danielattilasimon
commented
2 years ago Should we include shifting to Curve in the testnet version? If so, how realistic should it be?
- Option 0 (NOP): no shifting.
- Option 1 (very easy): set thresholds to 1 in both directions, so the entire permanent bucket can be shifted at will, in any direction.
- Option 2 (easy): simulate price movement (e.g. a pseudo-random walk) but no slippage, which would at any given time allow the entire permanent bucket to be shifted in one direction depending on the current pseudo-random price.
- Option 3 (harder): simulate both price movement and slippage, so the shiftable amount may at times be limited by the balance of the (fake) Curve pool.
- Bonus option: this is kind of "hard mode" for yield amplification. Hardcode a LUSD price of 1+ so that the entire permanent bucket can always be shifted into Curve, where it gets trapped. This would result in the lowest yield amplification.
I'm leaning towards something simple like option 1 or bonus option.
We want mocks for yield sources (B.Protocol and Curve) for testnet deployments, so we don’t need to rely on a reliable fully fledged stack (Chainlink + Liquity + B.Protocol + Curve).