mip-002 : Adjust the ETP reward rate of Deposit function

[betachen]

Refers to : https://github.com/mvs-org/metaverse/issues/203

[betachen]

In Chinese, will be translated in English:

总量以及分布情况

元界ETP总发行量1亿。其中：

持有者	数量
基金会初始资金	25,000,000
ICO投资者持币	25,000,000
挖矿Coinbase奖励	30,000,000
存款利息奖励	20,000,000

我们可以根据 此API实时查询当前已经产生的利息奖励总和。 单纯看存款功能，它是一个通胀模型，最大年化收益可达20%。 换句话说，ETP总量1亿的上限，肯定会在某个时间点被超过。

分析过程

以下是计算过程，作为结论并被引用，估算为主。

根据我们统计的数据，存款的分布大致是U型分布，存款1年占大多数，1年偏多，中间存款占少数。 我们取假设全网存款平均收益率为13%。 按照以往统计量，累计超过1000万的ETP参与过存款引用API1，目前全网新增的存款利息总量引用API2约144万。 可计算全网存款实际平均收益率为14.19% （当前时间实时），与我们估算吻合。

1000 万占流通量的17.94%，也就是假设每年总量的17.94% 会参与存款。 1年大概会有100万个新块产生，即约 285万个ETP线性新增，考虑ETP可用，我们取半为142.5万。 所以我们可估算每年在已有基础上新增约142.5万*17.94=25万个ETP锁仓，我们可进一步估算约14年才能到达总量1亿。

即使全网存款量增加一倍，即占比 35.88%，也需要超过7年才能达到全网总量1亿。

结论是，如果不做共识修改，预计在2032年超过1亿总量，极端情况2025 年总量才能到达1亿。 所以我们有足够的时间将存款功能作调整。

[betachen]

解决方案

假设我们按照挖矿的奖励进行相同的衰减模型设计， 设

• 每N个块衰减一次，衰减次数记为n
• 衰减系数m
• 目前存款利息总和为 s
• 每个块固定产出b个ETP 那么我们有

代入衰减系数0.95，衰减高度间隔50万，假设主网激活MIP002时利息总和150万，我们有：

我们可计算b=1.85个。

[betachen]

假设按照以上解决方案进行实施， 我们根据每块产生1.85个ETP作为利息，那么如何分配这1.85个ETP？ 有如下方案：

1. 按照当前块中产生存款交易的金额等比例分配
2. 按照当前块中产生存款交易的币龄等比例分配
3. 按照当前块中产生存款交易的金额的前几名随机选择一名奖励
4. 按照当前块中产生存款交易的币龄的前几名随机选择一名奖励

其中3，4属于博弈，类PoS系统，为了系统更稳定，我们需要设定在3或4的情况下进行证明。

同时考虑在此系统中，我们是否可以为MVS设计混合挖矿机制(Hybird PoW/PoS)系统，让PoS和PoW挖矿同时进行，对此我们研究了 TwinsCoin, HCash等多种混合挖矿的币种。

[betachen]

混合挖矿机制币列表

币种	结论
ETH	待研究
Hcash	文档不多，代码为主，需要进一步研究
nevacoin	基于BTC的altercoin，代码偏旧
steepcoin.net	开源，有白皮书，可参考模型
TwinsCoin	开源，有论文，可参考模型
emercoin	白皮书，可参考模型
byzcoin	待研究

[betachen]

方案1，2 具有被攻击的可能性，等比例会催生大量存款交易。 方案3，可以币数量可以重复使用，进行刷存款利息，这是不合适的。

[Daphneyan]

ETP Total Amount and its Distribution

ETP Total Amount and its Distribution

Holders	Amount
Initial Holdings of Foundation	25,000,000
ICO Investors	25,000,000
Mining and Coinbase Rewards	30,000,000
Interest Rewards for Deposit	20,000,000

We can query this API to query the total interest rewards generated in real time. Presently, the coinlock interest reward we have designed is always greater than 1. As such, if we consider the deposit function in isolation, it is an inflationary mechanism with the maximum annualized income being 20%. In other words, we will inevitably exceed the 100 million ## Analysis Process The following are calculations are predominantly based on estimates because the results will be referenced in other sections.

Analysis Process

The following are calculations are predominantly based on estimates because the results will be referenced in other sections.

According to our data, deposits have followed a roughly U-shaped distribution. 1-year deposits make up the majority, with middle-duration deposits being the least common. We will assume that the average yield of deposits across the entire network is 13%.

Based on previous statistics, over 10 million ETP have been deposited at some point of time, reference API 1. At the moment, interest rewards across the entire network total about 1.44 million ETP, reference API 2. From these figures, we can calculate that the observed average yield of deposits across the entire network is 14.19%, which is in line with our estimations.

10 million ETP makes up about 17.94% of the circulating supply, hence we assume that 17.94% of each year’s circulating supply will be deposited. ETP cap at some point in time in the future.

bout 1 million new blocks will be generated each year, which means that 2.85 million ETP will be added to the circulating supply in a linear fashion. Taking into account the availability of ETP, we take half the ETP mined = 1.425 million ETP. Hence, we estimate that an additional 1.425m * 17.94% = 250,000 ETP will be deposited each year on top of the baseline amount, and further estimate that it will take about 14 years for the total supply of ETP to reach 100 million . Even if deposits across the entire network double to 35.88%, it would require over 7 years for the total supply of ETP to reach 100 million. We conclude that with no change in the consensus mechanism, the total supply of ETP will exceed 100 million in 2032. In the extreme case, we will reach 100 million in 2025.
Thus, we have sufficient time to adjust the deposit function.

Solution

Suppose that use the same decay model as found in the mining reward mechanism. Assumptions:

1. The amount decays once every N blocks. Let the number of decayals be n.
2. Let the decay factor be m.
3. Let the current total interest be s.
4. Each block generates a fixed amount of ETP b. Then we have:

introduce the attenuation coefficient 0.95 and the attenuation height interval 500,000, assuming the total interest is 1.5 million when the main network activates MIP002: Let the decay factor = 0.95 and N = 500,000. Assuming that the current total interest s = 1.5 million when MIP 002 is activated, we have:

We can calculate b = 1.85.

Assuming that the above solutions are implemented, Let us assume that 1.85 ETP is generated each block as interest. Then, how should this 1.85 ETP be allocated? We have four options:

1. Allocate it proportionately according to the value of deposit-generating transactions in the current block
2. Allocate it proportionately according to the coin age of deposit-generating transactions in the current block
3. Randomly select one depositor from the higher-value deposit-generating transactions in the current block.
4. Randomly select one depositor from the higher-coin-age deposit-generating transactions in the current block. Like PoS, options 3 and 4 can be categorized as game-theoretic systems. For increased system stability, we need to provide a proof under situations 3 or 4.

Meanwhile, we are considering the possibility of designing a hybrid PoW/PoS system for Metaverse that would enable simultaneous PoW and PoS mining. With regards to this, we have analyzed TwinsCoin, Hcash and other mixed-mining tokens.

Mixed Mining Coin List

Coin	Conclusion
ETH	To be studied
Hcash	Not many documents, code-based, requires further analysis
nevacoin	Altcoin based on BTC, fairly old code
steepcoin.net	Open-source with white paper. Can refer to their model
TwinsCoin	Open-source with white paper. Can refer to their model
emercoin	Has white paper, can refer to their model
byzcoin	To be studied