Author Topic: Discussing the problems with bitUSD (smart coins)  (Read 14935 times)

0 Members and 1 Guest are viewing this topic.

Offline Helikopterben

  • Sr. Member
  • ****
  • Posts: 202
    • View Profile
Options buyers is traditional markets pay for premium only.  That is why we need some form of leveraged trading, if possible.

Offline tonyk

  • Hero Member
  • *****
  • Posts: 3308
    • View Profile
An option with a 10 year expiration is close enough to infinite. The further in the future you go the smaller the difference in price.

So do your estimation based upon the curve generated by 3 month, 6 month, and 12 month expirations.

Do you have any idea how much an option with 3 y expiration and the volatility of BTS will cost?

I bet it will be hitting 1usd per 1usd if not more.

I actually think that the option is not infinite in term because the option doesn't have a fixed price.  So I think that is the wrong algorithm to use.

When I go short, I am pricing the cost of finding someone else to take over my short position in the future relative to the BitUSD holder wishing to exercise.

With liquidity and no fixed price, the premium can be much lower.

1 bitCNY - is 1) 'smart instrument' and 2)the right to sell this instrument for a price of 1 CNY at any point in the future.

So this is definitely some instrument[we are not concerned with exactly here]  plus  a put option with strike of 1 CNY [ aka giving you the right to sell said instrument for 1 CNY = straight definition of an option]...except it has no expiration date.

When you are buying the bitCNY you are paying one CNY for the underlying contract (or smart instrument) and anything above 1 CNY is the premium for the put option.



And yes any option definitely has a price (theoretical value) at any point in time.

I think the place you went wrong is saying you have the right to sell said instrument for 1 CNY... which is no true,  you have the right to sell it for 1 CNY worth of BTS. 

So the person selling the BitCNY into circulation simply has a callable (on-demand) loan.  You borrow money that is callable on demand (24 hour notice)... considering the loan is interest-free itself we can consider the premium on BitCNY to be  PRICE_OF_OPTION - PREPAID_INTEREST where PREPAID_INTEREST is a variable that represents the value gained by the short and is proportional to the expected change in price of the collateral.   

There are a lot of variables in play, but I do not think it is fair to say the premium is just the cost of the option.  It factors in the cost of capital for collateral required to stay in the top 98% where you do not have to provide the option.

In other words the bottom 2% of least-collateralized shorts must assume the cost of the option.  The other 98% don't have to provide the option.  So the premium is pro-rated between these two costs.

I do not know if from this statement of yours is possible to directly reach the same conclusions [at least I did not] but it lead me to realize I am indeed wrong.

The reason is the contract [bitAsset short sell and premium] let's you actively adjust the  max loss [ which is a clear benefit and possible with each price movement in favour of the shorter] . Benefit which the option seller do not have* and which leads to different risk and consecutively price profiles (calculation).

*To simulate the behavior - At that very point it becomes the difference between the prices of 2 options with 2 strike prices - the original BTS price at the short sell and the current BTS price. [ if someone cares to know]
Lack of arbitrage is the problem, isn't it. And this 'should' solves it.

Offline bytemaster

An option with a 10 year expiration is close enough to infinite. The further in the future you go the smaller the difference in price.

So do your estimation based upon the curve generated by 3 month, 6 month, and 12 month expirations.

Do you have any idea how much an option with 3 y expiration and the volatility of BTS will cost?

I bet it will be hitting 1usd per 1usd if not more.

I actually think that the option is not infinite in term because the option doesn't have a fixed price.  So I think that is the wrong algorithm to use.

When I go short, I am pricing the cost of finding someone else to take over my short position in the future relative to the BitUSD holder wishing to exercise.

With liquidity and no fixed price, the premium can be much lower.

1 bitCNY - is 1) 'smart instrument' and 2)the right to sell this instrument for a price of 1 CNY at any point in the future.

So this is definitely some instrument[we are not concerned with exactly here]  plus  a put option with strike of 1 CNY [ aka giving you the right to sell said instrument for 1 CNY = straight definition of an option]...except it has no expiration date.

When you are buying the bitCNY you are paying one CNY for the underlying contract (or smart instrument) and anything above 1 CNY is the premium for the put option.



And yes any option definitely has a price (theoretical value) at any point in time.

I think the place you went wrong is saying you have the right to sell said instrument for 1 CNY... which is no true,  you have the right to sell it for 1 CNY worth of BTS. 

So the person selling the BitCNY into circulation simply has a callable (on-demand) loan.  You borrow money that is callable on demand (24 hour notice)... considering the loan is interest-free itself we can consider the premium on BitCNY to be  PRICE_OF_OPTION - PREPAID_INTEREST where PREPAID_INTEREST is a variable that represents the value gained by the short and is proportional to the expected change in price of the collateral.   

There are a lot of variables in play, but I do not think it is fair to say the premium is just the cost of the option.  It factors in the cost of capital for collateral required to stay in the top 98% where you do not have to provide the option.

In other words the bottom 2% of least-collateralized shorts must assume the cost of the option.  The other 98% don't have to provide the option.  So the premium is pro-rated between these two costs.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline tonyk

  • Hero Member
  • *****
  • Posts: 3308
    • View Profile
An option with a 10 year expiration is close enough to infinite. The further in the future you go the smaller the difference in price.

So do your estimation based upon the curve generated by 3 month, 6 month, and 12 month expirations.

Do you have any idea how much an option with 3 y expiration and the volatility of BTS will cost?

I bet it will be hitting 1usd per 1usd if not more.

I actually think that the option is not infinite in term because the option doesn't have a fixed price.  So I think that is the wrong algorithm to use.

When I go short, I am pricing the cost of finding someone else to take over my short position in the future relative to the BitUSD holder wishing to exercise.

With liquidity and no fixed price, the premium can be much lower.

1 bitCNY - is 1) 'smart instrument' and 2)the right to sell this instrument for a price of 1 CNY at any point in the future.

So this is definitely some instrument[we are not concerned with exactly here]  plus  a put option with strike of 1 CNY [ aka giving you the right to sell said instrument for 1 CNY = straight definition of an option]...except it has no expiration date.

When you are buying the bitCNY you are paying one CNY for the underlying contract (or smart instrument) and anything above 1 CNY is the premium for the put option.



And yes any option definitely has a price (theoretical value) at any point in time.
« Last Edit: December 02, 2015, 01:09:14 am by tonyk »
Lack of arbitrage is the problem, isn't it. And this 'should' solves it.

Offline bytemaster

An option with a 10 year expiration is close enough to infinite. The further in the future you go the smaller the difference in price.

So do your estimation based upon the curve generated by 3 month, 6 month, and 12 month expirations.

Do you have any idea how much an option with 3 y expiration and the volatility of BTS will cost?

I bet it will be hitting 1usd per 1usd if not more.

I actually think that the option is not infinite in term because the option doesn't have a fixed price.  So I think that is the wrong algorithm to use.

When I go short, I am pricing the cost of finding someone else to take over my short position in the future relative to the BitUSD holder wishing to exercise.

With liquidity and no fixed price, the premium can be much lower.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline tonyk

  • Hero Member
  • *****
  • Posts: 3308
    • View Profile
An option with a 10 year expiration is close enough to infinite. The further in the future you go the smaller the difference in price.

So do your estimation based upon the curve generated by 3 month, 6 month, and 12 month expirations.

Do you have any idea how much an option with 3 y expiration and the volatility of BTS will cost?

I bet it will be hitting 1usd per 1usd if not more.
Lack of arbitrage is the problem, isn't it. And this 'should' solves it.

Offline bytemaster

An option with a 10 year expiration is close enough to infinite. The further in the future you go the smaller the difference in price.

So do your estimation based upon the curve generated by 3 month, 6 month, and 12 month expirations.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline tonyk

  • Hero Member
  • *****
  • Posts: 3308
    • View Profile
Problem:
Mathematically, the value of the premium shorts sell to longs is a function of expected future variance in the value of that same premium - which makes it very difficult to calculate. However, markets do this all the time - it's basically pricing the derivative of a derivative. It's like trading options of VIX.
http://sixfigureinvesting.com/2010/01/trading-vix-options/

actually it is trading options with no expiration date on VIX...which makes the task...close to impossible.

otherwise great post + 1

No expiration?
Quote
7. Expiring In-the-Money VIX options give a cash payout.  The payout is determined by the difference between the strike price and the VRO quotation on the expiration day.  For example the payout would be $1.42 if the strike price of your call option was $15 and the VRO was $16.42.
8. The expiration or “print” amount when VIX options expire is given under the ^VRO symbol (Yahoo) or $VRO (Schwab).   This is the expiration value, not the opening cash VIX on the Wednesday morning of expiration.  VIX options expire at market open on expiration day, so they are not tradeable on that day.
9. VIX options do not expire on the same days as equity options. It is almost always on a Wednesday  See this post for upcoming expirations.  This odd timing is driven by the needs of a straightforward settlement process.  On the expiration Wednesday the only SPX options used in the VIX calculation are the ones that expire in exactly 30 days.  For more on this process see Calculating the VIX—the easy part.

no expiration in the BTS case. so the sentence should read:

However, markets do this all the time - it's basically pricing the derivative of a derivative. It's like trading options of VIX.  However in the BTS case those are options with no expiration date....which makes the task...close to impossible.

« Last Edit: December 01, 2015, 06:43:45 pm by tonyk »
Lack of arbitrage is the problem, isn't it. And this 'should' solves it.

Offline abit

  • Committee member
  • Hero Member
  • *
  • Posts: 4632
    • View Profile
    • Abit's Hive Blog
  • BitShares: abit
  • GitHub: abitmore
Problem:
Mathematically, the value of the premium shorts sell to longs is a function of expected future variance in the value of that same premium - which makes it very difficult to calculate. However, markets do this all the time - it's basically pricing the derivative of a derivative. It's like trading options of VIX.
http://sixfigureinvesting.com/2010/01/trading-vix-options/

actually it is trading options with no expiration date on VIX...which makes the task...close to impossible.

otherwise great post + 1

No expiration?
Quote
7. Expiring In-the-Money VIX options give a cash payout.  The payout is determined by the difference between the strike price and the VRO quotation on the expiration day.  For example the payout would be $1.42 if the strike price of your call option was $15 and the VRO was $16.42.
8. The expiration or “print” amount when VIX options expire is given under the ^VRO symbol (Yahoo) or $VRO (Schwab).   This is the expiration value, not the opening cash VIX on the Wednesday morning of expiration.  VIX options expire at market open on expiration day, so they are not tradeable on that day.
9. VIX options do not expire on the same days as equity options. It is almost always on a Wednesday  See this post for upcoming expirations.  This odd timing is driven by the needs of a straightforward settlement process.  On the expiration Wednesday the only SPX options used in the VIX calculation are the ones that expire in exactly 30 days.  For more on this process see Calculating the VIX—the easy part.
BitShares committee member: abit
BitShares witness: in.abit

Offline abit

  • Committee member
  • Hero Member
  • *
  • Posts: 4632
    • View Profile
    • Abit's Hive Blog
  • BitShares: abit
  • GitHub: abitmore

I agree with your definition of the requirements, but the conclusion is just the current description of how bitshares works...and I'm not sure this is the only possible way it *could* work.

We only need to know this way *could* or *could not* work. If we already knew it could not work, of course we'll try other ways. We've already run on this way so far, it costs a lot to turn to another way.
BitShares committee member: abit
BitShares witness: in.abit

Offline tonyk

  • Hero Member
  • *****
  • Posts: 3308
    • View Profile
Problem:
Mathematically, the value of the premium shorts sell to longs is a function of expected future variance in the value of that same premium - which makes it very difficult to calculate. However, markets do this all the time - it's basically pricing the derivative of a derivative. It's like trading options of VIX.
http://sixfigureinvesting.com/2010/01/trading-vix-options/

actually it is trading options with no expiration date on VIX...which makes the task...close to impossible.

otherwise great post + 1
« Last Edit: December 01, 2015, 06:09:07 pm by tonyk »
Lack of arbitrage is the problem, isn't it. And this 'should' solves it.

Offline monsterer

General rules:
1) When in oversupply, destruction (settling) should be incentivized
2) When in undersupply, creation (shorting) should be incentivized


I agree with your definition of the requirements, but the conclusion is just the current description of how bitshares works...and I'm not sure this is the only possible way it *could* work.

For example, starspirit demonstrated another model whereby supply and demand is controlled based on yield, where yield can be negative, such that holding bitUSD under a negative yield would naturally cause rational participants to reduce the overall supply of bitUSD, and likewise with an undersupply of bitUSD, a positive yield would have the opposite effect.
My opinions do not represent those of metaexchange unless explicitly stated.
https://metaexchange.info | Bitcoin<->Altcoin exchange | Instant | Safe | Low spreads

Offline maqifrnswa

  • Hero Member
  • *****
  • Posts: 661
    • View Profile
Q) Is there another design which doesn't have the same biased risk profile, or is this just a natural consequence of not having redeemability?

Mathematically, I believe it is a natural consequence. Here's the argument:

From game theory, there are two possible moves:
1) Create smartcoin (short)
2) Destroy smartcoin (settle)

We need to make a game such that the rules of the game encourage creation and destruction such that value is maintained.

Now, like all currencies, you can choose how you want to maintain value. You can have a centralized bank/group of people controlling supply (Federal Reserve), or you can have a fix value as currencies that are on the "Gold Standard." For the sake of merchants and liquidity, BitShares has made the decision that 1 smartcoin should always be redeemable for 1 underlying asset of BTS. That is, BitShares is on the "Gold Standard." (well, technically, bitGOLD is on the "gold standard," but you get what I mean  ;D ) This also means that a centralized group of people are NOT needed, and thus you can rely on decentralized data (feed) to maintain your gold standard (versus the federal reserve)

So now that BitShares is on the Gold Standard, what should the rules be such that creation and destruction incentivize the gold standard?

General rules:
1) When in oversupply, destruction (settling) should be incentivized
2) When in undersupply, creation (shorting) should be incentivized


We need to come up with a system, on a blockchain, that does that. There are really one two rules that can possible be implemented. However, the optimum implementation parameters are unknown and are thus parametrized in BitShares. However, these specific rules MUST be followed, as there really is no other way to achieve our goal given the boundary conditions.

Destruction rule:
1) At any time you are allowed to destroy a smartcoin by exchanging 1 smartcoin for its value in underlying asset

BitShares Implementation: Since underlying value asset is not determined within our system, we need to pull it from external sources that actually trade real versions of the underlying asset. Thus, settlement price should be a function of external price feed.
Arbitrage Profit for enforcing the rule: Finite and the difference between what you can buy the asset for now and what you can settle it for.
Long term profit: None if performed via arbitrage.
Risk to settler: None, besides change in feed between settlement request and execution (see below)

To reduce risk of manipulation via a "sneak attack" on the market: time delay between declaring a settle and settle
To reduce the risk of manipulation by "buying out" the entire market: settlement limits volume to a % of the total volume


Creation rule:
1) You can create (short) at any time to yourself

Implementation: Shorting to yourself at any time is allowed as long as you maintain collateral so that the smartcoin you create is "backed" by something of value
Arbitrage Profit: None
Long term Profit: Infinite and based on the difference in growth rate between smartcoin and BTS.
Risk: uncertainty in the future value of smartcoin relative BTS.

Outcome of rules:
Settlers (longs) have zero risk; shorts have non-zero risk. Therefore, shorts should sell to longs at a premium. The price of this premium is a function of variance in the expected BTS growth rate relative to the underlying asset as well as the variance in the premium itself.

There is one issue I've been struggling with for months:
Problem:
Mathematically, the value of the premium shorts sell to longs is a function of expected future variance in the value of that same premium - which makes it very difficult to calculate. However, markets do this all the time - it's basically pricing the derivative of a derivative. It's like trading options of VIX.
http://sixfigureinvesting.com/2010/01/trading-vix-options/

Conclusion:
Given the boundary condition of "BitShares wants to be on the Gold Standard," it directly follows that 1 smartcoin > underlying asset and forced settlement must exist.
« Last Edit: December 01, 2015, 04:12:49 pm by maqifrnswa »
maintains an Ubuntu PPA: https://launchpad.net/~showard314/+archive/ubuntu/bitshares [15% delegate] wallet_account_set_approval maqifrnswa true [50% delegate] wallet_account_set_approval delegate1.maqifrnswa true

Offline Samupaha

  • Sr. Member
  • ****
  • Posts: 479
    • View Profile
  • BitShares: samupaha
The problems as I see them:
  • BTS price continues to fall. The declining value of BTS cannot support the creation and maintenance of bitassets
  • The market mechanics are not well documented and over complicated for even the most advanced traders
  • The market mechanics are skewed towards bitasset holders instead of shorts. This innevitably leads to less liquidity and inefficient markets
  • As a result of number 3, it is far less risky and more profitable to margin trade on centralized exchanges than on Bitshares. Why would you short BTC releative to BTS on Bitshares when you can do so on Poloniex while incurring less of a cost
  • Market making for bitasset markets is not profitable and far too risky

I'm sure there are more problems that I have not enumerated. I have some idea of how to solve them, but I don't see the BM/community taking the necessary actions to fix these problems. We are wasting time adding features like stealth transfers to the GUI instead of making our core product actually work. I'm at my wits end and I've been here since the beginning.

Good list, I agree with all points.

It seems that MPA is really difficult subject and nobody knows how we can tweak the parameters to work so that they will provide good incentives in all situations. Maybe it would be best if we just set them aside for a while? Maybe it would be best to focus on products that are easily made useful for users, like stealth transfers, and try to get more users with them?

Its important to remember that much progress has been made and the building blocks are in place:  DPOS, basic market infrastructure, a (mostly polished) user-friendly web interface, ect.  We have to remember that this is programmable finance.  What does not work now can always be changed.  Eventually we will find the right formula.  I think there is a good chance we will see some solid adoption with the implementation of a bond/swap market.  Also, I have stated before that Smartcoins for physical commodities may be the ticket to real user adoption.  No other project can compete with that at this point in time. 

Yeah, most important thing here is the DAC platform. If one thing doesn't work we can either tweak it better or try something else. This is trial and error. If we do it enough times, eventually we'll come up with products that work.

Offline Helikopterben

  • Sr. Member
  • ****
  • Posts: 202
    • View Profile
The biggest difference will most likely come with the bond/swap market, essentially transferring risk from the lender to the borrower.  With a bond/swap market, the bts holders should be able to create smartcoins and lend them out to earn a nearly risk-free return instead of selling the smartcoin and taking on the added exchange rate risk.  The smartcoin creator's debt/collateral ratio would not change during the course of the loan.  When the loan is paid back, the smartcoin creator could simply adjust the debt and collateral back to 0 and collect the interest owed.  All the smartcoin creator would have to do is maintain sufficient collateral over the course of the loan and he would be in no danger of losing bts, essentially giving him a nearly risk-free return on his bts.  This makes sense as bts is the reserve currency of the system.  This could also potentially put upward pressure on prices as users realize they can get a nearly risk-free, market-driven return on their bts. 

Forced settlement is necessary as there are no time constraints to force settlement like there are in futures and options markets, although the parameters of forced settlement could be tweaked.  Maybe require a 7-day waiting period before the transaction executes and maybe execute at 98% or 95% of feed.  I never really thought guaranteeing settlement on a 1:1 basis was a great idea.  Merchant adoption won't happen unless we have fairly liquid markets and if we have sufficiently liquid markets, then redeeming value near the feed shouldn't be a problem.  Markets need to be liquid first.  They will become liquid if incentives are aligned.  SQP could possibly be adjusted to within 10% of the price feed, but this parameter is only relevant in illiquid markets. 

Its important to remember that much progress has been made and the building blocks are in place:  DPOS, basic market infrastructure, a (mostly polished) user-friendly web interface, ect.  We have to remember that this is programmable finance.  What does not work now can always be changed.  Eventually we will find the right formula.  I think there is a good chance we will see some solid adoption with the implementation of a bond/swap market.  Also, I have stated before that Smartcoins for physical commodities may be the ticket to real user adoption.  No other project can compete with that at this point in time.