Like you said, BM discussed this briefly in the latest dev hangout but only mentioned that there could be several price feeds different sized orders

OK, I didn't actually know where it had been raised or that the idea might be similar, apart from some unclear references in the forum about a change coming to price feeds

Does your suggested method rely on price feeds from external USD/BTS exchanges?

Ideally yes, although if it doesn't exist, any pathway from BTS to USD is sufficient to infer the purchase cost. e.g. BTS ->BTC->USD.

Quote

If the manipulator had the power to spark a self-fulfilling cycle of panicked selling from other market participants that pushed the price down even further, then they would make more profit waiting to get hit on lower bids than to settle bitUSD at the asks, which is effectively like crossing the spread a second time. In practice though, I would still probably have a 24 hour non-cancellable delay on settlements to ensure minimum opportunity to take advantage of short term market aberrations not reflected in the current scale.

Could you explain this further? Specifically "make more profit waiting to get hit on lower bids than to settle bitUSD at the asks"

In this case, the manipulator has a choice to sit at the bid and absorb the selling behaviour he has successfully instigated, or he can settle the bitUSD which will price the BTS based off the mid-price plus a liquidity premium related to expected depth in the ask. The latter is a more expensive price.

On the surface it sounds like BM's idea may be looking to accomplish something similar to what I suggested here, so maybe I'll wait to see how he describes that.

[Edit: I need to think more on this idea, not sure if it is actually a good idea or not yet]

We know that increasing the daily settlement limit will increase the premium, and decreasing it will decrease the premium. 

On the idea of using the settlement fee as a lever to manage the bitUSD:USD premium, I can't see why the market would price this in. If the market is at a premium, then the settlement mechanism is not needed. By the time the settlement mechanism is needed and used, the price must be at a discount where there is no longer a settlement fee. So there is no need for any settlement fee to ever be paid, nor any buyer to price in the risk of a higher fee, no matter how high the premium. So how would this affect supply or demand?

So lets just assume the market settles on $1.30  as the price for creating BitUSD with forced settlement at the feed which is $1.00.   Who cares so long as the premium is relatively stable?    The only thing that can move the premium is market forces based upon volume and market direction.   

I would care as a consumer because I would be better off just using real USD to make my purchase rather than using bitUSD which costs so much more to obtain. Now if merchants were willing to fully price in the higher value of a bitUSD into their product price, rather than just using face value, then that would adequately compensate. But then that forces merchants and consumers to keep constant track of the premium and adjust expectations accordingly, as well as accepting higher downside risk, and I don't know how we'd get everyone comfortable that the premium will be stable. Besides, it doesn't really appear true to label from a consumer/merchant angle.

I'm not underestimating the problem of dealing with premiums (or pegging in general), but I am still optimistic there is an improved solution...

Quote from: eagleeye on May 22, 2015, 04:13:35 am

Quote from: Helikopterben on May 22, 2015, 03:44:10 am

Quote from: bytemaster on May 21, 2015, 09:55:24 pm

Bottom line, when you look at the BitUSD price, the quantity available over $1.00 is VERY SMALL and thus meaningless.   That means that right now BitUSD is "perfectly pegged" and the premium for creating new BitUSD is VERY HIGH.   This is a good and healthy position and indicates what it will look like with force settlement enabled.    Everything else just depends upon trading volume and market making.   

Right now there is over 70,000 bitusd in bids at the feed price, which represents almost half of all bitusd in existence.  If this were force settled right now at the feed price and all else remaining constant, then nearly half of all short sellers would be forced into a transaction that they did not agree to, or at least a transaction that they did not initiate.  IMO, this would further shake the confidence of short sellers and the market in general.

How should we solve this?

I think a KISS method similar to the original design would work because the original design wasn't given enough time to work IMHO.  Force all trades to occur within 1% of the price feed.  Allow users to place orders outside of that 1% range but the order wont be activated until their price comes within 1% of the feed.  Don't force short sellers to cover at any time.  Give sell orders priority over short sell orders when being filled.    Allow users to place market orders that float at the price feed.  No one is forced into a trade unless margin called.  I don't see any reason why it wouldn't work. 

Adoption and liquidity are a function of time.  Everyone wants to try all these complex approaches when the simplest approach may be best.  Now that the bts market has settled down, it may be worth a try.

Helikopterben, you may have missed a previous response I made to this idea when you raised it in another thread - there is nothing to stop people exchanging these assets somewhere else at very different prices in a free market. [edit: I think it could be achievable though if you allowed something else to float to change the relative attractiveness for longs and shorts]

Bottom line, when you look at the BitUSD price, the quantity available over $1.00 is VERY SMALL and thus meaningless.   That means that right now BitUSD is "perfectly pegged" and the premium for creating new BitUSD is VERY HIGH.   This is a good and healthy position and indicates what it will look like with force settlement enabled.    Everything else just depends upon trading volume and market making.   

I'm not satisfied. I would want a bitAsset that can be traded on any day against the real asset with a tight spread. If I'm just a moderate sized buyer, I want to be able to buy a bitUSD today or any other day for not much more than $1. And have the knowledge that I can sell it again on any future day for not much less. Sure big volume transactions need to pay a higher spread, but there should be arbitragers tripping over themselves to meet this at any meaningful deviation from fair value. I won't stop working toward a better solution until this is what we have, or its shown to be impossible.

Problem of the idea that BitAsset holder can settle for bts
BTA holders have an advantage: can settle back BTS whenever they want. This leads to the same result as the old rules.
For a healthy market:

• if market mainly want to sell, sellers need place lower price orders to match buyers'.This leads the price to go down.
  if market mainly want to buy, buyers need place higher price orders to match sellers'. This leads the price to go up.

In the settlement rule, the settlement function make the market do not work normally. Hold BitAsset to settle for BTS, will not leads BTS price go up, but hold BTS to get BitAsset leads BTS price to go down.

xiahui, I agree this concern is worth exploring further. It's one I'm thinking about too.
The pegging approach you suggest would not create the right incentives however...

Placing orders will affect the price instantly, and this affect the peg directly.

• if the bitUSD price is lower than realUSD,  new short orders pay instant interest. (to the system fund, not to the bitUSD seller. bitUSD can be sold to both short and buyer with BTS)
  if the bitUSD price is premium than realUSD, new short orders get instant interest. (from the system fund. not from the bitUSD holder.)

If the bitUSD is at a premium to realUSD, and users are not inclined to open new short orders in the absence of incentive, then the instant interest incentive would merely encourage them to open a new short, take the instant interest, then close their short again, ending up back where they started. So the premium will not close.
In the case of a discount, a fee on new shorts orders discourages new shorts, but does not incentivise any existing shorts to close, so the discount would remain.

You can always "self short" which is what he is saying.

BitUSD backed by BTS  is not fungible with BitUSD backed by BitGLD.      In a BTS black swan you need to settle BitUSD backed by BTS but not BitUSD backed by BitGLD. 

In other words, the quality and risks of the collateral impact the nature of BitUSD.   Having a "basket of different types of collateral" means there is no natural way to settle the market in a Black Swan.

Given a bond market that allows individuals to lend BitUSD backed by BitGLD the problem goes away. 

So far the only thing "different" I see is removing the ability of BitUSD holders to call the loan on demand.   Removing this ability to call the loan creates a power struggle that pushes BitUSD below the peg and potentially to 0.

Dan, this misunderstands the intent of the post.

The purpose of this post is a thought-piece to offer a wider perspective of bitAssets and how they could be used. In turn, that could guide the way we describe bitAssets in our marketing as well as their future design. This OP was never structured as a design proposal. Discussion of settlements, black swans etc is not relevant to the message, so their absence does not imply anything about proposed design.

The main (but not only) message is that bitAssets can be issued in a way that satisfies a more compelling and broader market need than simply offering leverage to BTS bulls. It can be issued to create loans for margin lending or any other purpose, or to earn market making spreads between bitUSD and real USD. The first is a borrowing motive, and the second is an income motive. We have a new way of describing bitAssets to the public and for marketing the short side without having to rely on the motive of BTS bulls. And we may wish to evolve the design to better suit those ends.

The question of what design implications this has for bitAssets is a secondary question which I will raise (in detail!) separately.

Oh BTW... re: self-short... you still have the issue of shrinking collateral so it's not risk-less.  I was going to discuss that when expanding on the contract for difference explanation the other day.   Anyways.....both designs probably become very similar in the end...

Agree, and as you can see I raised that you still need to manage the collateral. However the borrower does not profit or lose from this, as they would have held this collateral in their own hands anyway.

Quote from: Thom on May 20, 2015, 12:46:41 am

You lost me at CFDs.

Contracts for Difference (CFD's).  Maybe that helps

Ahh, yeah...sorry about that...merivercap recently gave a really good run-down of CFDs here...
https://bitsharestalk.org/index.php/topic,16391.0.html

I really like your idea starspirit.....as long as there is collateral, we can bring as much bitgold or bitsilver into existence as we like without the limitation of needing two parties that have a different view of the future price of bts. So if my Bts was worth 2 bitgold, I could bring 1 bitgold into existence, loan it to myself and use it to buy something else or keep it? If bitgold went up in price relative to bts, could I buy more bts, close the position and keep the difference in bitgold? If I bought BItSilver with the bitgold, I need to buy back enough bitgold at some future point to cover the loan and release the bts collateral?

Sorry, I find the web of consequences with this stuff and the implications of changing relative values so hard to get my head around!

You are simply borrowing in the bitAsset, with BTS as collateral. You would self-create your loan in bitGold if you thought bitGold was going to be the cheapest source of funding for your loan (i.e. decline in price). Your financial exposure depends only on where you invest those funds, and how that performs relative to bitGold, because eventually you have to buy back the bitGold to cancel the loan and release your BTS (*). You are never exposed to profit or loss from the price movement of bitGold relative to BTS. The BTS just sits there as collateral waiting for you to reclaim, and you just need to make sure you keep it topped up as necessary if it falls in value relative to the size of your bitGold loan, so that your loan doesn't get margin called.

To help understand this, when you self-create the bitAsset loan, you are at that point equally long and short, which is a neutral position. You are not exposed to price movement one way or the other, and you can cancel the two positions at any time. Its only when you exchange the bitGold funds for something else, that you are exposed to movements in bitGold because eventually you have to buy it back to cancel your loan. You have not taken any position at all between bitGold and BTS.

To xeroc's point, if you decided to specifically sell the bitGold for bitUSD, yes it would be similar to shorting bitGold with bitUSD collateral. But this is much more general than that -  instead you could decide to invest or spend the bitGold wherever you like, as long as you maintain collateral.

This is all possible with BitAssets today, we've just not promoted it this way. And if we think about it in these terms, it suggests new ways to think about the product design.

(* Technically, AFAIK, in the current system it is not possible to self-cancel a long position against your short position, and release the BTS, although I would propose this feature to allow exactly what is discussed above. Instead at the moment there is a more convoluted way to achieve the same end result - you need to cover by buying bitAsset in the market from the BTS collateral, returning the residual BTS, and then you would need to sell your long bitAsset separately for BTS to get your original BTS holding back.)

Its time we lost a lot of our old thinking and look differently at what our offering can provide to the world. Its time we stopped talking about bitAssets as CFDs (of which its only a mutated version anyway). There is no need to invoke the requirement that currency is created by speculative shorts. There is no need for all currencies and assets to be based on a one-size-fits-all bitAsset structure. They had a place once, but the external world does not care for and may even fear these technical curiosities. It wants familiar and better services offering greater freedom and flexibility. BitAssets are so much more than these things, and thinking of them in this way is heavily constraining their required evolution, and our path to self-funding profitability.

What is a bitUSD? Its a transferable token that represents a collateralised USD loan to a set of borrowers. Anybody can borrow in bitUSD, for any purpose, by providing acceptable digital collateral (*i). Anybody can use their borrowing as a margin loan if they choose to buy other digital assets, including BTS (*ii). Anybody can use the funds for market-making income by switching between bitUSD and real USD(*iii). Anybody can use the transferable token as a substitute USD.

The bond market (to come) is a way to lend and borrow USD at different terms. Again these USD funds can be used for any purpose as long as collateral is provided (*iv).

Combined, these services provide competition against Bitfinex, Poloniex, and others, with the advantages of eliminating exchange risk, allowing general purpose loans, and allowing loan receipts to be used as a substitute stable currency.

The trading exchange (to come) is a way to take margined positions on speculative assets. These can either be done in derivative form (expiry of instruments) or traditional CFD form (with an intermediary taking a spread). But the current BitAsset structure is not ideal for these instruments and needs to be discarded (*v). For a start, bitUSD should be the currency of denomination, not BTS. This service would compete against every asset exchange in the world with the advantage of eliminating exchange risk and lowering costs.

Unshackle the existing bitAsset paradigm, and suddenly we may see whole new ways to compete and earn profit. I see the potential for a more rewarding roadmap. It's not constant and evolves. But I think a roadmap on the services we are striving to offer is critical to making the right choices about our product design.

(*i Self-create a bitUSD by depositing the required collateral, sell the bitUSD for real USD funds)
(*ii Want leverage to BTS? Use the real USD funds to buy more BTS. Or why not BTC or LTC or DASH....)
(*iii Self-create bitUSD, sell bitUSD for real USD at premium, buy back at discount, etc)
(*iv Self-create a bond by depositing the required collateral, sell the bond for bitUSD or real USD funds).
(*v Require bitUSD deposit/margin accounts, leverage and long/short symmetry.)