Should settlement prices have a size-based scale?

[starspirit]

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.

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]

[Permie]

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

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

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"

[starspirit]

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

Dan (& others), here is another idea that may affect the way you plan to implement price feeds and settlements. I know there is a mooted change with price feeds taking account of liquidity in some way, but not clear what this is yet.

In principle, a bitUSD that has the same purchasing power as a real USD should only be able to settle for as many BTS as could be bought in the external market. This implies that the settlement price should have a sliding scale that varies with the size of the settlement being requested, and this scale might vary over time according to the market depth on the ask/sell side of BTS in the external market.

Anybody settling $1 gets settled at the external mid-market price of BTS. $100000 would require paying a premium for the BTS. In this way, whale bitUSD holders cannot claim strategic stakes in BTS without paying up for it as they would in external markets.

Its also not possible for settling parties to manipulate the external BTS price to seek more a more favourable settlement price. If they try to sell BTS down by taking out bids this has no direct impact on the sell side of the market. If other sellers follow this lead and drop their asking prices, their ask prices will likely still be higher than where the manipulator was forced to sell. 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.

With this approach, there might be no need to limit the size of settlements. This would be self-reinforcing for any whales that wanted to obtain a better price by splitting their settlements. It also means settlers don't need to wait long periods before other settlements are filled. There would also be no need to change the price feed from the mid-price to something based around higher liquidity like $100000. Small settlers would still get the usual price feed (and so still maintain a bid close to parity with a real USD in the external market), while larger settlements incur a fair price based on the scale.

Just throwing this one around for thoughts.

[Edit: This approach may need to be modified to allow for situations where there are hordes of simultaneous small settlements, which should also ideally require a higher premium. I've got an idea on this too, but would like to discuss above as a principle first, even though it may not be perfect.]