Ok this is going to be another long one but few points I want to address:
-
vBNT vs BNT burning
After reading the comments here, it makes sense to me to keep burning in vBNT.
However, both solutions can be modified to allow vBNT vs BNT burning in any desired ratio.
So for the sake of efficiency, I suggest we assume we’ll have another discussion around that later on and for now let’s focus on the proposed solutions themselves. -
DAI distribution solution
I think it’s a good solution but one thing I don’t quite understand is how that has an advantage in terms of getting new liquidity in.
As soon as an LP deposits into a pool with a deficit, they take a % haircut, and then they get DAI distribution that’s initially supposed to cover that haircut. Does it matter that they get something in their wallet if it’s still lower than the deficit? -
New liquidity
It’s true that while pools are in deficit, they aren’t very appealing to new LPs.
However, not all pools are in deficit, the DAO can whitelist more new pools - these won’t be in deficit and can potentially even get into a surplus if the solution is viable. If we focus on these pools (and many of them), that can actually be a significant amount of new liquidity.
The fee distribution (in both solutions) can also be dynamic based on the deficit in the pool, so that LPs in pools with lower deficit or surplus can enjoy higher portion of the fees.
Re. LPs leaving when the deficit is lower - it’s true that there are probably many LPs that are waiting for the deficit to close in order to leave, but on the other hand, if the deficit closes, it means that these pools are actually getting profitable for LPs, so this might also present new opportunity for other LPs. -
Fees vs deficit
Very important to note that in any case the bancor community has lots of very bright minds that are going to keep building!
New planned features that are mentioned above are going to get the focus, features that improve the health of the protocol and increase fees, so any number that’s based on current/historical state shouldn’t be assumed as the baseline for the path to recovery/growth. -
Implementation details
Ok, since there is already sufficient positive feedback for these solutions, I think it’s time to start discussing some of the implementation details, as these are an important part of the decision making process here / time to market.
Both solutions have some open issues that we still need to finalize.
Both solutions require tokens to be traded to DAI (or ETH) through an external system, so that not to affect the pools.
Both solutions require fees in pools to stay off curve (v3 is unique in its ability to do so), so that fees aren’t accruing IL as well.
a. DAI distribution
How do we do the fee distribution?
Fees need to be normalized between all participants. Probably based on TVL contribution and potentially take deficit in pools into account - this is one issue we need to figure out.
Distribution should probably work similar to a standard rewards contract but with a variable distribution rate.
There were some discussions around an epoch based system - you need to stake your bn tokens into a contract and distribution is done after each epoch is over (day / week / month etc.).
It means tokens are staked for participation, a bit challenging in terms of UX etc.
b. Protection mechanism
How do we distribute the protection?
Assuming that fees are traded into ETH/DAI and then sit in the mutual protection bucket, how do LPs get access to it?
One of the options discussed is a dutch auction/leaky bucket approach (credit isn’t mine 
) -
- every day (for example), X amount of tokens can be accessed when an LP withdraws
- let’s say we start with 10% (parameter) coverage
- the next day, if the bucket isn’t empty, we add X amount of tokens again but this time the coverage is increased to 11%
- whenever the bucket gets empty, the next day the coverage drops by 1% again
- we can potentially also take the deficit in the specific pool into account, giving some advantage to LPs in pools with more deficit
This creates an interesting dynamic that LPs have an incentive to wait in order to get more coverage while LPs that leave earlier cause the deficit to reduce.
Of course all of the above are just preliminary ideas that should be hashed out.