After reading @lesigh 's response, I decided to further analyze his data.
The following data-points are only related to the v2.1 LINK pool.
Since this pool is one of the biggest (if not biggest) in surplus, it is assumed that other big pools that are in deficit (e.g. BTC, ETH) in v2.1 will even have a higher % of active positions that are better off by simply migrating to v3. Thus speaking for a forced migration of v2.1 pools
It is assumed that the v2.1 withdrawals will happen without BNT minting
After fetching every active position and extracting the data, I fetched the deposit date and added the estimated IL based on @lesigh’s numbers
74,86% of positions (524) have a bigger IL than 28% and profit off a full migration
Where the data could be NOT 100% accurate:
Wallets with multiple positions and less withdrawals than deposits (85 of 700 positions to be exact). Excluding those wallets result in less than 1% deviation though.
If you find occurences of wrong data, please let me know
TL;DR
Found 5284 ProtectionAdded Events
Found 4584 ProtectionRemoved Events
Resulting in 700 active v2.1 LINK pool positions
Calculated the following with data provided by: @yudi (~28% v3 LINK deficit after a full migration) and @lesigh (IL % numbers per date)
74,86% of positions (525/700 in total) would be better off migrating to v3
Therefore a majority of Link v2.1 LPs would be in disadvantage, if they withdrew from v2.1 directly and v3 migrations get disabled.
Tried my best to make the data as accurate as possible, results also kinda match with lesighs results
Proof of analyzed data:
Dune Query to get the amount of LINK v2.1 pool protected positions: Result = 700
I ran some quick numbers on the scenario you laid out and I don’t believe the math is accurate. From my calculation the improvement to the deficit of v3 is almost immaterial in all scenarios I ran. The best case scenario assuming all 2.1 LPs migrating have no IL only deems a 5.2% improvement to the v3 deficit, not the ~16% improvement implied in your analysis which factors in the 75/25% split of impacted LPs with IL.
You can see a current snap shot of the state of v2.1 and v3 LINK pools in the top 2 grey tables. The tables in yellow are scenarios based on the information presented in your summary, and additional numbers running more realistic scenarios. All scenarios are run on deposits only and do not factor surplus, as this is protocol owned revenue and should not be factored into the analysis, as neither v2.1 or v3 LPs have any claim to it. It’s important to note, that given your analysis was run on split of total deposits and not LINK qty, I ran the 75/25 split scenario based on the total LINK deposits.
Scenario 1 (5.19% improvement): This is an unrealistic baseline showing a full migration of v2.1 deposits w/ no IL in aggregate
Scenario 2 (2.62% improvement): This is the scenario you stated in the post of LPs having greater than 28% IL. I ran this at 29% conservatively. Please note, the statement “74,86% of positions have a bigger IL than 28% and profit off a full migration” makes no sense, as the v3 deficit is 43.8%. This would imply that almost all v2.1 LPs would be at a detriment migrating.
Scenario 3 (1.2% improvement): Based on your statement that 75% of LPs would be better off migrating, this implies that their IL is greater than the 43.8% deficit currently in v3. I ran this at 1% greater than v3 deficit.
Scenario 4 (0.4% improvement): This is more realistic assuming 75% of LPs have 10% IL than than v3, benefiting greatly from migration.
Scenario 5 (-0.9% impact to v3): This shows an extreme benefit to the 75% of LPs assuming they have a 25% worse IL than v3.
What seems abundantly clear to me now is that the most heated debate with this whole fiasco revolves around users withdrawing their funds from either v2 or v3, and for just reason, simply because everyone cares for their own funds.
The solution for resolving bancors problems does not involve users ability to leave from a particular version, so IMO I would like to drop this topic, not take any action regarding this topic for the time being, and focus ALL of our brain power on resolving the deficit.
We need to pivot discussions from this to something more constructive. If we find other solutions, then I believe we will all be in a better position at the end of the day. So lets simply agree to disagree and revisit this at a later date after we come to an agreement on a few other changes. Should the conversation continue to revolve around this topic, I don’t see us moving forward, when right here right now, this is the time for us all to come to an agreement on various fixes, have the DAO vote on them and start to execute asap.
I feel like we need a resolution to be able to move on from this. If not for the fact that maintaining v2.1 could be a liability, I’d say let’s vote to revisit this in 6 months. Because it seems to pose a risk, let’s put it to a vote and be done with it.
I’m for a full migration as it is the fairest solution I’ve seen by far.
I also think that a resolution here is required. What I suggest is that rather than voting on multiple different proposals (since there are many offered here), that this gets broken down into smaller actionable items. The first vote should be a simple proposal whether we should enable or disable withdrawals from v2.1. Whatever the outcome of this vote is will decide what the next steps are and how to deal with potential options.
I recommend a separate thread for this proposal so that it is kept clean with arguments for and against. We should give folks enough time to comment before it goes to Snapshot so it would make sense to have this up next weekend or the week after that.
if we are to take action, id personally prefer to bite the bullet and get it over with sooner than that. If we are to deal with this now, The sooner we deal with these issues, the sooner we can focus our attention on matters that will help bancor as a protocol recover. Bancor recovering is good for everyone here regardless of which version theyre in
A vote to forcibly move users funds without consent is egregious. The fact of the matter is that user funds should never be up for discussion to the DAO given the lack of information available and technical/economic expertise from members to make informed decisions on matters this material.
The upgradeability clause clearly states v2.1 LPs have always been given the right to opt-out of upgrades. I have opted out of v3.
It is negligent that the team is even allowing this to potentially go up for vote. The stakeholders effected by this outcome don’t even have a say because of a flawed governance model, when in reality they should have voting weight 7.5x the current active DAO members given the DEBT owed to LPs. This is a gross misrepresentation of stakeholders.
There is only 3-4M vBNT actively voting on the outcomes of proposals since the 6/19/2022 shut off. This is about $2M of voting power determining the future of over $150M+ in total TVL across both versions.
This $2M of voting power is now faced with the task of determining whether to forcibly move user funds without their explicit consent, which across all of v2.1 is a -$15M deficit [need to double check figure].
This deficit is actually a DEBT owed in BNT to all 2.1 TKN holders due to rebalancing not being paid out. This DEBT is effectively a staked BNT position in the system, yet we are not issued vBNT to have a say in our own outcome.
The 2.1 DEBT holders should have a total voting power close to 30M vBNT ($15M deficit owed), which is 7.5X the voting weight of the vBNT voting power of its active community. This is such a misrepresentation of stakeholders interest it is comical.
TLDR:
There is only 4M BNT ($2M) actively voting on the outcome of $150M+ TVL
TKN LPs should have voting power as the DEBT owed to them from rebalancing is effectively a staked BNT position.
This DEBT represents 100M ($50M) BNT in aggregate voting power across versions for TKN LPs that is not represented in the fate of their own outcomes.
The TKN LPs DEBT in aggregate represent over 25X the voting weight of the current active vBNT holders participating in the outcomes. This is gross misrepresentation.
As mentioned above, there is no “team”, we are all community members. And nobody makes protocol level decisions for the community other than the DAO.
No one can prevent anything from going on a vote.
I agree that forcing migrations without consent through an upgrade is an extreme measure that should only be brought up in case of emergency in order to protect those same funds, and this is clearly not the case.
I agree with @alphavalion that this discussion includes 5-6 different topics and that it probably makes sense to separate them into different discussions. But I also think that practically nothing that is discussed here other than the previous point can be proven without a doubt objectively. Nothing in the math is objective and time cannot be rolled backwards. The surplus/deficit numbers were affected by different things - user migrations, withdrawals from v3, rate changes in the pools in both v2.1 and v3 and the design of each version of the protocol. And add a time component to any of these points which pretty much makes it a chaotic nondeterministic system. As such, there’s (sadly) no single approach that can be taken that will be fully objective and that can satisfy everyone.
Do you see the issue in your 1st point? v2.1 LPs are community members, but not part of the DAO. Many v2.1 LPs have more assets at stake than even the voting vBNT members, but they have no voting power or say in what happens to their money.
I don’t think there’s necessarily and issue with that - that’s how every DAO operates and anyone can participate by getting vBNT.
I do think that the DAO should define some kind of “code of conduct” but that’s for another discussion
I do believe my 3rd point covers your concern.
It is different when a protocol is facing insolvency like Bancor is. When it becomes obvious that revenues are not enough to pay off the deficits, the interests of the creditors, in this case the Token depositors, should have more value. The fact that other DAOs do not have a mechanism in place for this scenario just shows the nascency of this industry and should not be used as the established norm.
I wouldn’t go as extreme. The protocol will eventually recover.
It’s just a young industry and DAOs will evolve with it. There’s no standard yet but there will be.
I agree that since it’s not an extreme situation, a forced migration wouldn’t set the right precedent going forward.
I think giving an option to migrate to v3 after seeing your IL on v2.1 isn’t fair to v3 LPs for e.g. the ETH ones. The Link LPs wouldn’t mind as the surplus is more on v2.1.
Getting into it pool-by-pool basis would be over-engineering and not the right way to dedicate engineering bandwidth.
I propose giving 2 weeks to v2.1 LPs to move to v3 after the proposal has passed, after which withdrawals should be only on v2.1. We can share the latest deficit data to help LPs make a decision on this, but individual LP level IL shouldn’t be shown as that would require engineering effort, would delay the whole process more and wouldn’t be fair.
I think the order for this should be different for the “For” vote:
2 weeks to withdraw from v2.1
close withdrawals
return to the current status quo
Also, the “Against” vote here should clearly state that the status quo be kept and ratify that withdrawals are not allowed from v2.1 going forward. For v2.1 LPs to withdraw, migration to v3 is the only path possible going forward.
Has already been mentioned that this refers to the opt-in upgradeability by the pool owner.
And the option that LPs can opt-in to upgrades simply does not (and never did) exist in the code.
and does not mean what you are suggesting.
It is negligent that the team is even allowing this to potentially go up for vote. The stakeholders effected by this outcome don’t even have a say because of a flawed governance model, when in reality they should have voting weight 7.5x the current active DAO members given the DEBT owed to LPs. This is a gross misrepresentation of stakeholders.
Bancor is a protocol that is DAO controlled and anyone can participate in the DAO if they are a vBNT holder. AFAIK, the following token holders $veCRV, $COMP, $AAVE, $UNI, $SUSHI, $BAL determine the fate of their protocols and liquidity providers for these platforms don’t have a say in any governance decisions for these protocols either. Feel free to bring up a proposal that amends the governance requirement on Bancor to allow non token holders to decide the future of the protocol (might also want to change this across all of DeFi as well).
I think maintaining the status quo for v2.1 is probably a good choice as long as there is a possible way to withdraw which there currently exist via v3.
probably best to keep the vote as simple and clean as possible I think and migrating all liquidity to v3 can be a separate vote to not pile on to many things into the proposal.
Should really be what lesign suggested:
but instead of migrate should be open withdraws for 2 weeks, close withdraws, back to the status quo. The against vote should be to keep the current status quo with withdrawals only occurring via version 3.
Equalising the surplus wouldn’t be fair to v3 LPs at all unless it is done daily. One of the versions will definitely have more deficit but only v2.1 LPs would have the chance to migrate to v3 and reduce theirs.
Also means that v2.1 stays alive for much longer, which again is more effort in my opinion.
Giving v2.1 LPs the window to leave at their respective IL and moving the surplus to v3 is the best way forward.
