keep n^2 because game theory" is exactly a way to shut down conversation, and its not real game theory at all. its just avarice masquarading as it.
Sure... people do that a lot. Yet that is not Game Theory shutting it down... that was actually an appeal to authority fallacy being employed. Much like saying "that is heresy". It didn't actually make Game Theory the source of the problem. It is people accepting appeals to authority and giving in that is the issue.
Be advised the context. While you are talking about flagging, the context of OP and mine is about
n^2that intends cooperation on purpose.Actually that is only where it ended up. I was only responding to the initial quote of yours I quoted where you asked them to stop referring to game unless it was going to turn into a gambling site.
That is what I was responding to as Game Theory has very little to do with that and does play very much into a lot of the discussions here.
As to the rest of what you had to say, I didn't necessarily disagree with any of that. I only thought it important to clarify that use of the term GAME is much larger than I think you were considering and it is appropriate. Game Theory is a rather back alley study that has been gaining in favor and applies to human interaction, animals, logic, and computers, and since this system deals with all of those and runs on a computer it is relevant.
As far as n^2 and such though it won't really matter one bit. n, n^2, nLog2 in Game Theory doesn't mean shit as long as they are consistently applied.
I know you've written that in the context of author rewards, but this is why I'm against curation rewards.
basically I support a flatter reward curve. however, an idea occurs to me: if some want linear (n^1) and some want or want to keep longer n^2, is it possible to fork it like n^a and 1<a<2? a can be decided like feed prices determined by witnesses.
I know this doesn't make things simpler. Just a thought and welcome any criticism.
1.x is technically not desirable as far as I am heard.
It could be possible as I understand it as long as it's on the same blockchain, but that would mean new accounts and how would that translate to transferring steem from one version to another as they run in tandem and one can move their steem to the more profitable flatter reward curve if they are a minnow and the whales will see no value in doing that effectively fracturing the community.
indeed, that's a problem. didn't think of that. thanks for the feedback.
"Hmm... that's a grounded view of things, for you see, without the ground one could not tell where the sky begins." - Justin Harvey John Ashby
The curve is the biggest reason steem has failed to capitalize on the July august sign ups..