Quote from: bitcrab on December 07, 2018, 08:08:37 am

发展智能货币与发展其它生态不矛盾。发展生态的事情从来都没人拦着。

向交易所推广从来都不是妄想。

bsip42有固有缺陷，新版的负反馈方案见这里 https://bitsharestalk.org/index.php?topic=27506.0

当然，现在的形势下，新版有没有机会获得足够支持来实施，是另外一个问题。

蟹哥，本来市场就风云变幻，不可预测的影响因素很多，机制越简单越具有鲁棒性，无论是负反馈还是动态调整MCR，引入了动态变量，结果的不可预知性被不断放大。单纯调MSSR引起的变量最少，市场对于机制可以有明确的认识。

另外，在BTS目前这个市值情况下，锚定精准意义真的不大。

bsip42有一些缺点，这使得它招致了很多反感。

但不能泼洗澡水的时候把孩子也一起泼出去。

负反馈机制的作用不仅仅在于锚定精准，更在于根据市场调整供需。

前几天的情况表明，即使有了目标抵押率，极端情况下连环爆仓的正反馈依然会起作用，这种情况下如果没有一种对应的机制来在风险可控的范围内扩大供应，智能货币生态会快速衰落。

“机制约简单越有鲁棒性”从没得到过证明。

不说了，随它去吧。

now bitCNY has a global settlement price of 0.2568.

the current BTS price is 0.247CNY.

this means if now witnesses feed real market price bitCNY is already in black swan status.

now most of witness add "black swan protection" logic to ensure they feed price > global settlement price*1.11

this help to avoid black swan, and now bitCNY still has a premium of about 4%.

anyway, I think this is much better than just let global settlement happen.

actually we have suggest to add same protection logic in bitUSD price feeding, but seems only little witnesses has done that.

中文贴在这里 https://bitsharestalk.org/index.php?topic=27506.0 中文请到该贴回复。

Chinese version is here https://bitsharestalk.org/index.php?topic=27506.0 please switch to there if you use Chinese.

It seems we kind of agree on the general direction so my suggestion would be to implement things in an incremental way.

To start:
Settlement offset as low as possible (would like it to even be 1% on par with all other bitassets)
Some kind of GS protection

and personally I think our first attempt to improve things should be to allow for variable MSSR according to circumstances.

in most of the cases margin call orders are filled at the market price in CNY market, so it make almost no difference either to set MSSR to 10% or 5%, maybe only when in very serious bear market 5% or 10% make some difference.

we learnt a lot after BSIP42 implementation, now we need to consider more on how to update bitCNY, actually there are a lot discussion on this topic in the community, especially in Chinese community, here I'd like to propose some new solution summarizing all the opinions.

1. change the way to handle black swan, keep the borrowing operation alive.

global settlement is a too direct  and too hard way to handle black swan, it stop the smartcoin borrowing business, while it happen it always give big negative impact on smartcoin ecosystem, we can see that bitUSD devalued a lot after the global settlement, the community require a more soft  way that has more little negative impact on the smartcoin ecosystem.

abit has suggested 2 new ways to handle black swan, one is to settle the accounts independently: https://bitsharestalk.org/index.php?topic=27273.0, one point is that this way may cause one phenomena that "to be settled earlier is better than later".

the other is to set feed price floor to provide black swan from happening: https://bitsharestalk.org/index.php?topic=27449.0
the current bitCNY is similar to be handled in the second way.

either way we will choose, it is important not to stop the smartcoin borrowing operation.

2. replace BSIP42 with dynamic MCR solution

the key point of dynamic MCR is also to introduce negative feedback mechanism to adjust smartcoin supply, different with BSIP42, in dynamic MCR solution the feed price will return to reflect the market price, then this solution will not bring the troubles of "margin call orders not being eaten" or "exploiting debt position owners with force settlement".

considering that in BSIP witnesses take the key responsibility and authority, this introduce much uncertainty and put witnesses at the place that undertake much doubts of feed price manipulation, here I suggest a new way to implement dynamic MCR:

2.1.  other than settlement price, CER, MCR, MSSR, introduce a new parameter premium to be fed by witnesses。witnesses need not to design complex negative feedback algorithm, they just need to get the price data from the market and feed to system.

2.2 add committee controlled parameters as below to each smartcoin:

low premium limit: when the fed premium is lower than this parameter, the process to make MCR higher will be triggered, default value: -2%
high premium limit：when the fed premium is higher than this parameter, the process to make MCR lower will be triggered, default value: 1%
max MCR：MCR will not increase while it reach this value, default value: 1.75
min MCR：MCR will not decrease while it reach this value, default value: 1.4
MCR tuning step：the value that MCR can be added or subtracted in each administrative period(1 hour), default value:0.001

at each administrative time system will check the feed premium and decide how to adjust MCR, if the parameters are set as default values shown above and at the time MCR=1.75, and premium=2%, then MCR will be adjusted to 1.749.

in this way the traders can expect the adjustment time and step, the can reflect in time to avoid being hurt, no manipulation doubt can be put to witnesses.

3. adjust MSSR to 105%

4.change force settlement offset to 2%, this will ensure that the discount of bitCNY will be no higher than 2% even after MCR reach the max limit.

what will this solution bring?

1. good peg of bitCNY
2. less volatility of BTS price
3. more risk?
4. less free market?

easy to understand that this solution will lead to good peg of bitCNY, as BSIP42 has proved, negative feedback mechanism works well on making a good pegging.

as MCR will change according to the premium of bitCNY, when bitCNY is in shortage and has a big premium, system will increase MCR and make more bitCNY supply possible and vice versa, for example, if we implement the solution at the current market status, we can expect that the MCR will be tuned to min value(maybe 1.4) in about 2 weeks, this will increase bitCNY supply obviously and can prevent BTS price from falling rapidly.

so will this solution introduce more risk to the system?
I believe this will depend on whether the margin call orders will be eaten quickly, the main issue of BSIP42 is that is make the margin call order price higher than the market price and accumulate big quantity of margin call orders without being eaten. but in the dynamic MCR solution, as feed price reflect the market price, we believe the margin called orders will be eaten just after they comes out, if necessary MSSR can be keep unchanged to ensure this.

and, I don't agree that this rule will bring a manipulated market that has been free when there is no such solution, they are just markets with different rules,  if the rules are certain and base on strong consensus, it will not bring manipulation, it just bring better user experience.   

I also suggest that witnesses can implement updated BSIP42 before community can reach consensus on above described solution and implement it.https://bitsharestalk.org/index.php?topic=27492.0

I didn't see any point to  get profit from it, if you have you should do it.
As a trader I will never borrow a cent  if you can make me margin call without  price of BTS change, sometimes even when BTS price increase.

can I manipulate the feed price at will? or any witness can? if I can why I let my accounts be margin called?

I understand that because bsip42 let witnesses feed price following a complex logic, it introduce some uncertainty that make traders do not feel good, so in the newly advised alternative solution witnesses only feed price and premium data, which are easy to verify accurate or not, the negative feedback logic will be done by the system in a low frequency.

the system need negative feedback logic, the rule of original system is far from satisfactory and cannot sustain the dream of Bitshares. and I don't think we can call DEX free market only when we make no change to the rules.

Quote from: bitcrab on December 03, 2018, 02:06:07 pm

Quote from: btspp on December 03, 2018, 01:52:02 pm

2.1和2.2部分，我不觉得这样喂溢价合适。如果有人觉得见证人操作喂价，他们也会认定操作溢价。没什么区别。反而这样会失去灵活性，如果这种方式最后发现不太合适。要修改，估计又需要等到下次分叉了。

关于mssr我建议动态调整，可以根据一定规则评估黑天鹅风险系数。风险系数大的时候mssr增加，鼓励搬砖吃爆仓单。风险系数小的时候mssr降低。

那你觉得怎样合适呢？

MSSR也搞动态的话感觉太复杂了。

喂的数据还是直接喂“价格”就行，只是算法调整，保留灵活性。
MSSR动态确实需要一些规则，复杂性会上升。目前来说就固定测试一些参数就可。105 102都可以测测看到底哪些适合。

没有一种反馈机制，就无从通过自适应调整来匹配市场需求。
不获取溢价信息，系统就无从知道该向什么方向调整。
价格和溢价都是很明确的，见证人喂的质量如何是不难判断的。
MSSR起的作用不如MCR明显，直接。