Author Topic: Discussing an attack on BitUSD  (Read 4529 times)

0 Members and 1 Guest are viewing this topic.

Offline merivercap

  • Hero Member
  • *****
  • Posts: 661
    • View Profile
    • BitCash
Recently at Decentral in Vancouver we had a meetup with Max Wright to discuss BitShares and BitUSD:

https://www.youtube.com/watch?v=p023lanFsJE

This inspired me to write a blog about a possible attack on BitUSD and similar derivative currencies:

http://tpbit.blogspot.ca/2015/05/inert-versus-volatile-currencies.html

I am wondering - has there been some analysis done on this subject by the BitShares developers and if so, how do they assess the risk involved with a similar attack?

I think it's a good exercise to look at these attacks in general. 

Few points:
1) Not sure about your statement that a currency is inert.  Currencies are dynamic with a lot of factors that move them in the forex market.  BTS is also an asset like company shares so it's even more dynamic. 

2) There are traders/speculators/investors that try to evaluate pricing every moment so the idea that you can move the market and it will remain the way you left it is an odd introductory statement.*

3) I have to read Currency Wars, but a lot of times governments try to artificially manage their currencies and the free market eventually puts governments back in their places.  As long as the design allows free market participation, any type of manipulation will be difficult and can work itself out. 

4)  I do agree rapid moves up and down of BTS can cause some margin pressure, but as long as trades are at the market it's not that easy to manipulate because accumulating or selling aggressively will push the market against you.   Also anyone can enter the market at anytime so new collateralized positions will be entering/exiting continually. 

5)  Your example should be logarithmic because the returns from one level to each level 200 to 300 to 400 is 100%, 50%, 33% returns respectively...and going back down is -25%, -33%, -50%...  not sure how realistic the increasing margin levels are in your example... Rather than escalating it could very well be pretty level or random as market participants move in and out at various times with various levels of collateral.  The CFD is just a wealth transfer back and forth.. you can double your returns or double your losses... the nature of volatility is a bigger issue than margin calls and that's something shorts/bitUSD creators have to manage...

6)  Your first scenario of shorting in an external market is incredibly difficult.  You need traders on both sides of a market.  A prediction market or offline exchange needs someone going long to your short.  I remember someone mentioning Bitcoin miners can short off-exchange, attack the Bitcoin network and make money.  While theoretically possible, you still have to generate large enough buy orders off-market and this would be an incredible difficult challenge, especially since a large short position will eventually leak and be a signal to the larger exchanges what is happening. It should raise eyebrows of anyone off-exchange to take the other half of a massive short bet.   

*Note: I do concede and have mentioned before that prices and avg shareholder's perceived value are not always intimately connected, especially in an illiquid stock.   For example a company with 10 million shares can have 9.9 million shares owned by long term investors that think the shares are worth $2/sh for a total perceived market cap of $20 million.  If 100k shares are traded among speculators and traders at 50 cents and $1/sh, and if the long term owners don't support the market and buy more or care what the market exchange rate is, the perceived market cap would be $5 - $10 million.  Hence traders can dictate the market cap...that's why there should be a caveat with external price feeds especially in illiquid markets.  Or perhaps collateralization should be tied to liquidity/trading volume.  It is good to consider various kinds of attacks considering the disconnect between prices and value.   
BitCash - http://www.bitcash.org 
Beta: bitCash Wallet / p2p Gateway: (https://m.bitcash.org)
Beta: bitCash Trade (https://trade.bitcash.org)

Offline merivercap

  • Hero Member
  • *****
  • Posts: 661
    • View Profile
    • BitCash
This is actually a general attribute of all assets. BitUSD can be considered a debt instrument that provides credit to holders of bitshares (that they then use to buy more bitshares with). If the value of BTS falls, the access to credit of the holders of BTS also goes down. This also happens with all companies that use credit, or the holders of any asset that is used as collateral for loans, so it is a type of speculative attack that already exists. I think the big question is if the liquidity of an asset is high enough to support the amount of it that is used as collateral at any point in time, which is why I think there needs to be "debt ceilings" for collateral types the size of which correlates with their total market depth.

Interesting idea to use market depth to manage collateral requirements, but it's probably hard to measure true market depth without incentivizing makers to place orders to fill up the book.  Anyways I think BM is allowing other collateral to back Privatized Bit Assets and  BitUSD.  Asset/commodity baskets will be much less volatile.  Don't quote me on that though... not sure exactly what's coming...
BitCash - http://www.bitcash.org 
Beta: bitCash Wallet / p2p Gateway: (https://m.bitcash.org)
Beta: bitCash Trade (https://trade.bitcash.org)

Offline xeroc

  • Board Moderator
  • Hero Member
  • *****
  • Posts: 12922
  • ChainSquad GmbH
    • View Profile
    • ChainSquad GmbH
  • BitShares: xeroc
  • GitHub: xeroc
By definition, a margin call forces the purchase of a needed commodity. It's like saying "I need to settle this debt right now, the price has changed too much". If the order is executed by putting a buy option at the current price that does not result in the bitUSD being purchased right away, then you can potentially have an open order indefinitely even past the time when the bitUSD needs to be extinguished. If this is the way this system works, then I think some naming convention is not appropriate.

I might be wrong, so I'd like to learn more on this subject.

Hi ThePiachu,
nice to have you around and any kind of input is appreciated:

First of all:
Margin calls have to buy at market price PLUS up to 10%!!
expired shorts buy up at the market price! you can see this behavior currently at the bitUSD/BTS market

Also, see these threads for bitassets 2.0
https://bitsharestalk.org/index.php/topic,15766.msg202569/topicseen.html#new
 https://bitsharestalk.org/index.php/topic,16143.new.html#new

Offline Permie

  • Hero Member
  • *****
  • Posts: 606
  • BitShares is the mycelium of the financial-earth
    • View Profile
  • BitShares: krimduss
https://bitsharestalk.org/index.php/topic,16498.0.html

Fixed in the next release.

Would be interested to know more how it will be fixed without changing how the system operates...
AFAIK The system is changing.
Extensive documentation to be released in the next few weeks
JonnyBitcoin votes for liquidity and simplicity. Make him your proxy?
BTSDEX.COM

Offline ThePiachu

  • Newbie
  • *
  • Posts: 5
    • View Profile
https://bitsharestalk.org/index.php/topic,16498.0.html

Fixed in the next release.

Would be interested to know more how it will be fixed without changing how the system operates...

Offline ThePiachu

  • Newbie
  • *
  • Posts: 5
    • View Profile
What you claim is that there is some multiplier when the short bitUSD are called and this puts a pressure on the BTS price???

In reality when the shorts are called they do try to buy bitUSD [i.e sell BTS if you wish to call it that], but it:
- happens on the internal market [as opposed on external one, where they would actual drive the price down in comparison with say BTC]
- they use BTS held in collateral for this operation
- and most importantly they do not drive the price down...they just sit there trying to buy bitUSD at the current feed [or 0.9 of it, unfortunately, in the current design].

I really do not see/understand how your multiplier is supposed to work.
Thanks.

By definition, a margin call forces the purchase of a needed commodity. It's like saying "I need to settle this debt right now, the price has changed too much". If the order is executed by putting a buy option at the current price that does not result in the bitUSD being purchased right away, then you can potentially have an open order indefinitely even past the time when the bitUSD needs to be extinguished. If this is the way this system works, then I think some naming convention is not appropriate.

I might be wrong, so I'd like to learn more on this subject.

Offline Rune

  • Hero Member
  • *****
  • Posts: 1120
    • View Profile
This is actually a general attribute of all assets. BitUSD can be considered a debt instrument that provides credit to holders of bitshares (that they then use to buy more bitshares with). If the value of BTS falls, the access to credit of the holders of BTS also goes down. This also happens with all companies that use credit, or the holders of any asset that is used as collateral for loans, so it is a type of speculative attack that already exists. I think the big question is if the liquidity of an asset is high enough to support the amount of it that is used as collateral at any point in time, which is why I think there needs to be "debt ceilings" for collateral types the size of which correlates with their total market depth.

Offline jshow5555

  • Full Member
  • ***
  • Posts: 57
    • View Profile
Recently at Decentral in Vancouver we had a meetup with Max Wright to discuss BitShares and BitUSD:

https://www.youtube.com/watch?v=p023lanFsJE

This inspired me to write a blog about a possible attack on BitUSD and similar derivative currencies:

http://tpbit.blogspot.ca/2015/05/inert-versus-volatile-currencies.html

I am wondering - has there been some analysis done on this subject by the BitShares developers and if so, how do they assess the risk involved with a similar attack?

I wish you explain your attack with a bit more detail.

What you claim is that there is some multiplier when the short bitUSD are called and this puts a pressure on the BTS price???

In reality when the shorts are called they do try to buy bitUSD [i.e sell BTS if you wish to call it that], but it:
- happens on the internal market [as opposed on external one, where they would actual drive the price down in comparison with say BTC]
- they use BTS held in collateral for this operation
- and most importantly they do not drive the price down...they just sit there trying to buy bitUSD at the current feed [or 0.9 of it, unfortunately, in the current design].

I really do not see/understand how your multiplier is supposed to work.
Thanks.

Offline nomoreheroes7

  • Hero Member
  • *****
  • Posts: 756
  • King of all the land
    • View Profile
  • BitShares: nomoreheroes7
https://bitsharestalk.org/index.php/topic,16498.0.html

Fixed in the next release.

Quote
Under normal conditions, the futures market should be stable. You might see some spikes every now and then and perhaps some margin calls being made on an infrequent basis, but the market should track the proper values.

However, it is also possible that a market might have a tipping point - a price at which a snowball effect might take place. Lets say the current market is like this:

    The current price is 100 BTS per BitUSD
    It takes:
        10k BTS to move the price from 100 to 200 BTS per BitUSD
        10k BTS to move the price from 200 to 300 BTS per BitUSD
        10k BTS to move the price from 300 to 400 BTS per BitUSD
    When the price:
        Reaches 200, 5k BTS will be used on margin calls
        Reaches 300, 15k BTS will be used on margin calls
        Reaches 400, 25k BTC will be used on margin calls

If we throw 5k BTS into the market, not much happens - we moved the price a bit. If we throw 10k BTS, we force the market to spend another 5k BTS. We moved the price by 100, and the margin calls moved it further, multiplying the strength of our move by 1.5.
Excerpt from the blog, I hope that's ok

Dan discussed this in the dev hangout yesterday.
From what I understand; in the new system there will be several price feeds that take the liquidity available at a particular price into account.
So if you want to sell a large amount of bitUSD on the market you will not have the same 'snowball effect' power as in the current bitAsset design.
So the 'attack' described against bitUSD in Piachu's blog should be fixed in the new release.

Offline ThePiachu

  • Newbie
  • *
  • Posts: 5
    • View Profile
Recently at Decentral in Vancouver we had a meetup with Max Wright to discuss BitShares and BitUSD:

https://www.youtube.com/watch?v=p023lanFsJE

This inspired me to write a blog about a possible attack on BitUSD and similar derivative currencies:

http://tpbit.blogspot.ca/2015/05/inert-versus-volatile-currencies.html

I am wondering - has there been some analysis done on this subject by the BitShares developers and if so, how do they assess the risk involved with a similar attack?