A nice example of one of Steemit's problems (edited)

in steemit •  8 years ago  (edited)

This post, containing the wise words "fwe fe", has already made more than many a good post with real content in it. (Edit: some rewards have been flagged away now.)

It's not about being annoyed because somebody makes more than you, it's about how ridiculous it makes people feel when noise makes more than content, even when that content gets over 100 votes.

Bots gone wild? Following votes? Sock puppet? Who knows.

(Edit, thanks @nikolai:) The problem goes deeper than votes-for-noise; posts with 200+ votes making under $1,- are hard to explain to new users when they see noise or even equally good posts make a lot more, with under 20 votes. The gap is too wide and doesn't seem to be closing at all.

Would it be possible to change the reward formula into something that also includes both the number of upvotes and the reputation of the upvoters? Or would that just invite even more botvoting?

I was thinking about rewarding, for each vote, (voter reputation / a) + (current formula reward / b), with a and b calibrated to give a more even result without upsetting the total reward pool too much. Or something along those lines 8-).

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:  
  ·  8 years ago (edited)

There's no easy solution, unfortunately.

The reputation system is really only useful for negative reputation, to hide spammers. High reputation does not mean much other than the user has been previously rewarded.

Number of votes means absolutely nothing. I can upvote this comment hundreds of times if I want to, just with a small fraction of my accounts: this is a sybil attack and steem is designed not not reward sybil attacks, but rather prevent it by using stake weighted voting.

For example: My comment here has more upvotes than your post right now, does that mean I deserve a larger reward?

The post you call into question I brought to attention to some people. It's an unfortunate aspect of current vote bots that are not the most intelligent, yet.

For example: My comment here has more upvotes than your post right now, does that mean I deserve a larger reward?

I got your point before 8-). It wouldn't get a larger reward if the reputations of all those upvoters would be taken into consideration, though. Quoting myself:

I was thinking about rewarding, for each vote, (voter reputation / a) + (current formula reward / b), with a and b calibrated to give a more even result without upsetting the total reward pool too much.

How would that fail? Just asking, am learning.

It's very easy to create multiple accounts with high reputation.

hi im sorry for bothering you but i did not mean to make a mistake on my artical i deleted my bad links can you please explain to me why i was downvoted

fuck you and fuck your cheetah stupid bot too..

Could you say that using any other metric than "wallet size" for determining what a vote is worth would open things up even further for gaming the system?

Could you say that using any other available metric than "wallet size" for determining what a vote is worth would open things up even further for gaming the system?

I don't necessarily think that steem power as a linear value alone has to be the only way to calculate vote worth, but I do think it has to be at the core.

For example, another system that MIGHT work is using a pagerank type setup where everyone declares trustworthy users, and essentially a web of trust forms. Then, the power of a users vote also is rated based on the value of people who trust them. This is still open to sybil attacks, however people can choose not to trust (or even distrust) people who are clearly only trusting their sockpuppets.

Sometimes I wonder, only sometimes, mind, what would happen if voters wouldn't get curation rewards at all, and the only way of making something would be posting content and get votes. I only think about that because I would like to see an end to all the gaming and bot voting for money and sock puppets and other shit going on troubling the waters. I must admit I haven't had any coffee yet.

These are the questions that are worth asking, but you have to think like an adversary in order to determine the viability. :)

One important aspect of curation rewards is that it is economically incentivized to vote for a post that will be voted on by other people. The reward for doing this is greater than the reward for you creating a post of your own and using that vote on your own post instead. For example, say, I could upvote my own comment and earn 5 cents... or I could use that vote to curate and earn maybe 10 cents.

If there wasn't curation rewards, one would be economically incentivized to only use their voting power on their own posts. Why should I upvote someone and earn 0, when I can still upvote my own comment and earn that 5 cents?

We're stuck between a rock and a hard place.

Unless .. you have a population that votes because they actually like something in stead of one that votes for making curation rewards. There are other motivations for voting than just the economic ones. No idea how it would play out in practice, though, so feel free to call me naive 8-). I know I vote for things I like without considering curation rewards at all, and not just because I hardly get any anyway with 5000 SP.

Consider this problem as noise to be ignored. Although it IS frustrating, the way for Steemit to succeed is to keep generating GOOD content...and hope a wave of robots upvotes you until the real curators show up

  ·  8 years ago (edited)

You're probably right, and I don't really mind much for myself, I'm fine here, but I'm worried about what this does to new user retention.

I agree that it probably does drive away some new users, but those people would also likely leave when that dont get paid big bucks on their first post.

the whale/robot folks can be on Steemit for the short/medium term. creators have to look at this in long term mode

In the long term, we'll all be dead 8-). But I get your point and I hope for the best. I don't see things shifting the right way yet, though, however slowly.

Wow, why would anyone upvote a post with only that as its content? I am betting that it is voting bots.

Unless we see the frequency of posts like this growing I'd assume it's bots misbehaving and not an exploit of the rewards mechanism. In that case I would actually be happy about it because it's just normal market feedback to cut back on inefficient bots and reward bots that add value, even if someone accidentally made some money.

  ·  8 years ago (edited)

You may be right. The problem goes deeper than just empty posts like this, though; it's also about the distribution. Posts with 200+ votes making under $1,- are hard to explain to new users when they see noise or even equally good posts make a lot more, with under 20 votes. The gap is too wide and doesn't seem to be closing at all.

The gap is too wide and doesn't seem to be closing at all.

Do you know of any numbers to support this? Paper napkin theory math says whales need to gather upvotes worth 10% of their SP to maintain their voting power, which seems like it would be a force that dissipates voting rewards

  ·  8 years ago (edited)

I have compared the tables at http://steemwhales.com/ over time for a while to get an idea of the development of voting power as distributed in the population, and concluded that you still need at least one whale vote to equal about 150 to 200 average votes (excl. whales). The numbers are off the top of my head, as I buggered up the spreadsheet a week ago and couldn't be bothered to repeat the exercise.

I'm not concerned about current distribution, only about centralization vs diffusion effects that the voting system has. Has the power of the top 10 or 100 voters been trending down or up? How about average value at 20,000th voter?

Resteem this...