Bid bots aren't the devil and neither are Linear Rewards

Ok, people. I've seen some folks around these here parts saying bad things about bid bots and also some posts recently dumping on linear rewards. I've had mixed feelings about HF19. You may recall I launched a few posts into trending mentioning "clusterfuck" and "circle-jerk" a week prior to launch (I was admittedly a little late to that party).

Linear Rewards

Here's my take on why distribution is still a little fucked. A eon ago in cryptoland or approximately 2 years for normies Dan and Ned launched Steemit.com and the Steem blockchain. When they did this Dan looked at legislation surrounding ICOs and said "You know... it actually seems like if we do this the wrong way we'll create and unregistered and unliscend security. We can either register it and do this that way or we can get around those requirements by working within the law a different way and simply starting from scratch and mining from zero point while not promising any equity returns so it doesn't look like a security." Faced with that decision they opted out of registering Steem and choose to mine from the start.

Now, they were excited to own and manage a big blockchain company so they didn't make it super easy for everyone to mine. They put it out there how to do and then didn't offer extensive help. That's because they didn't' want a lot of mining competition. It was fair. Anyone could mine it, but it was also unfair because it wasn't obvious how to setup the miner. Dan helped some folks out that were devs or pretty close, but otherwise you either knew how or you didn't.

Getting started

Immediately people started complaining both because of a restart with a glitch and later because Steem was spending more money mining than everyone else. As a result they accumulated 75|% of the steem at one point. Most of the rest went to a handful of people who enough dev knowledge to setup a miner, and later they shut off mining as an option.

Then they followed through on a shitty plan where they did stake weighted inflation. That means the people that already had steem received the new steem. It's like staking coins, but Steem already had a shit distribution and this made it worse because the haves became the mega wahles and the have-nots became plankton. It also tanked the price of Steem. The value of your holdings didn't change much, but post rewards were basically nothing.

Back in my day Steem was seven cents each and it took me 5 posts to make 0.07 Steem from them!

So, that created a really shitty distribution on the platform, but the challenge didn't end there. We also had this thing called exponential rewards. Basically when figuring out post rewards you took your stake and squared it. This took the broken distribution on the chain and made it "exponentially" worse.

Example

Minnow with 5 SP votes full power- rewards shares ~ 25
Dolphin with 5,000 SP votes full power - rewards share ~25,000,000
Whale with 50,000 SP votes full power- rewards share ~ 2,500,000,000

Get the picture? That minnow ain't shit.

Why have exponential rewards? Mostly because of abuse. The point is to find consensus on what a good post is. Generally, the assumption under exponential rewards is that good content will be liked by lots of people so a post with very few people supporting it ain't worth much even with a large vote at the start, but posts with lots of likes end up being worth a ton.

So, in this version you have a few people that are making a killing on Steem because the whales like them and also because of how curation could be gamed. You needed to find an author that was likely to be voted high and trending, get on early, and try to ride exponential curation rewards up. It made it so autovoters became prominant as people gamed the system to upvote the trenders early to catch curation. This made it so that fewer and fewer people got more and more rewards as everyone piled in.

So, that was sorta broken

Ok, to be clear, I think exponential rewards is best, but it's horrible with our current distribution. Something like 93% of the Steem is in the top 100 accounts. You square the influence of the top holders and it gets worse. There were experiments done to try to combat this like limiting how much whales could vote or else they'd get flagged, but ultimately that wasn't cool either as it generally discouraged holding a lot of Steem.

Backdoor hack

I think the impotus for linear rewards was really trying to work around the distribution that was created. By switching to linear it did actually give the minnow accounts a lot more say on the platform. It's still miniscule compared to a whale, but it's roughly an order of magnitude more than what it was before. I remember when my vote became worth 2 cents. It was a giant deal and it took me months on the platform. You ungrateful bastards don't know how good you have (read that in a grumpy old man voice please).

So, they switched to linear rewards and it fixed everything... or not quite

I tracked how rewards started changing right after the change. The most shocking piece was that way more votes were going into comment rewards. Now, I'd love to say this had to do with people getting genuinely interested in other people's posts, but the honest truth is that people started self voting in comments. You couldn't do that before. It wasn't worth shit... even for a mega whale. That's not a challenge now, so you can get massive votes in comments and the value of the comments collectively roughly tripled.

What's the point. We have a lot more self voting. it's now profitable. Generally speaking this is highly frowned upon behavior. I call it autovoterotic. You're getting yourself off with your votes. Am I talking about all votes? No. I tend to vote for my own posts frequently. I wrote something I think it's worth giving myself 1-2 dollars for it. But not my "hey this art is lovely." Especially when we first got going with the change I was watching people dump $100 upvotes in their own comments frequently.

What about Bid bots?

Bid bots weren't prevalent during exponential voting. They are on golos and apparently had a financial impact there, but exponential rewards woudln't make it profitable unless your post was already doing well.

What about vote selling? It was fucking rampant! Vote swapping and vote selling was happening all the time, but it just wasn't a transparent activity that anyone could engage in. You had to be wealthy and connected to even get a shot at it. It wasn't an accident that the same 10-20 people were always in trending.

So, now we have bid bots. Yes, people can buy their way into trending. I like to think I actually pioneered it. While it's not perfectly egalitarian because now you need money to get into trending it's certainly better than it used to be when it was the same faces every time. Now anyone with cash can get into trending and anyone with money can buy a vote; not just the limited few.

I think that's an improvement. I also kinda dig aspects of linear rewards. You like Utopian? Me too. It only operates as it does because of linear rewards. Same with minnowsupport. We couldnt' offer votes to everyone on the platform because they would be worth literally nothing instead of something small.

I do think the vote weight distribution is better under linear rewards too.

Fixes I'd suggest

This thing isn't perfect. Here are some ways I'd suggest making it better. Some of these have bene floated around before. Some I feel like i'm just shouting to myself over and over.

1. Have a separate pool for downvoting and have an option to not hurt reputation. Essentially offer 3 votes
   upvote
   downvote
   flag

Upvote gives rewards
downvote removes rewards
flag removes rewards and hurts reputation

The one number on the platform that really helps determine how much you earn around here (outside of SP) is rep. If you're looking to earn through blogging you're gonna need a high rep. Downvotes that adjust rep have a lasting impact. So, not only are you saying with your flag "I don't like this content" you automatically say "and I hope you don't earn shit here motherfucker" because you effect their long term ability to earn.

It would be better to have a non rep altering flag so you can disagree with a reward without impacting them indefinately.

It would also be good for this to have a separate pool because right now if yhou flag you get no curation reward. That's bad. You're asking whales to give up money which they have the right to earn through their investment to clean up the community. If flagging didn't cost them anything I think it would be more likely and especially if it didn't hurt rep I think it would be used a lot more to actually level set content on the blockchain. With the stakes a little lower maybe that would reduce revenge flagging too.

1. Stake weighted inflation

The worst decision made was when they set the infation to 100% or more and did it all through stake weighted distribution. It was fucking terrible to watch. Rich people got way more steem. Post rewards went way down. Steem price tanked. it sucked. The stake weighted inflation went way down, but it's still there. It's currently set at 15% and has an effective rate a little lower than that. Since Steemit and the exchanges hold a fat percentage of Steem we're giving them an assload more.

260M steem. 9% inflation. that's 23.4M steem. 15% goes to stake weighted inflation. That's roughly 3.51M steem. Something like 60% of that goes to three entities for 2.1M steem. That's roughly $6M dollars right now and potentially a lot more. This seems like a waste to me.

It would take a hardfork to change, but the code should be really fucking simple to execute. In the code change the setting from 15% to 0% for the stake weighted inflation out of overall platform inflation.

What would happen then? That 3.5M steem would be distributed mostly to post authors and curators with a smidge going to witnesses. I'd even be down to shrink witness percent a little to make sure they/we dont' get any additional benefit from this. Ultimately I think that 3.5M steem is better off in the hands of active users than primarily in the wallets of just Steem and the exchanges.

1. Overall platform inflation

I think we should set this to 10% and leave it there. It's currently at 9% and set to shrink by half a percent every 6 months for the next 20 years or so. As that happens the amount of steem distributed will go up for a little bit, and then start shrinking. There will be less and less distributed through posting. I think that's bad. I think it will discourage signups, and I think signups drive growth.

General conclusions

Looks. Shit's working! Price is going up, people are signing up, active users are increasing. Life is good. We have some amazing hard forks coming down at us. I think those are going to drive price which will in turn drive users which will later drive price. I think we're in for a great year.

That said, linear rewards is prone to abuse and could use more downvotes for price discovery on content.

Bit bots aren't the devil and going back to exponential rewards while we still have a terrible distribution will make things go from ok to bad.

It could use less stake weighted inflation to Steemit and teh exchanges.

it could use a slight tweak up for platform inflation and let it stay there.

I know some very smart people that don't like some or all of these plans I've listed and who knows how much of this would or can happen. It might be that Steemit's other priorities keeps them from happening regardless of what the platform and witnesses think at least for the next year or so. That's ok, I happen to think they're working on a good list of priorities right now and the platform is experiencing healthy growth even though btc and Steem have been correcting pretty hard.

$0.02,

hope the background helps some frustrated folks on the chain