Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - bitcrab

Pages: 1 ... 54 55 56 57 58 59 60 [61] 62 63 64 65 66 67 68 ... 129
901
中文 (Chinese) / BSIP42反思与展望
« on: November 29, 2018, 02:15:02 pm »
BSIP42负反馈喂价改革的实验现在可以说是告一段落,那么也应该对其做一个整体的反思。

关于负反馈喂价的大部分讨论是在中国社区进行的,最多的讨论在https://bitsharestalk.org/index.php?topic=26315.0 ,bitCNY一直都存在“如何保证供应,如何博阿正锚定”的问题,“过剩则减,短缺则加”的负反馈思路很自然地被人提出来了,首先想到的还是通过调节MCR来调节智能货币的供给,但是由于涉及到一些bug的修复而难以尽快实施,就把思路转到了通过调整喂价来调整供给上面。

通过调整喂价来调整供给有一些天然的缺点:

1.喂价本来的定义就是市场价格,如过实施负反馈喂价那么喂价就被重新定义为一种“指导价”,这非常难以向用户解释。

2.由于喂价和爆仓,强清等直接相关,喂价偏离市场价格后也影响到爆仓和强清的正常执行,如后来所看到的,但喂价高于市场价幅度达到10%以上时,爆仓单的价格就会大于市场价,因而导致了大量的爆仓单难以被市场消化,造成风险累积。

最初,我也因为这些原因反对负反馈喂价,但由于市场确实需要一种调节机制并且一时难以从MCR入手,后来我被abit说服,同意负反馈喂价方案并开始积极推进,再后来abit起草了相应的BSIP,投票通过后开始着手实施。

BSIP42的实施是从bitCNY开始,因为bitCNY的市场深度和活跃度都比bitUSD好很多,更适合做这样的实验。

八月底BSIP42开始在bitCNY实施的时候,正值BTS价格刚从1.2元跌至0.8元左右,bitCNY出现了明显的溢价,BSIP42的实施立竿见影,很快消除了溢价,但是始料未及的是,市场对bitCNY的需求也一路走低,导致当初高于内盘价格的喂价不得不被调整至低于内盘价格,导致一些原本按市场价不应爆仓的仓位爆仓,引起了很多怨言。

BSIP42的实施也使得BTS内盘价格与外盘价格更紧密的绑定在一起,按照负反馈的逻辑,内盘价格低于内盘视为bitCNY有溢价,反之则视为有折价,溢价上调喂价,折价下调喂价,这就引起内盘价格上升会导致喂价下调,外盘价格降低也会导致喂价下调,内盘价格需要保持比外盘低才能稳定喂价,在一定程度上这种现象缩小了市场空间,降低了市场活力。虽然后来通过折价容忍的手段尝试解决过,但还是没有从根本上上消除影响。

还有一个问题是,市场价格是瞬息万变的,由见证人实时提供动态的喂价调整使得喂价变化难以预测,很多时候让交易者无所适从。

但从效果上看,BSIP42确实让bitCNY锚定得更好了,在稳定币成为市场风口的时候,这一现象是让人兴奋的,阿姆斯特丹bitFEST期间,我同xeroc等西方理事讨论在bitUSD实施BSIP42的计划,但xeroc这种典型的“保守的德国人”基本上总是在说“投票,八周后看结果”之类的话,而当时一些俄罗斯和新加坡的人士已经在讨论做bitUSD承兑的事情,我觉得BSIP42需要再bitUSD上尽快实施。

于是后来我就以手中的选票为筹码,要求见证人们在bitUSD上实施BSIP42,后来引起了xeroc的不满,讨论之后达成妥协:xeroc原则同意在bitUSD上实施BSIP42,而我也同意见证人可以自己选择是否在bitUSD上实施BSIP42,后来bitUSD就进入了一种不是很完整的BSIP42实施状态。

之后市场一路走熊,更严重的危机在慢慢积累,由于负反馈喂价导致喂价一直都比内盘价格高不少,于是市场积累了许多没有消化的爆仓单,而在bitUSD市场这一现象又比bitCNY市场严重得多,终于在几次大跌之后,bitUSD市场在2018年11月25日发生了黑天鹅,所有的债仓被全部清算。

大的市场形势的极度恶劣是黑天鹅的首要原因,当然BSIP42本身也增大了黑天鹅发生的可能性。

之后bitCNY也屡屡陷入黑天鹅边缘,但因见证人们及时做了黑天鹅防护,最终没有发生黑天鹅。

但是让人遗憾的是,市场的换手导致反对BSIP都投票多过了支持反对BSIP42的投票,按理说,对于一个已经过去许久的投票,即便要重新根据投票结果执行,也应该理好足够平滑的过渡路线,可是许多见证人选择了直接转换,其结果就是爆仓单对价格再度形成下压之势,而bitCNY再现5%-10%的高溢价。

这都是在719硬分叉之前没少看见的场景,本来不应该在这个时候重现,遗憾的是又一次出现了。

BSIP42有不少的问题,但是这实验也至少证明了,在一定范围内bitCNY的锚定是可以保证的,脱锚的害处大家都理解,不用多讲,在这个场景下也很容易理解BSIP42的必要性:bitCNY极度短缺,溢价高企,巨大的爆仓单压制的内盘价格并顺带也压制了外盘价格,这种时候,不正是采取措施释放出更多bitCNY的时候吗?明明可以有刺激生态活力的机会,却在那里干等着,让bitCNY持有者们慢慢等待最好的吸筹机会?

大换手没什么好抱怨的,那都是市场行为的结果,既然上了桌那就认赌服输,但是让bitCNY再次陷入毫无必要的高溢价,让市场再度对bitCNY产生怀疑,到底有啥必要啊?

我的建议是在这个时候重启BSIP42,当然加入一些更新:

1.增加黑天鹅防护,确保喂价 >全局清算价*1.11

2.确保喂价高于市场价

3.确保喂价低于市场价*MSSR

4. 如果1与3冲突,舍弃3

5.对-2%-1%之间的溢价作容忍,不更改喂价。

当然,这也是一个临时的方案,长远看,还是要推动态MCR方案。

但就算是动态MCR方案,如果依然是用见证人喂的方式来决定最终MCR,依然有着以前存在的问题 - MCR变动太频繁而且难以预期,交易者难以适应。

如果技术上可以解决,我更倾向于让见证人喂价格和溢价,而由系统根据溢价来自动调整MCR,就象POW系统里自动调整难度一样。

比如规则可以设定为,系统每小时更新一次MCR,溢价在-2%-1%区间不改动MCR,如果超出范围则更新时向相应方向改动一个千分点。

这样MCR的更改是可以预期的,而且一天最多改动2.4个百分点,方便交易者应对。

902
Quote
1.price fall continue, bitCNY holders seems to wait for lower price, no big will to eat current margin call orders.

That is just speculation... Nobody can guarantee they will not eat current margin. Actually I personally believe if you let the market free as possible it will happen in no time. Raise MSSR to 20% for start and then you will see how margin will get eaten!

yes, some more margin order will be eaten, and people can buy cheaper BTS in DEX, BTS price in CEX will fall, and then feed price will become lower, more debt position margin called, bitCNY will be in higher premium - this is positive feedback, we have seen similar process before 719 hard fork, unfortunately we now see this again. we have better way to revive the whole ecosystem.

903
中文 (Chinese) / Re: BISP42 紧急修正案
« on: November 29, 2018, 12:26:48 pm »

904
woker 1.14.140 created, kindly vote if you support this.

905
as some time ago the voting power that oppose BSIP42 overcome that support BSIP42, most of the witnesses return to feed just market price, immediately the BTS/bitCNY price fall about 8%, although more margin call orders are eaten, but:

1. price fall continue, bitCNY holders seems to wait for lower price, no big will to eat current margin call orders.

2. bitCNY has a premium of about 10%, it's a disaster for bitCNY business.

3.more debt position margin called, not obvious margin call quantity reduction.

actually witnesses need to be careful while doing a big change, it is not a good behavior to just change while seeing the voting result - no one host/claim/announce the voting activity, it is likely that some balance owner change lead to this result, please be careful while seeing this!

here I propose to restart BSIP42, with adding some new logic:

1.add the black swan protection logic, ensure feed price >global settlement price*1.11

2.ensure feed price > market price.

3.ensure feed price < market price*MSSR

4. if 1 conflict with 3, ignore 3.

5.tolerate a -2%-1% premium, do not adjust feed price if bitCNY premium is in this interval.

this poll will be regarded successful if it get more voting weight than 1.14.119


(2) only in case of premium imho...

actually in Chinese community there's a lot complaints against the fact that feed price can be lower than market price, some think this hurt them by margin calling a debt position with enough collateral.

906
as some time ago the voting power that oppose BSIP42 overcome that support BSIP42, most of the witnesses return to feed just market price, immediately the BTS/bitCNY price fall about 8%, although more margin call orders are eaten, but:

1. price fall continue, bitCNY holders seems to wait for lower price, no big will to eat current margin call orders.

2. bitCNY has a premium of about 10%, it's a disaster for bitCNY business.

3.more debt position margin called, not obvious margin call quantity reduction.

actually witnesses need to be careful while doing a big change, it is not a good behavior to just change while seeing the voting result - no one host/claim/announce the voting activity, it is likely that some balance owner change lead to this result, please be careful while seeing this!

here I propose to restart BSIP42, with adding some new logic:

1.add the black swan protection logic, ensure feed price >global settlement price*1.11

2.ensure feed price > market price.

3.ensure feed price < market price*MSSR

4. if 1 conflict with 3, ignore 3.

5.tolerate a -2%-1% premium, do not adjust feed price if bitCNY premium is in this interval.

this poll will be regarded successful if it get more voting weight than 1.14.119
-----------------------------------------------------------------------------------------------------------

因为BSIP42投票刚刚反对票多过了支持票,于是一些见证人就直接改回了真实市场价喂价,导致内盘BTS价格瞬间跌了7,8个点,虽然帮助消化了一些低抵押率爆仓单,但是造成的恶果是:

1.价格再持续下跌,bitCNY吃爆仓单的意愿并没有足够强,期待更低价格的情绪明显。

2.bitCNY再次溢价10%,大家可以想象带来的恶果是什么。

3.喂价持续下跌带来了更多的爆仓单。

BSIP42投票有特定的计票日,即便要取消BSIP42也不应该采取这么生硬的方式。而且在现在这样的非常时期是需要BSIP42来提升活力并确保bitCNY供应的,我提议对BSIP42进行更新之后重新启动,更新如下:

1.增加黑天鹅防护,确保喂价 >全局清算价*1.11

2.确保喂价高于市场价

3.确保喂价低于市场价*MSSR

4. 如果1与3冲突,舍弃3

5.对-2%-1%之间的溢价作容忍,不更改喂价。

本投票当获得支持超过1.14.119时生效。

907
中文 (Chinese) / Re: 呼吁见证人重启BSIP42喂价
« on: November 29, 2018, 11:46:19 am »
算了,我来发一个英文的吧

https://bitsharestalk.org/index.php?topic=27492.0

908
中文 (Chinese) / Re: 呼吁见证人重启BSIP42喂价
« on: November 29, 2018, 11:44:41 am »
兄弟啊,你就不能写成英文的啊,你这让老外怎么看得懂啊?

909
中文 (Chinese) / Re: 呼吁见证人重启BSIP42喂价
« on: November 29, 2018, 09:25:14 am »

910
中文 (Chinese) / Re: 喂价负反馈紧急修正案
« on: November 29, 2018, 07:38:15 am »
如3和2抵触,则3失效。

支持。

911
General Discussion / please restart BSIP42!!!
« on: November 29, 2018, 06:40:04 am »
as some time ago the voting power that oppose BSIP42 overcome that support BSIP42, most of the witnesses return to feed just market price, immediately the BTS/bitCNY price fall about 8%, although more margin call orders are eaten, but:

1. price fall continue, bitCNY holders seems to wait for lower price, no big will to eat current margin call orders.

2. bitCNY has a premium of about 10%, it's a disaster for bitCNY business.

3.more debt position margin called, not obvious margin call quantity reduction.

actually witnesses need to be careful while doing a big change, it is not a good behavior to just change while seeing the voting result - no one host/claim/announce the voting activity, it is likely that some balance owner change lead to this result, please be careful while seeing this!

now I greatly propose: 
 
1. witnesses please restart BSIP42, however according to what we learn in the past, I suggest to add black swan protection logic, and ensure the feed price is not lower than market price and not higher than market price*MSSR.

2. please vote to support BSIP42.

3.start the process to draft updated solution to relevant issues.

912
中文 (Chinese) / 呼吁见证人重启BSIP42喂价
« on: November 29, 2018, 06:27:41 am »
因为BSIP42投票刚刚反对票多过了支持票,于是一些见证人就直接改回了真实市场价喂价,导致内盘BTS价格瞬间跌了7,8个点,虽然帮助消化了一些低抵押率爆仓单,但是造成的恶果是:

1.价格再持续下跌,bitCNY吃爆仓单的意愿并没有足够强,期待更低价格的情绪明显。

2.bitCNY再次溢价10%,大家可以想象带来的恶果是什么。

3.喂价持续下跌带来了更多的爆仓单。

BSIP42投票有特定的计票日,即便要取消BSIP42也不应该采取这么生硬的方式。而且在现在这样的非常时期,取消BSIP42如何做好过渡至关重要,当前,我呼吁:

1.见证人重启BSIP42喂价,但接受前面教训,喂价应该保证不低于黑天鹅保护价以及市场价,同时不高于市场价*MSSR。

2.社区投票支持BSIP42。

3.启动新版解决方案的制定。

913
中文 (Chinese) / Re: 喂价反馈紧急修正案
« on: November 29, 2018, 06:02:55 am »
巨蟹讲的有道理,估计难通过。那能否把上限定在20%?毕竟若就10%那就跟下调MSSR到0没区别了。
关于速率问题,我觉得CNY的允许波动范围应当能缓解下降速率。

有区别,调高喂价可直接增加可借出量,MSSR调为0没这个效果。

我知道。修正案的核心就是允许喂价在cny溢价的时候上调,往往是现价大跌的时候出现溢价。现在就是允许现价跌的时候,喂价在一定范围内不跌。喂价0.61的时候发生了现价大跌的情况,那时喂价不跌很好的保护了bts,没有造成连环爆仓。但是错就错在没有一定的折价容忍,使喂价缓慢下调。现在若只允许喂价高于现价一个MSSR,则连环爆仓还是会发生呀。至少应该比MSSR大那么一点吧。

不会,当喂价=市场价*MSSR的时候,爆仓单卖价也就是市场价,如果bitCNY有溢价那么爆仓单卖价比内盘价格还高一点点,带不来搬砖机会,也就没有连环爆仓。

914
中文 (Chinese) / Re: 喂价反馈紧急修正案
« on: November 29, 2018, 05:18:37 am »
巨蟹讲的有道理,估计难通过。那能否把上限定在20%?毕竟若就10%那就跟下调MSSR到0没区别了。
关于速率问题,我觉得CNY的允许波动范围应当能缓解下降速率。

有区别,调高喂价可直接增加可借出量,MSSR调为0没这个效果。

915
General Discussion / Re: how to revive bitCNY?
« on: November 29, 2018, 04:45:22 am »
gdex-witness plan to update feed price algorithm for bitCNY as below:

Code: [Select]
plan for next update on bitCNY:

Pdex:BTS price in DEX in smartcoin
Pf: current feed price
premium: current premium
GS_price: global settlement price

scale= 0.5;
get Pdex, Pf, premium, GS_price;

black_swan_protection_price = GS_price*MSSR*1.01

while True:
   
   get Pdex, Pf, premium;
   if 0.5%>premium>-1%: ##just adopt the current median if the absolute premium is low enough.
       feed price = Pf;
   else:
       feed price = Pf*(1+premium*scale);
   feed price = min(feed price, Pdex*MSSR)
   feed price = max(feed price, Pdex, black_swan_protection_price)
 
   time.sleep(120); ##update every 2 minutes.

keep the BSIP42 logic, added GS protection logic, ensure the feed price is higher than market price and lower than market price*MSSR

bitCNY now has a premium of 2%+, if removing BSIP42 logic, the positive feed back effect will make price even lower and bitCNY in a higher premium, that is a disaster that actually can be avoided.

hope big proxy can support this updated BSIP42 as a temporary solution for this special moment.

Pages: 1 ... 54 55 56 57 58 59 60 [61] 62 63 64 65 66 67 68 ... 129