Self-creation, self-cancelling, self-rolling and auto-rolling functions

[starspirit]

The potential uses of self-creation, self-cancelling, self-rolling, and auto-rolling functions


These ideas have been touched on in a couple of recent threads, and also been discussed to some extent in the past. My idea here is that although many of these functions can theoretically be performed through a series of manual steps in the current system, having simple functions that automate these would have the following benefits -

- they simplify the processes to "one-click" processes
- they would offer simultaneity of matching trades, which eliminates the risk of other parties stepping in to take one side of the order
- as such, they facilitate market-making and short-side management
- there is no risk to the system if the conditions of these self-matching orders meet the conditions that would have existed in implementing them manually

The last point is important. As long as the incentives in the system are set up right, we should have no problem with taking out the practical inconveniences that rational actors would experience in trying to manually replicate these strategies.

There's no doubt issues I have not thought of, as I've not read all the previous history, so apologies in advance.

Self-cancellation

Self-cancellation does not alter the ability of sellers to achieve the feed price, since it buys from the pool of sellers that might exist below the price feed. Actually I discuss here...https://bitsharestalk.org/index.php?topic=14835.msg195515#msg195515...why I currently think it is necessary to have a mechanism for sellers to be able to set a sell order that moves with the price feed in order to guarantee they will receive the feed price level within 30 days. In any case, self-cancellation orders (in the current system, covered short orders) do not jeopardise this at all.

There are alternate ways to accomplish self-cancellation other than the current short-cover procedure. Self-cancellation could also be achieved by presenting a unit of the bitAsset (from any source) against the short position. This would be useful in certain strategies such as the one described in self-creation below.

Self-creation

Self-creation provides a neat way for profit-seeking market makers to exploit bitAsset premiums when there is not enough bitAsset remaining in inventory to sell. Simply, the market-maker takes some of their BTS inventory, self-creates bitAsset (long and short) then sells the bitAsset for the real asset on a centralised exchange. Should the premium contract at some future time, the market-maker can exchange the real asset for the BitAsset again, and present the bitAsset against their short position for self-cancellation (alternative means of self-cancellation described above). The big advantage of this approach for the market-maker is that at no time are they exposed to the directional movement of the asset, and they unlock just as many BTS as they put in to start with. Their profit comes in the form of units of the real asset, which of course can be converted afterward to USD, BTS or whatever exposure is preferred.

To do this in the current system requires the market-maker to manually match a long to a short order. In slow-moving, illiquid markets, that may seem easy enough. If fast-moving liquid markets (which we hope to see) there are risks of others taking one side or the other before it is matched. This creates an exposure to both BTS and the bitAsset that needs to be quickly unwound (potentially at a loss) or hedged (at risk and cost). From a pegging perspective, facilitating and simplifying this type of market-making strategy helps to manage the potential excessiveness of bitAsset premiums from time to time.

In practice self-creation requires there to be no sellers below the price feed. It also requires the new short order to be at the front of the short queue, which means either the highest interest rate (if at the price feed) or the lowest price (if above the price feed). So it ought to be possible for anybody to queue a self-creation order at an interest rate of X%, and when it's the highest it gets created. Note that these limits are not limits at all when the bitAsset is at a premium, because there are no sell orders below the price feed, and with no shorts at the price feed, shorts are queued on price and can pay 0% interest.

But the process could be even more sophisticated than this. In practice if there were a small short order at a very high interest rate at the price feed (that blocks reasonable self-creation), one could manually buy out this self-order, sell the bitAsset at the highest bid, then do the self-creation. This comes at a cost equal to the volume of the order to be taken out, times the price spread. But by taking this out, the interest rate required for self-creation might be much lower. By specifying a maximum cost that one is willing to bear in the self-creation process, along with the interest rate, this could all happen automatically, without the market shifting or getting in the middle as the various steps are attempted manually.

Self-rolling

Self-rolling is essentially self-cancelling plus self-creation occurring simultaneously. So the conditions on self-creation also apply to self-rolling. Self-rolling orders could be placed with a maximum interest rate and roll cost, and when the market allows this, the roll occurs.

Auto-rolling

Auto-rolling could be a permanent (switch on, switch off) feature that always self-rolls when it can, subject to one's interest and cost parameters.