RE: Algorithm Based Assets on the Bitshares DEX!
I've written extensively about HERO. I think it would be fascinating to create one with a very high algorithmic appreciation, something insane like 100%. The purpose wouldn't be to actually offer 100% gains; rather, it would be a market experiment to see what average returns it would actually offer.
Obviously (I hope), nobody in their right mind would short it at the feed price. At what price would they short it? 2x the feed price? 10x the feed price?
On the other side of it, what price would people pay for them? Think of it this way: if it traded consistently at 2x the feed price, it would appreciate at a rate of 50% annually. Would you buy that? I certainly would; in the worst-case, I could just request settlement a year later and cover my investment.
Or maybe the whole thing would globally-settle at the first sign of trouble.
I've been thinking about very high appreciation rates recently, let's assume 365% p/a (1% per day) - a shorter may sell it at the feed price because they believe that BTS is going to appreciate by more than 1% that day and they don't want to wait for someone to buy their high spread bid.
If there was such a high appreciation rate, there would surely be a huge buy pressure at the feed price enabling greater shorter opportunities however there would be the risk of global settlement if holders refused to sell the token (due to high appreciation rate) so the shorter (debt holder) would potentially be unable to close their position unless they bought far greater than the feed price..
So yeah, the shorters would be best advised selling greater than the feed price because its possible that by selling at the feed price they would have to buy back their debt greater than the feed price which would eat into their shorting profit..
With BSIP18 improving global settlement recovery we can investigate such possible assets!
Global settlement risk is hard for me to quantify, and you're right - that might be a major factor in aggressively-appreciating FPAs. There's also the issue that margin calls are implemented too conservatively; a +1% per day asset would likely need to have margin calls that buy at much more than 10% above the feed. Conservative margin calls are black swan magnets.
BSIP18 will be a good step, but I still think it's a bandaid. The fact that a black swan is caused by a single undercollateralized debt position, however small, is absolutely insane to me. I could owe $1, the next guy above me could have a 6000% collateral ratio, and if my collateral ratio goes below 100% the bitAsset would be shut down and all that juicy collateral returned. It's madness.
How do you believe we could improve upon this current system? Are there comparisons we could make to centralized systems? There are no bad ideas when brainstorming! :)
I haven't thought about it yet as thoroughly as I'd like. But it does seem like a global liquidity pool might not be a bad idea. Maybe you aren't allowed to short something until you pay a "membership fee" or something, and that fee pool is used to cover undercollateralized assets. If you're the account responsible for the black swan, then you get kicked out until you pay back into the pool. Maybe?