HF21 - once all together

This is a translation of my last contribution to the HF21. Unfortunately my English skills are limited. If there are any mistakes, please let me know and I will try to correct them.

---

What is a HardFork

A hardfork is a separation from a blockchain by a new version of the software. Because a blockchain, unlike a traditional database, cannot be updated easily, there are so-called hardforks. Here the contributors of the nodes can decide whether they accept the new version or not. An example of this is the split into Bitcoin and Bitcoin Cash a few years ago, a part of the node operators switched to Bitcoin Cash, another part remained with Bitcoin.

Which version the network uses is decided by a majority vote. At Steem, 17 out of 20 of the top 20 Witnesses must confirm the Hardfork in order for it to become the standard for the network.

---

The EIP system

The most discussed change in HF21 is likely to be the EIP system. Put simply, this system is about the distribution of rewards for authors and curators. This section first explains the previous system, then presents each change and concludes.

The basics

There is a reward pool, which contains a certain amount of Steem. A certain amount of this reward pool is paid out in a certain period of time. How this amount is distributed is determined by the vShares. This is the weighting of a vote.
The weight of a vote depends on various factors. The SP of the voter, his vote power and a weighting of the vote determined by him.

At the moment we have a linear distribution system, whoever gets twice as many vShares as another gets twice as much reward. In the following I will use a simplified system to demonstrate this system.

Examples
We have eight users, Bob, Alice, Max, Dan, Ned, Iglo, Martin and Louis. There are 100 Steem in the pool to be paid out.

We assume that each vote has the same weighting.

Example A
We have two authors who have written contributions.
The votes were distributed as follows:

• Bob: 5 Votes
• Alice: 10 Votes

Bob gets 33,333 steem from the pool, Alice gets 66,667 steem from the pool. She received twice as many votes as Bob and therefore twice as many steems.

Example B
In this case, for example, all eight user posts have been added.
The votes have been distributed as follows:

• Bob: 0 Votes
• Alice: 2 Votes
• Max: 5 Votes
• Dan: 5 Votes
• Ned 10 Votes
• Iglo: 18 Votes
• Martin: 25 Votes
• Louis: 35 Votes

In this case, for example, all users have been retained as many times as there have been votes.

Example C
Let's assume Dan published a shit mail. For this reason five users give one downvote to Dan, so Louis, Martin, Iglo, Ned and Max each get one vote less.
The votes were distributed as follows:

-Bob: 0 Votes

• Alice: 2 Votes
• Max: 4 Votes
• Dan: 0 Votes
• Ned 9 Votes
• Iglo: 17 Votes
• Martin: 24 Votes
• Louis: 34 votes

Effectively 10 votes were taken out of the system, whereby the steems are distributed as follows:

• Bob: 0 Steem
• Alice: 2.222 Steem
• Max: 4,444 Steem
• Dan: 0 Steem
• Ned 10 Steem
• Iglo: 18.889 Steem
• Martin: 26,667 Steem
• Louis: 37,778 Steem
  Note: This list may vary slightly due to rounding errors.

As you can see, five users got one vote less each, but except for Max, all users got at least as much or more Steem than before. Alice also profited, as she did not get a vote deducted from her vote, she still has a profit from the downvotes as the smallest recipient.
Max, in turn, has the problem that he really loses Steem to the Downvote because of the current distribution system, but not because he was downvoted, but because his vote was used elsewhere.

Example D

In this example, we take a closer look at the impact of bidbots. This time 50 votes were distributed by a bidbot to Bob.
The votes were distributed as follows:

• Bob: 50 votes
• Alice: 2 Votes
• Max: 3 Votes
• Dan: 3 Votes
• Ned 5 Votes
• Iglo: 8 Votes
• Martin: 13 Votes
• Louis: 16 Votes

The Steem are distributed as follows:

• Bob: 50 Steem
• Alice: 2 Steem
• Max: 3 Steem
• Dan: 3 Steem
• Ned 5 Steem
• Iglo: 8 Steem
• Martin: 13 Steem
• Louis: 16 Steem

One sees a clear upward distribution caused by bitbots, this becomes clearer in the example D2.

Example D2

We still have 8 users and 100 Steem. However, 1000 votes were used, 900 of which were purchased from Bob. This example is deliberately overstated to further illustrate the impact of bitbots.
The votes were distributed as follows:

• Bob: 900 Votes
• Alice: 2 Votes
• Max: 5 Votes
• Dan: 5 Votes
• Ned 10 Votes
• Iglo: 18 Votes
• Martin: 25 Votes
• Louis: 35 votes

The Steem are distributed as follows:

• Bob: 90 Steem
• Alice: 0.2 Steem
• Max: 0.5 Steem
• Dan: 0.5 Steem
• Ned 1 Steem
• Iglo: 1.8 Steem
• Martin: 2.5 Steem
• Louis: 3.5 Steem

One sees the problem in the current system, through bidbots or a large stake it is possible to exploit the rewards pool, this leads to the fact that it is especially lucrative for whales to vote themselves and thus take the SP out of the system.

Weighting of the vote
Now that we have taken a closer look at the distribution of the Reward Pool, we go to the weighting of the votes. In our example each post was of equal value. In reality, votes have a different value. This results from three factors, the own Steem Power, the Votingpower and the weighting of the vote.

Steem Power
Steem Power is the basic value from which the vShares are determined.

Voting power
Voting power was introduced to prevent individual users from voting excessively and thus influencing the distribution of rewards too much. It regenerates over 5 days and allows 50 votes during this period. Thus 10 votes can be assigned per day and the voting power regenerates to the value at the beginning of the day.
The lower the voting power, the lower the value of the vote.

Weighting of the vote
To cast more than 10 votes a day, it is possible to change the weighting of the vote, which makes it possible to cast up to 1000 1% votes a day without loss.

examples
The user has 1000 SP, these generate a value of 1 STU with a 100% Vote(VG) to 100% Votingpower(VP).

• Example A: 100% Vote, 50% Votingpower: 1 STU * 0,5 VP * 1 VG= 0,5 STU
• Example B: 20% Vote, 30% Votingpower: 1 STU * 0,3 VP, 0,2 VG= 0,06 STU
• Example C: 50% Vote, 80% Votingpower: 1 STU * 0,8 VP * 0,5 VG= 0,4 STU

Distribution of the Reward of a Post
If the rewards were distributed to the contributions, not everything goes to the author. At the moment 25% of the rewards are distributed to the curators. The first curators get more than the others. It can be said that the first 10 voters get a good 90% of the curation. This leads to the fact that there are always candidates who try to pick up this curation. From about the twentieth vote the share of the curation is about 1-3%. Unfortunately I could not find the exact distribution.

To allow more users to get among the first votes there is a voting window. This is a penalty for voting too early. In the first 15 minutes the value of the vote increases linearly. If you vote in the first second, you get virtually nothing from your curation. If one votes after 5 minutes, one gets 1/3 of the curation due to one etc..

Other Recipients in the Reward Pool

In addition to the rewards for authors and curators, there are two other beneficiaries in the Rewardpool. The first is Benificts, which is a fee or donation where part of the reward is transferred to another person for a contribution. This is mainly used by platforms like e.g. d.tube to finance themselves.

The other recipient of the pool are the Witnesses, 10% of the Rewards go to these for the operation of the Nodes.

50:50

h3>

The first change of the HF21 in the EIP system is the change of the reward distribution between authors and curators. In the future, the rewards will be distributed equally between authors and curators.

Background
The change from the current 75/25 system to the 50/50 system should make it more attractive for investors to curate. At the moment, many investors delegate a large part of the SP to bidbots and similar services, as this is often more attractive than curating. This will change after the change.

Another point that this change should improve is the demand situation on the markets. At the moment, more Steem is sold on the markets than bought, which tends to lower prices. After this change this could change as it becomes more attractive to invest again.

A last aspect that this change should bring has the @balte in a comment under @meins0815 contribution mentioned:

I experience this as a "curator" since palnet, makes of course as a curator much more fun to curate at 50/50

You can hear this opinion recently more often when you read reports about palnet. There are several reasons why curating there is fun for many, but in a diluted form it should also work on the Steem itself.

downvote pool

The second controversial change in the HF21 is the addition of a downvote pool to the voting power. This voting pool will amount to between 10% - 25% of the voting pool, i.e. 1-2.5 100% downvotes.

This will be placed on top of the normal voting pool. D.H. each user has more voting power available, since he can now downvote without disadvantages in a limited extent.

We already had an example for downvotes and their effect on the reward pool in the basics in example C.

Example C
Let's assume Dan published a shit mail. For this reason five users give one downvote to Dan, so Louis, Martin, Iglo, Ned and Max each get one vote less.
The votes were distributed as follows:

-Bob: 0 Votes

• Alice: 2 Votes
• Max: 4 Votes
• Dan: 0 Votes
• Ned 9 Votes
• Iglo: 17 Votes
• Martin: 24 Votes
• Louis: 34 votes

Effectively 10 votes were taken out of the system, whereby the steems are distributed as follows:

• Bob: 0 Steem
• Alice: 2.222 Steem
• Max: 4,444 Steem
• Dan: 0 Steem
• Ned 10 Steem
• Iglo: 18.889 Steem
• Martin: 26,667 Steem
• Louis: 37,778 Steem
  Note: This list may vary slightly due to rounding errors.

Under the new system, the five votes do not have to be converted into downvotes. Instead of 100 votes, 105 votes are now in the pool.

In the new system the votes would be distributed as follows:

• Bob: 0 Votes
• Alice: 2 Votes
• Max: 5 Votes
• Dan: 0 Votes
• Ned 10 Votes
• Iglo: 18 Votes
• Martin: 25 Votes
• Louis: 35 votes

The distribution of the steem in this system would look as follows:

• Bob: 0 Steem
• Alice: 2,105 Steem
• Max: 5,263 Steem
• Dan: 0 Steem
• Ned 10.526 Steem
• Iglo: 18.947 Steem
• Martin: 26,3156Steem
• Louis: 36,842 Steem
  Note: This list may vary slightly due to rounding errors.

As you can see here, no user has a disadvantage because of the additional votes, even Max, who would get less rewards in the previous system, gets more rewards in the new system than before, just like Ned, who would get more rewards after the HF21 instead of as much. What's particularly remarkable is that the three users at the top of the reward distribution no longer benefit as much from a downvote as before.

Overall, it can be seen that a downvote is effectively just another form of reward distribution. It strengthens the previous distribution of votes and it is therefore not necessarily a purely negative matter.

background
At the moment there are three forms of voting. The upvote, the downvote and the vote change. Only the upvote brings a profit to the curator, which is why it is actually the only vote form used worth mentioning. Downvotes are only used to a limited extent, they cost money effectively at the moment. Since one can reward a downvote badly, this would promote the abuse only, one decided to put a downvote pool on the regular Votingpool on it. This includes intentionally only 10% - 25% of the general voting pool to prevent abuse on a large scale. Who wants to distribute more Downvotes, must pay for it - as so far also. For all others, this offers the chance to take active action against plagiarism, Nazis and spam and to distribute the rewards to honest authors.

Further note
This part refers to the previous distribution curve. In the next chapter we look at how downvotes affect the new distribution curve.

New distribution function for rewards

So far we use the formula f(x)=x for the distribution of votes and put this in relation. We have seen the result above. In the future the formula f(x)=x^2/(x+1) will be used. Let's have a look at what this means. As in the first part we have 8 users, Bob, Alice, Max, Dan, Ned, Iglo, Martin and Louis.

Example E
We've got 100 votes giving out 100 steem.

• Bob: 0 Votes
• Alice: 2 Votes
• Max: 5 Votes
• Dan: 5 Votes
• Ned 10 Votes
• Iglo: 18 Votes
• Martin: 25 Votes
• Louis: 35 votes

The steems are distributed according to the above formula:

• Bob: 0 Steem
• Alice: 1.385 Steem
• Max: 4,432 Steem
• Dan: 4,432 Steem
• Ned 9.685 Steem
• Iglo: 18.167 Steem
• Martin: 25.61 Steem
• Louis: 36,253 Steem

We're looking at an upward shift in payouts. At this point it has to be said that this model is not suitable to illustrate how the formula affects the future reward distribution. For this reason we will work in the example F with other values closer to the Steem.

Example F
There are 1000 Steem in the pool.
To support new users we extend the formula to f(x)=x^2/(x/250 + 1). This gives users with less than 250 a small bonus.
We go away from the individual users and create the following players:

• A whale
• A group of dolphins
• A group of Minnows
• A group of plankton

The groups are made up as follows:

• a whale
• 50 dolphin
• 200 Minnows
• 500 plankton

We assume that every contribution from a group receives the same value.

The votes are distributed as follows:

• one whale: 10.000 votes
• a dolphin: 250 votes
• one Monnow: 50 Votes
• one plankton: 1 vote

Under the old system, the rewards would have been distributed as follows:

• a whale: 430,108 Steem
• one Deflin: 10,753 Steem
• one Minnow: 2.151 Steem
• one plankton: 0,043 Steem

Under the new system, contributions would be distributed as follows:

• one whale: 551,987 Steem
• a dolphin: 7,072 Steem
• one Minnow: 0.471 Steem
• one plankton: 0 Steem

At first glance, it brings more benefits for large users than before.
Let's take a closer look.

• Alice gets 5 votes
• Bob gets 10 votes
• Roland gets 20 votes

There's another 1,000 votes in the pot.
The 1000 Steem in the pot would be distributed as follows:

• Alice: 4,03 Steem
• Bob: 8,817 Steem
• Roland: 18,474 Steem

Bob got the 2.19 from Alice.
Roland received the 2.1 times from Bob.

You can see that the curve slowly flattens out. If you want to get the maximum out of this system, it is advisable to distribute your votes among many small ones instead of some large ones.

This is where the 50:50 system comes in. Now it will be more interesting for investors to curate small contributions and less the tip of the iceberg. At the top remains a vote of 0.01 STU 0.01 STU. At the contribution of a Minnow, which is just at 0.4 STU, the vote of 0.01 STU is now 0.02 STU value, whereby the curator gets double the payout.

I won't go into downvotes any further. Their behavior will not change significantly due to the new distribution curve.

New voting window

At the moment you can only vote for a contribution after 15 minutes without penalty. At the moment, this means that many posts that are found in the first 15 minutes will not be voted at the end. The curators do not want to receive a penalty for their curation and are therefore waiting for the 15 minute voting window to expire.

This is where the human element comes into play. One looks further, goes some food or watches a good porno - and forgets that one wanted to vote the contribution. This has the consequence that the author is not curated and he escapes income.

To solve this problem, the voting window is reduced to 60 seconds. This gives the curators a better chance of leaving a vote - but it is more likely that the bots will try to pick up the curation.

conclusion

We have a problem. This should be clear to everyone who is concerned with the Steem. EOS is becoming more and more a competitor, they have more capital - but fewer users. This is the strength of the steem, we are people who actively use it. This is a factor that hardly any platforms have.

The new EIP system is primarily designed to leverage demand. Steem consists to a large extent of two groups. The investors and the authors. One group cannot do without the other. At the moment we have the big problem that more Steem is sold than bought. This causes the price to fall slowly and is actually only kept stable by the rising Bitcoin price.

If we want to change this it has to become more attractive to invest in the Steem. This is where the 50:50 system comes in. A side effect should be a rising Steem price, whereby in the end all should receive more Rewards in the form of SBD.

Another point is the new distribution curve, yes it is not perfect and complicated, but if it is provided with the right set screws, it can ensure that voting small contributions for curators will be much more lucrative and so many users who currently look into the tube give the chance to be rewarded for their work.

A last point is the Downvotepool. Like some knowledge I am an advocate of the downvotes. This has the background that I don't see it that users plunder the reward pool for plagiarism, hate, spam or other shit and so honest authors take the reward. There are already some users who take active action against such players, but they pay indirectly for it. In the future, anyone can do something about it. You have something against the same picture being posted 100 times? Flagg it away. You have something against my face? Flagg me away.

All in all I am confident that the EIP system will have a positive influence on the Steem. We can't do more than speculate right now. But if we don't do anything now, this platform may not exist in a year.

---

The SPS - the Steem Proporsal System

Steem wage system. This system allows groups, developers and other stakeholders around the Steem Blockchain to make a request and be rewarded with 10% of the reward pool.

background

At the moment there are few who can afford to work on the Steem on a grand scale. Many projects are financed either by Posts, a Witnessnode or a usage fee. The problem with this is that you can't pay your rent regularly with contributions, running a Witnessnode is also a story in itself and if you don't run your own frontend, you won't get any usage fees. The goal of the SPS is to give a certain security to those developers who only finance themselves through posts at the moment.

Functionality

The system takes 10% of the reward pool and distributes it to the above groups. The disbursement takes place in SBD, as it is better to plan long-term than with Steem. In order to be financed by the system, an application for 10 SBD must be made. This application looks like this:

• account_being_funded (recipient of the money, usually the person or group submitting the proposal, but does not have to be)
• daily_pay (quantity of SBD to be paid out on the day)
• start_date (date from which disbursement is to begin)
• end_date (date on which the disbursement is to be terminated)
• subject (a short description of what you need the money for)
• url (Link to the project)

Subsequently, it will be possible for users to vote for contributions. This system will work just like the Witness system, with two exceptions.

1. There are as many votes as you like
2. Projects are funded as long as they are within the budget.

The weight of a vote depends on the own SP, without consideration of delegations. On the basis of this weighting and the number of votes, a list is created, the projects in this list each have a sum X selected which they need.

This list could look like this:

• Project A: 100 SBD per day
• Project B: 3000 SBD per day
• Project C: 10 SBD per day

In the pot are 3101 SBD a day. Project A and B have received more voting power than Project C and are financed. The SBD that should still go to C goes back to the pool and increases it.

The pool is increased by the reward pool and a certain amount of X from the pool is paid out. The larger the SPS pool is, the larger the payout amount and the more projects can be supported. The disbursement takes place hourly.

conclusion

Altogether I find the SPS system is a function that should be appreciated. It further cuts the author pool but at the same time allows it to work more decentralized. There is a chance that a group of whales will exploit this pool, but at the same time we can support the work on the Steem. How about a project that supports the Steem in the DACH area?

At the current rate, the pool comprises about 1,000,000 USD per year. This sum can be used to finance a whole series of projects. In addition, the pool can be extended by donations, Steemit has already announced a donation of 100.000 SBD if the system works.

---

Custom JSON

h2>

A final point in the HF21 is the increase in custom JSON per block and user.

Background

A Custom JSON is one of four actions on the blockchain, there are posts, comments, votes and Custom JSON. A Custom JSON is everything that doesn't fall into the first three categories. It can be many things, the information of a Witness how much the Steem is worth, a Follow on Steemit, a new building at @nextcolony or an @splinterlands match.

At the moment, the number of custom JSONs per user and block is limited to one JSON, so only one can be done every three seconds. The background to this decision is a DoS protection, which can now be guaranteed by the RC system.

The discussion

h3>

This article discusses how the Custom JSON system should be changed. In my last post on this topic I made a mistake. There is no result yet what @oliverschmid pointed me to.

At the moment there are two options being discussed:

• 5 Custom JSON per block and user

This would have the advantage that a certain DoS protection is still given but at the same time projects around the Steem Blockchain could interact better with the Blockchain.

• Eliminate the Custom JSON limitation completely

This has the advantage that you could handle a lot of actions in one block. However, there is a risk of DoS vulnerability if a user with a lot of SP is doing it.

In the article linked above you can discuss this and leave your opinion.

---

Final words

After more than 24 hours of sitting at this post I hope that the HF21 will be more understandable for one or the other. I myself have learned a lot here. If you notice any mistakes or have any questions, please contact me. I will now take care of my exams and my Witnessnode.