cunicula is correct, BitAssets will go to 0.
It is very easy to explain. Both the short and the long have an incentive to ask and bid (respectively) a lower price. Both seller and buyer want a lower price (at the moment they transact). Period.
Short wants to sell for a lower price, so he is protected on the downside. Long wants to buy for a lower price, so he can gain more on the upside.
If I am going to go short, I want to sell for as high a price as possible because I have borrowed the asset and will have to buy back at a lower price in the future. Thus the foundation of the rest of your conjecture that BitAssets will go to 0 is gone.
Damn, I brain farted (in bed), inverted my logic on long needs higher price, short needs lower price (trying to do too many things at same time).
Okay excellent. Now I can excited about BitAssets again.
So then I continue with the thought process I was doing to redesign BitAssets to make symmetric the stochastic range of preferences on both sides, long and short.
If instead of borrowing to go short, both short and long put their collateral as backing for the BitAsset, and they agree to subtract the (L x) movement in price (higher from short, lower from long) from their half of the collateral and give it to the other party. Then they can agree to any term of contract expiration. We can even support leverage, by making L > 1.
Example, if short ask and long bid meet at 1 BitUSD for 10 BitShares. They both put up 10 BitShares of collateral. If L = 1, and at the end of the agreed expiration, the market price for 1 BitUSD is 6 BitShares, then the long gives 4 of his collateral to the short, and the BitUSD (BitAsset) is retired. (Problem: how do we determine market price?)
My expectation is that by balancing it this way, instead of your current asymmetric design (your design short gets margin, but long never does and no expiration date), then the shorter-term expirations will much more closely track the value of the designated BitAsset.
Why?
1. Both long and short have a balanced set time preference.
2. Shorter-term positions remove the secular trend of the designated asset from the equation.
3. The short and long are more frequently renewing their positions in the market place of bid / ask, thus resetting their target price to the current real-time price of the designated asset.
The problem with my design above as compared to bytemaster's design is that there is no buying of a BitUSD to cover, thus there isn't a 1-to-1 trade for each covering. Rather my design expects a "market price", but such doesn't really exist.
So let's fix my design.
My solution is in the above example, the long must buy a BitUSD and give it to the short. The long keeps the remainder of his collateral, plus takes 2 BitShares from the short's collateral. This is done automatically by the miner of the block. Note this is what I originally thought BitAssets were designed to do. I later learned from bytemaster's example upthread, that his design is asymmetric. In this case, leverage can still be paid out, by giving the appropriate calculation of collateral to the short, along with the BitUSD.
But my design still has a problem. The BitUSD the long buys can't be one that expires. So this creates N classes of each BitAsset, one for non-expiration, and N-1 others for different expiration terms. But how does the BitUSD with no expiration get created? We are back to bytemaster's design in order to create the BitUSD with unlimited expiration time. So what have I just described? Options on the BitAsset?
The asymmetric design has the advantage that the long can hold his/her position indefinitely. Both designs could be offered in the system.
Can anyone make any other observations of ramifications between my design and bytemaster's design for BitAssets?
