Steemit how can this thing be workable long term? - page 7.

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: magicalacademy on July 22, 2016, 03:16:19 PM

iamnotback, before investing your time in this, have you considered the possibility that the devs's ultimate goal is to enrich themselves instead of making steem a success?

To me everything points to the former, closing registration + extending payout time in a last ditch effort to retain value of the hyper-inflated currency...

Of course. Yet there may be something worthy here... but might need to be forked ... not sure yet...

magicalacademy

member

Activity: 66

Merit: 10

iamnotback, before investing your time in this, have you considered the possibility that the devs's ultimate goal is to enrich themselves instead of making steem a success?

To me everything points to the former, closing registration + extending payout time in a last ditch effort to retain value of the hyper-inflated currency + no change to the flawed voting algo cuz it will lower their power

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: smooth on July 22, 2016, 03:08:57 PM

I'm not convinced much changes if you divide into N like-minded segments. More relevant ranking results, yes. But still stake-weighting means that the few will have the bulk of the influence.

My prior post Edit means I choose who I want to have influence on me. I can choose which stake I want to follow. Unbounded degrees-of-freedom. TADA! And solves the potential Grassmannian issue also.

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: smooth on July 22, 2016, 03:00:20 PM

If the issue is the eventual distribution then the current whales and their stakes don't really matter. Either stake-weighed systems work, or they don't.

...

Quote from: iamnotback on July 22, 2016, 02:58:36 PM

We need a smarter algorithm which can measure like-mindedness so that we can relevant ranking results, not stake game theory wars against the power-law distribution.

Edit: new idea. We don't need to compute the k-clusters. Each user can form their own cluster for their personalized ranking computation based on who they follow. We can introduce some randomization so they can find new people to follow (which somewhat comes from recommendations of whom they already follow).

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: ?? on ??

Quote from: smooth on July 22, 2016, 03:00:20 PM

If the issue is the eventual distribution then the current whales and their stakes don't really matter. Either stake-weighed systems work, or they don't.

...

Quote from: iamnotback on July 22, 2016, 02:58:36 PM

We need a smarter algorithm which can measure like-mindedness so that we can relevant ranking results, not stake game theory wars against the power-law distribution.

I'm not convinced much changes if you divide into N like-minded segments. More relevant ranking results, yes. But still stake-weighting means that the few will have the bulk of the influence.

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: iamnotback on July 22, 2016, 02:58:36 PM

Quote from: smooth on July 22, 2016, 02:56:40 PM

Quote from: iamnotback on July 22, 2016, 02:55:20 PM

Quote from: smooth on July 22, 2016, 02:50:20 PM

Quote from: iamnotback on July 22, 2016, 02:46:27 PM

So with a post with less than say 1000+ votes, you should be able to double or more its ranking.

Ah, but look, already with the relatively tiny user base the top posts have 500-700 votes. Multiply the user base by 100 and they will have not 500 votes by 50000 votes.

Will number of users voting increase faster than number of blog posts? I think it should be the reverse, if people post more than one blog simultaneously (which I assume they eventually will do).

The number of people voting will probably increase much more than the number of blog posts in the short term, because the site is currently dominated by active bloggers. Longer term it will probably converge where they increase at a comparable rate.

Increasing the number of users dilutes their share of the debasement. It is a zero-sum. Bottom line is the whales must be diluted, but the problem is the power-law distribution will dominate because you can't 100% annual debasement forever.

If the issue is the eventual distribution then the current whales and their stakes don't really matter. Either stake-weighed systems work, or they don't (and maybe they don't, but we don't know that).

Honestly I really think crypto nerds are highly obsessed with the whole issue of flattening the rewards and influence and whales and such. Users sign up, post, have fun, vote, get a few bucks, stare in awe at what the stars are getting, and they're basically going to be happy with it. So far that bears out.

It isn't really going to be a site to make a steady living for non-stars. Just like how Instagram has a few stars who essentially make their living on it, even if not paid directly, but most users don't. It provides infotainment (like other social web sites) and some rewards to keep it a bit more interesting (unlike other social web sites). And even the small rewards distribute at least starter amounts cryptocurrency which people can use. Potentially orders of magnitude more than have ever seen it before. Should be interesting.

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: smooth on July 22, 2016, 02:56:40 PM

Quote from: iamnotback on July 22, 2016, 02:55:20 PM

Quote from: smooth on July 22, 2016, 02:50:20 PM

Quote from: iamnotback on July 22, 2016, 02:46:27 PM

So with a post with less than say 1000+ votes, you should be able to double or more its ranking.

Ah, but look, already with the relatively tiny user base the top posts have 500-700 votes. Multiply the user base by 100 and they will have not 500 votes by 50000 votes.

Will number of users voting increase faster than number of blog posts? I think it should be the reverse, if people post more than one blog simultaneously (which I assume they eventually will do).

The number of people voting will probably increase much more than the number of blog posts in the short term, because the site is currently dominated by active bloggers. Longer term it will probably converge where they increase at a comparable rate.

Increasing the number of users dilutes their share of the debasement. It is a zero-sum. Bottom line is the whales must be diluted, but the problem is the power-law distribution will dominate because you can't 100% annual debasement forever.

We need a smarter algorithm which can measure like-mindedness so that we can relevant ranking results, not stake game theory wars against the power-law distribution.

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: iamnotback on July 22, 2016, 02:55:20 PM

Quote from: smooth on July 22, 2016, 02:50:20 PM

Quote from: iamnotback on July 22, 2016, 02:46:27 PM

So with a post with less than say 1000+ votes, you should be able to double or more its ranking.

Ah, but look, already with the relatively tiny user base the top posts have 500-700 votes. Multiply the user base by 100 and they will have not 500 votes by 50000 votes.

Will number of users voting increase faster than number of blog posts? I think it should be the reverse, if people post more than one blog simultaneously (which I assume they eventually will do).

The number of people voting will probably increase much more than the number of blog posts in the short term, because the site is currently dominated by active bloggers. Longer term it will probably converge where they increase at a comparable rate.

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: smooth on July 22, 2016, 02:50:20 PM

Quote from: iamnotback on July 22, 2016, 02:46:27 PM

So with a post with less than say 1000+ votes, you should be able to double or more its ranking.

Ah, but look, already with the relatively tiny user base the top posts have 500-700 votes. Multiply the user base by 100 and they will have not 500 votes by 50000 votes.

Will number of users voting increase faster than number of blog posts? I think it should be the reverse, if people post more than one blog simultaneously (which I assume they eventually will do).

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: iamnotback on July 22, 2016, 02:46:27 PM

Quote from: smooth on July 22, 2016, 02:38:12 PM

When there are 100x as many users voting on posts, the influence of whales will decline by at least 10x.

Smooth what happened to your math skills Huh

If you have a million SP, then it requires 100,000 free signup users (10 SP each) to match your voting power.

Of course, but active users don't stay entry level for long. Look at the above charts. There are 1700 people with 1 MV (= 230 SP) and 500 with 10x that.

Quote

So with a post with less than say 1000+ votes, you should be able to double or more its ranking.

Ah, but look, already with the relatively tiny user base the top posts have 500-700 votes. Multiply the user base by 100 and they will have not 500 votes by 50000 votes.

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: smooth on July 22, 2016, 02:38:12 PM

When there are 100x as many users voting on posts, the influence of whales will decline by at least 10x.

Smooth what happened to your math skills Huh

If you have a million SP, then it requires 100,000 free signup users (10 SP each) to match your voting power.

If you argue there are 10,000 users with 1000 SP each, then it requires 10% (1000) of them (voting on the same post) to match your voting power.

So with a post with less than say 1000+ votes, you should be able to double or more its ranking. Actually quadruple or more, because of the total being squared (quadratic not linear).

(I am not saying you personally are benevolent of course)

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: iamnotback on July 22, 2016, 01:33:28 PM

Dump the pegged asset STEEM DOLLAR. It is entirely unneeded for where we are headed...

If we don't get users to use STEEM POWER as their unit-of-account, then the entire concept is DOA.

Quote from: https://steemit.com/steemit/@anonymint/it-s-so-easy-to-become-a-millionaire-with-steem#@repholder/re-anonymint-it-s-so-easy-to-become-a-millionaire-with-steem-20160722t184534553z

The hard part is getting out Wink

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: CoinHoarder on July 22, 2016, 02:36:22 PM

Quote from: smooth on July 22, 2016, 02:15:59 PM

If I vote on something new, it gets maybe to the 5th or 6th page of trending. Multiply the volume on the site by 10x and that becomes the 50th page. The really highly ranked posts will have many votes from non-whales, just a they do now, but by an even greater degree.

Volume doesn't matter, as the content is organized by the SP power behind the up votes. 10x the volume does not mean your up vote will go 10x pages down, as you have way more SP than most.

Volume does matter, because there are only a small number of whales and they can only vote a limited number of times. When there are more users, there votes will count proportionally more. There is no quadratic weighting by user, only by post. When there are 100x as many users voting on posts, the influence of whales will decline by at least 10x.

Mark my words, as the site grows "whale sightings" will become rare. Watch it happen. I already notice the change taking place.

Quote

after omitting the Steemit account considering it doesn't vote.

Remember the steemit account is explicitly being distributed to entry level users and to buyers (who mostly won't be whales).

CoinHoarder

legendary

Activity: 1484

Merit: 1026

In Cryptocoins I Trust

Quote from: smooth on July 22, 2016, 02:15:59 PM

If I vote on something new, it gets maybe to the 5th or 6th page of trending. Multiply the volume on the site by 10x and that becomes the 50th page. The really highly ranked posts will have many votes from non-whales, just a they do now, but by an even greater degree.

Volume doesn't matter, as the content is organized by the SP power behind the up votes. 10x the volume does not mean your up vote will go 10x pages down, as you have way more SP than most.

Also, you are ignoring bandwagon voting. A post which has a whale vote is much more likely to get dolphin/minnow votes because it is much more likely to make the front page and make a decent curation reward. The dolphin, fish and minnow votes are mainly a windfall result of the whale vote.

As long as content is displayed by whatever content has the most SP behind its up votes, you can't seriously look at this chart and say that whales will become irrelevant...

Doing the math myself... the top 24 accounts own 47% of the SP in the voting pool (after omitting the Steemit account considering it doesn't vote.)

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: iamnotback on July 22, 2016, 02:26:54 PM

Quote from: smooth on July 22, 2016, 02:15:59 PM

When there are more users whales won't be able to get anyone near the front page.

Fact is if 80% of the stake votes together, they can influence which content goes on the front page.

Sure. That's the nature of a stake weighted system. 80% is a huge supermajority, it can clearly vote anything it wants. The reduced number of winning posts change this to a degree, but ultimately 20% is not going to be able to outvote 80%. Period.

If stake weighting is unworkable then we'll have to just accept that any system on the Internet must be either KYC-based or overrun with sock puppets. The answer to that experiment is not clear yet.

(replies to the rest of your comment as well)

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: smooth on July 22, 2016, 02:15:59 PM

When there are more users whales won't be able to get anyone near the front page.

Fact is if 80% of the stake votes together, they can influence which content goes on the front page.

Quote from: iamnotback on July 22, 2016, 01:44:23 PM

You presume all whales are benevolent. Stake will eventually end up in a power-law distribution, because perpetual 100% debasement is not sustainable (even the transactions fees of the entire earth could not support it).

Quote from: smooth on July 22, 2016, 02:15:59 PM

but the more you divide things up into more pieces, the more likely it is that the number of whales present in any given segment will be zero.

Hypothetically the solution is to divvy up by like-mindedness, so that no group can collude to control all the rankings. With k-means, it is impossible (unless you can somehow game the algorithm by splitting up your accounts) to have your voting power in more than one of the k clusters simultaneously.

I do worry though that the power-law distribution will always find a way to win.

Even if we allow users to control their own clusters (e.g. via a Follow button), the power-law distribution may influence which users they are exposed to.

I'd suggest adding some randomization.

BlindMayorBitcorn

legendary

Activity: 1260

Merit: 1116

^Great job. https://np.reddit.com/r/Bitcoin/comments/4u3vov/steem_turning_into_an_inadvertent_ponzi_funded_by/

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: CoinHoarder on July 22, 2016, 02:06:04 PM

Quote from: smooth on July 22, 2016, 01:38:23 PM

Whales will become totally irrelevant when the user base and posting volume increases by 10x or 100x or even 1000x. There just aren't that many whales and whales have the same voting power restrictions as anyone else. A rare whale sighting will be a big deal.

Whales will still have a huge amount of sway as far as deciding what content gets put on the front page.

When there are 4 to 10 posts a minute, the "new" tab is useless and it is impossible to read all or even a majority of new posts.

To show up on all the other tabs, you need to have a substantial amount of SP up voting your post.

The dolphins, fish and minnows aren't able to see a post to be able to up vote it unless it already has some serious SP behind it. The most likely source is votes from whale(s). Enough dolphins, fish and minnows are not going to be able to find a post to give it enough up votes to outpace that of one or a few whale up votes.

When there are more users whales won't be able to get anyone near the front page. If I vote on something new, it gets maybe to the 5th or 6th page of trending. Multiply the volume on the site by 10x and that becomes the 50th page. The really highly ranked posts will have many votes from non-whales, just a they do now, but by an even greater degree.

This will be somewhat different if people follow tags where the competition isn't as great, but the more you divide things up into more pieces, the more likely it is that the number of whales present in any given segment will be zero.

Also, as I understand it the front page is to be phased out in favor of seeing posts from those you follow. How you will find new people to follow I don't know, maybe via suggestions like Facebook or something?

CoinHoarder

legendary

Activity: 1484

Merit: 1026

In Cryptocoins I Trust

Quote from: smooth on July 22, 2016, 01:38:23 PM

Whales will become totally irrelevant when the user base and posting volume increases by 10x or 100x or even 1000x. There just aren't that many whales and whales have the same voting power restrictions as anyone else. A rare whale sighting will be a big deal.

Whales will still have a huge amount of sway as far as deciding what content gets put on the front page.

When there are 4 to 10 posts a minute, the "new" tab is useless and it is impossible to read all or even a majority of new posts.

To show up on all the other tabs, you need to have a substantial amount of SP up voting your post.

The dolphins, fish and minnows aren't able to see a post to be able to up vote it unless it already has some serious SP behind it. The most likely source is votes from whale(s). Enough dolphins, fish and minnows are not going to be able to find a post to give it enough up votes to outpace that of one or a few whale up votes.

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: smooth on July 22, 2016, 01:52:05 PM

Quote from: iamnotback on July 22, 2016, 01:44:23 PM

Quote from: smooth on July 22, 2016, 01:38:23 PM

Whales will become totally irrelevant when the user base and posting volume increases by 10x or 100x or even 1000x. There just aren't that many whales and whales have the same voting power restrictions as anyone else. A rare whale sighting will be a big deal.

You presume all whales are benevolent. Stake will eventually end up in a power-law distribution, because perpetual 100% debasement is not sustainable (even the transactions fees of the entire earth could not support it).

Also it is incorrect to say they play by the same rules. 3 x 3 = 9 is not 1/2 of 6 x 6 = 36.

Any way, I have proposed a partial solution.

By voting power I was referring to the declining weight according to frequency of votes placed. Just like anyone else, whales can only make about 40 votes per day without depleting their voting power. So with millions of posts and comments and a handful of whales making at most 40 votes per day each, do the math.

But those 40 (times many whales) are enough to influence (with quadratic weighting) the Top 100+ ranked posts on the site which earn perhaps 90% of the rewards.

I have an idea to add the recent blogging rewards as a significant component of voting power. Those who are generating good content should have more weight even if their stake hasn't accumulated yet. It would be more dynamic. We can't assume whales are benevolent any way, so we need to have game theory that won't allow whales to increase their percentage of stake via voting.

Topic: Steemit how can this thing be workable long term? - page 7. (Read 32368 times)