BitAssets 3.0 - For Community Review

[arhag]

Wait how do you prevent shorts to cover with owned (already) bitAsset. By completely removing such operation?

Well they can still cover with already owned BitAssets, but then they wouldn't really be short. I am assuming they don't own enough of the BitAsset to cover their existing short (so that they profit from BTS price increases). Then to rollover they would need to use their spare BTS to buy BitUSD to cover the existing short and to buy into their new short order. However these BitUSD buy orders would first match with the activated redeem orders (if any exist) before matching with any other sell (or short) orders.

[arhag]

.

Black swan liquidation rules still exist and are triggered when the collateral ratio of any short position falls below 100%.

.

Arhag, falls below 100% during redemption process?

Can't we see this situation from an attacking whale who attempts to swing the internal price dramatically downward seconds before his bts is reclaimed at a higher price feed?

The black swan liquidation is independent of the redemption process and already (in theory) exists with the current BitAsset system. The risks associated with that are price feed manipulation or market manipulation to trigger unfair black swan liquidation to the advantage of BitAsset holders and at the expense of the shorts. These are the same risks that exist with the current system because it is the same exact mechanism. The shorts holding BTS as collateral would never benefit from falsely triggering black swan liquidation because the BTS price cannot falsely be higher than its true price and still trigger an undercollateralization event that wouldn't have been triggered with the true price anyway.  So in this case, the attacking whale would need to be a large BitUSD holder and would then want to manipulate the price feed (perhaps by dumping some BTS in thin markets and hoping the delegates update corresponding prices to trigger undercollateralization). But I think this would be a difficult attack to profitably pull off.

[clayop]

Too high-level conversation for me.    I may need Max's kind explanation. Will "BitAssets 101" be published? 

[zerosum]

Can you summarize for the less smart of us what your model achieves and or how it differs from the original proposal 3.0?

Mostly the same objectives as the ones listed in bytemaster's post. I have added some detail for a possible implementation. However there are a few notable differences.

First, I reintroduce a grace period where the shorts cannot be called. This is similar to our current system of expiration. I'm thinking it would be useful for shorts to have 2 weeks where they do not have to worry about being forced out of their position (assuming they maintain enough collateral to not get margin called). It is only after that grace period that they need to worry about their ranking in the regular call queue and keep track of any pending unactivated redeem orders they may potentially force call their short position in a day or two.

Since, the activated redeem queues have priority over any other order (especially shorts), there is no way for the existing shorts to roll over to a new short without first satisfying the activated redeem order. So it may, in the worst case, delay settlement of the redeem order from 2 days to 2 weeks, but I think that isn't a big deal for the long. I for one would find a 2 week grace period incredibly useful since I have often "purchased" BTS at (what I thought was) a nice price (via short) and then took my time to settle the payment with new BitUSD purchased using new Bitcoin (and that has delays associated with ACH transfers if you don't already have fiat stored with a Bitcoin exchange).

Wait how do you prevent shorts to cover with owned (already) bitAsset. By completely removing such operation?

There are two other major differences that have to do with collateral requirements and margin calls. I would not force any particular initial collateral ratio, such as 200%. Instead the margin call ratio limit would be defined as part of the BitAsset (and if we have "privatized" BitAssets then each creator can specify their own), and the blockchain only requires the initial collateral ratio to not be less than the margin call ratio limit defined for the BitAsset. Other than that restriction, the short seller has freedom to specify the initial collateral ratio depending on how much margin call risk they are willing to tolerate.

As I have said before I like this idea of yours a lot. Maybe it is a good time to really implement it, with the rest of all those changes.

Also, even when a short position is margin called, it does not automatically buy up BitUSD up to a price of 10% above the feed price in my proposed tweaks. Instead the price limit it is willing to cover at grows monotonically from the feed price all the way to 20% above the feed price as the collateral ratio continues to drop lower and closer to 100%.

I do hope we get rid of this 'up to 10%' penalty, altogether.

[Ander]

I really like these changes!  Well done all those who contributed to this idea!!


Improvements:
* No forced cover at any particular amount of time.  (30 day rule is gone).  The least collateralized short may be "margin called" with 24 hour notice by the bitAsset longs, but they have to pay him 1% to do so at least.

* bitAsset holders can be guaranteed a return with only 1% loss in 24 hours, which is much better than up to 30 days! 

* Probably a better peg!

* Collateral is much more fair, even colalteral between shorts and longs, instead of favoring bitAsset longs.




We are going through a process of testing and improving the bitshares product.  This iteration will be superior to the previous one.

[Bitcoinfan]

.

Black swan liquidation rules still exist and are triggered when the collateral ratio of any short position falls below 100%.

.

Arhag, falls below 100% during redemption process?

Can't we see this situation from an attacking whale who attempts to swing the internal price dramatically downward seconds before his bts is reclaimed at a higher price feed?


Sent from my iPhone using Tapatalk

[zerosum]

That said a system that reduces restrictions while improving liquidity (no more logjams ) is all good in my book.

Aside from that there is a very important detail not mentioned by anybody up to now (I do not know if the 3.0 designer did it intentionally or it is side effect) BUT IMHO this 3.0 things has a very good chance to also improve the peg (remember a few months back when for days and weeks bitUSD was selling for 0.98...0.97 or less?)

For the first time the bitAsset holders will have the actual power to sell at about the feed price. It is significant (in my book) difference from the current model where the shorts have to cover at the peg and only as a consequence of that the longs can sell at the peg (in 30 days or less).

[arhag]

Can you summarize for the less smart of us what your model achieves and or how it differs from the original proposal 3.0?

Mostly the same objectives as the ones listed in bytemaster's post. I have added some detail for a possible implementation. However there are a few notable differences.

First, I reintroduce a grace period where the shorts cannot be called. This is similar to our current system of expiration. I'm thinking it would be useful for shorts to have 2 weeks where they do not have to worry about being forced out of their position (assuming they maintain enough collateral to not get margin called). It is only after that grace period that they need to worry about their ranking in the regular call queue and keep track of any pending unactivated redeem orders they may potentially force call their short position in a day or two.

Since, the activated redeem queues have priority over any other order (especially shorts), there is no way for the existing shorts to roll over to a new short without first satisfying the activated redeem order. So it may, in the worst case, delay settlement of the redeem order from 2 days to 2 weeks, but I think that isn't a big deal for the long. I for one would find a 2 week grace period incredibly useful since I have often "purchased" BTS at (what I thought was) a nice price (via short) and then took my time to settle the payment with new BitUSD purchased using new Bitcoin (and that has delays associated with ACH transfers if you don't already have fiat stored with a Bitcoin exchange).

There are two other major differences that have to do with collateral requirements and margin calls. I would not force any particular initial collateral ratio, such as 200%. Instead the margin call ratio limit would be defined as part of the BitAsset (and if we have "privatized" BitAssets then each creator can specify their own), and the blockchain only requires the initial collateral ratio to not be less than the margin call ratio limit defined for the BitAsset. Other than that restriction, the short seller has freedom to specify the initial collateral ratio depending on how much margin call risk they are willing to tolerate. Also, even when a short position is margin called, it does not automatically buy up BitUSD up to a price of 10% above the feed price in my proposed tweaks. Instead the price limit it is willing to cover at grows monotonically from the feed price all the way to 20% above the feed price as the collateral ratio continues to drop lower and closer to 100%.

[Riverhead]

I like this a LOT better than proposal 2.0. This gets us back to one bitAsset per currency and frees up the market. I'll comment in the 2.0 thread why that model gives me the heebeejeebees.

However, can we first see how the market works once more bugs are worked out? While the recent bitUSD issue was a confluence of events the short matching bug prevented us from seeing how the market may have worked itself out.

That said a system that reduces restrictions while improving liquidity (no more logjams ) is all good in my book.

[vlight]

So as i understand this is the system where the price feeds are only forced onto participants if they 'misbehave' and deviate from the fair price too much. If everyone play nice there will be no need for the blockchain to police the market.

 

[oldman]

Its fine if yield goes away.  It can be re-introduced with loans and leveraging like in Bitfinex.  I'm sure that's the aim.  And yield in this matter would be much larger.

exactly

Loans, leverage and credit creation are the keys to the rocket ship.

When integrated into the BTS platform there will be literally nothing on the planet that can compete... and no more powerful force for social reform.

Decentralized finance on a global scale.

[zerosum]

I have a few proposed tweaks.

First I will make a note of five main queues:

• Regular call queue
• Margin call queue
• Activated redeem queue
• Sell queue
• Buy queue

A short position has a grace period G (e.g. G = 14 days) during which it is not callable unless margin called. If the collateral ratio of a short position is below the margin call ratio limit M (e.g. M = 175% = 1.75), it will be placed in the margin call queue. If a short position does not belong in the margin call queue but it has existed for longer than the grace period, it is placed in the regular call queue. Both the regular call queue and margin call queue are ordered in ascending collateral ratio order (meaning the head of each queue is the short position that has a lower collateral ratio than any other short position in the same queue).

The buy queue contains all the orders that want to buy the BitAsset by selling its corresponding collateral asset (e.g. BTS). This is actually a virtual queue that encompasses all the other real buy queues such as the ones for absolute buy orders and relative (to the price feed) buy orders. The sell queue contains all the orders that want to sell the BitAsset for the corresponding collateral asset. It too is actually a virtual queue that encompasses all the other real sell queues such as the ones for absolute sell orders, relative sell orders, absolute short orders, relative short orders, and unactivated redeem orders.

An unactivated redeem order is very similar to a relative sell order except that its offset is restricted to be no greater than -1% (meaning offering the order at a price no better than 99% of the price feed). Once an unactivated redeem order has existed continuously for an X period (e.g. X = 48 hours), it is promoted to an activated redeem order.

The buy queue is in descending price order. The sell queue and activated redeem queue are in ascending price order.

Matching rules are as follows:

• If the activated redeem queue is not empty, first attempt matching the head of the activated redeem queue with the order with the highest bid from the set of orders composed from the head of the buy queue and margin call queue (all margin call queue items offer a limit bid price up to Y fraction above the feed price, where Y is a monotonic function of the collateral ratio, e.g. Y = 20% * (1 - (COLLATERAL_RATIO - 1)/(M-1))^2 ). If there was a (partial) match, repeat with the new set of heads until there is no match anymore. If the activated redeem queue is still not empty at that point (which, by the way, indicates that the margin call queue must be empty), match the head of the activate redeem queue with the head of the regular call queue (this will guarantee a match as long as the regular call queue is not empty) until either the activated redeem queue is empty or the regular call queue is empty.
• Attempt matching the order with the highest bid from the set of orders composed from the head of the buy queue and margin call queue with the head of the sell queue. Repeat until no more matches are possible.

A short order has no price restriction. However, the short order has to provide the initial collateral ratio it wants to match with and the blockchain requires that this ratio is not less than the margin call ratio limit imposed by the BitAsset (a reasonable value for the margin call ratio limit, for now, might be something between 150% and 180%). So if the margin call ratio limit for the BitAsset was set to 175%, the short order could specify an initial collateral ratio of 175%, but then the risk of margin call would be insanely high, so perhaps the short seller would choose a more sensible initial collateral ratio of 190%.

Black swan liquidation rules still exist and are triggered when the collateral ratio of any short position falls below 100%.

Can you summarize for the less smart of us what your model achieves and or how it differs from the original proposal 3.0?

As I read the proposal 3.0 - the (forced) settlement process would be completely independent of the market matching engine.
 And although the OP does not talk about margin calls at all, I assume they would be simple buy (the bitAsset) @ feed price orders with no special rules other than they will change its price with the feed price changes.

[bytemaster]

1) All short positions can "vote" and if more than 51% of the shorts think the market has gone bad then the forced liquidation of the long side can be initiated with a 10% premium to the USD positions.   51% of the shorts would equal 25% of the stakeholders in the asset and if they are all willing to take a 10% hit then chances are the long positions are being unreasonable.   

2) Allow the price feed producers to trigger it any time they feel it makes sense (with 30 day notice).   These producers are the "guardians of the feed" and the peg.

3) Require both the feed producers AND the shorts to agree.

[arhag]

I have a few proposed tweaks.

First I will make a note of five main queues:

• Regular call queue
• Margin call queue
• Activated redeem queue
• Sell queue
• Buy queue

A short position has a grace period G (e.g. G = 14 days) during which it is not callable unless margin called. If the collateral ratio of a short position is below the margin call ratio limit M (e.g. M = 175% = 1.75), it will be placed in the margin call queue. If a short position does not belong in the margin call queue but it has existed for longer than the grace period, it is placed in the regular call queue. Both the regular call queue and margin call queue are ordered in ascending collateral ratio order (meaning the head of each queue is the short position that has a lower collateral ratio than any other short position in the same queue).

The buy queue contains all the orders that want to buy the BitAsset by selling its corresponding collateral asset (e.g. BTS). This is actually a virtual queue that encompasses all the other real buy queues such as the ones for absolute buy orders and relative (to the price feed) buy orders. The sell queue contains all the orders that want to sell the BitAsset for the corresponding collateral asset. It too is actually a virtual queue that encompasses all the other real sell queues such as the ones for absolute sell orders, relative sell orders, absolute short orders, relative short orders, and unactivated redeem orders.

An unactivated redeem order is very similar to a relative sell order except that its offset is restricted to be no greater than -1% (meaning offering the order at a price no better than 99% of the price feed). Once an unactivated redeem order has existed continuously for an X period (e.g. X = 48 hours), it is promoted to an activated redeem order.

The buy queue is in descending price order. The sell queue and activated redeem queue are in ascending price order.

Matching rules are as follows:

• If the activated redeem queue is not empty, first attempt matching the head of the activated redeem queue with the order with the highest bid from the set of orders composed from the head of the buy queue and margin call queue (all margin call queue items offer a limit bid price up to Y fraction above the feed price, where Y is a monotonic function of the collateral ratio, e.g. Y = 20% * (1 - (COLLATERAL_RATIO - 1)/(M-1))^2 ). If there was a (partial) match, repeat with the new set of heads until there is no match anymore. If the activated redeem queue is still not empty at that point (which, by the way, indicates that the margin call queue must be empty), match the head of the activate redeem queue with the head of the regular call queue (this will guarantee a match as long as the regular call queue is not empty) until either the activated redeem queue is empty or the regular call queue is empty.
• Attempt matching the order with the highest bid from the set of orders composed from the head of the buy queue and margin call queue with the head of the sell queue. Repeat until no more matches are possible.

A short order has no price restriction. However, the short order has to provide the initial collateral ratio it wants to match with and the blockchain requires that this ratio is not less than the margin call ratio limit imposed by the BitAsset (a reasonable value for the margin call ratio limit, for now, might be something between 150% and 180%). So if the margin call ratio limit for the BitAsset was set to 175%, the short order could specify an initial collateral ratio of 175%, but then the risk of margin call would be insanely high, so perhaps the short seller would choose a more sensible initial collateral ratio of 190%.

Black swan liquidation rules still exist and are triggered when the collateral ratio of any short position falls below 100%.

Finally, the mechanism that triggers forced settlement of all longs with shorts needs to be made more concrete. How exactly do we determine whether BitUSD asks are not fair and also not fair for a long enough period of time (30 days?) to justify triggering the forced settlement? I'm still thinking about the details of this one, but would love to hear what you suggest.

[bytemaster]

Its fine if yield goes away.  It can be re-introduced with loans and leveraging like in Bitfinex.  I'm sure that's the aim.  And yield in this matter would be much larger.

exactly