LobsterDAO 🦞

24 Feb 2025 23:34

Not sure why you’re suddenly getting salty. It’s a tweet about the topic we were discussing that goes into more detail. You implied every protocol leaks a shit load of OEV and said “solutions aren’t being used”, which just isn’t true. That was my point.

LobsterDAO 🦞

24 Feb 2025 23:30

Hi, am I allowed to recruit here for a project I am building?

LobsterDAO 🦞

24 Feb 2025 23:14

Look I only wanted to chime in because there seemed to be an implication that liquidation fee extraction from protocols is somehow inevitable or even positive. By now you're just advertising Euler and you're the CEO. Of course your shit is the best in DeFi in your opinion. It's not a revelation.

LobsterDAO 🦞

24 Feb 2025 23:10

They are being used. Euler’s liquidation mechanism had an average liquidation penalty of something like 0.7% for loans above $10k in v1. Even smaller for larger borrowers. It’s a very effective system and the best in DeFi in my opinion.

https://x.com/euler_mab/status/1841429778026434573?s=46

LobsterDAO 🦞

24 Feb 2025 23:05

Any recommendations on Keepers? Is Gelato good enough for triggering regular rebalances programmatically - or is there risk of opinionated action on that network?

LobsterDAO 🦞

24 Feb 2025 22:40

I keep hinting at OEV because we built exactly this and its effectiveness has been well proven in the recent major liquidation events (Compound captured all of their liquidations on Mantle using it and got 90%+ of their liq fees returned to them). But it’s a bit of a rhetorical question in the sense that the protocols are for sure aware that this exists but choose to use Chainlink instead because reasons.

I mean we extracted >$165k in liq fees in one day using OEV and returned it to Compound. That’s Compound on Mantle which is peanuts. In this year alone their protocol has leaked well over $5M across all chains. God knows what that number is for Aave, not to mention the lending protocol vertical in aggregate. It’s massive and totally preventable.

LobsterDAO 🦞

24 Feb 2025 22:28

Aren’t most liquidations captured through blockspace auctions? I mean in effect gas wars in which searchers bid for the right to extract the liquidation fee/penalty. If there is a liquidation incentive, it only makes sense for there to be an auction to capture that incentive (as there effectively is). So the only remaining question is at what level the auction happens and who captures the proceeds - the protocol or the chain validators. Atm it’s the chain validators in the vast majority of cases, though that value could very simply be returned to the protocols. And it’s a huge $ figure in aggregate.

LobsterDAO 🦞

24 Feb 2025 21:53

Aave is in quite a unique position as it has seen multiple market cycles. Where usually protocols tend to design fees or not have fees at all until there is more risk that’s been realized during a downturn market and its too late.

LobsterDAO 🦞

24 Feb 2025 21:50

getting into a debate of who has the biggest/smallest proverbial X usually doesn’t matter to anyone else

What really matters is: security and trust when you drop serious capital into these contracts

LobsterDAO 🦞

24 Feb 2025 21:48

Feels like these terms are used a bit interchangeably, but yeah

LobsterDAO 🦞

24 Feb 2025 21:45

You’re confusing liquidation penalty and liquidation fee. Aave adds an explicit fee on top of the liquidation penalty.

LobsterDAO 🦞

24 Feb 2025 21:44

I mean if it’s using an OEV-enabled oracle. Which Aave isn’t.

LobsterDAO 🦞

24 Feb 2025 21:44

The lending protocol can also recover the vast majority (up to 90%) of the liq fees by using an OEV-enabled oracle, so it doesn’t get extracted as MEV but rather returns back to the protocol.

LobsterDAO 🦞

24 Feb 2025 21:33

For the liquidation fees? 🤣

LobsterDAO 🦞

24 Feb 2025 19:38

Borrowed 1B vs 13B

But fees are like 40x different

LobsterDAO 🦞

24 Feb 2025 23:31

Last time I checked this wasn’t a looking for a job group 🤯

LobsterDAO 🦞

24 Feb 2025 23:14

You're quoting your own tweets about your own protocol dude.

LobsterDAO 🦞

24 Feb 2025 23:13

Best is of course having an ability to be deliquidated

LobsterDAO 🦞

24 Feb 2025 23:05

I disagree. Auctions on health score do guarantee the best outcome. They don’t delay liquidation. They find the fair price beyond the cost to liquidate. They lead to the lowest liquidation penalties for the largest borrowers. Whenever you take additional collateral from borrowers unnecessarily you only drive them unnecessarily closer to insolvency. That’s not just bad for the borrower, it hurts the lenders too.

LobsterDAO 🦞

24 Feb 2025 22:41

Anywho, just figured I’d chime in since it’s not like there are no solutions, they’re just not being used.

LobsterDAO 🦞

24 Feb 2025 22:30

https://www.paradigm.xyz/2024/06/priority-is-all-you-need this might be a design space to improve a bit on

eg charging additional fees based on priority

LobsterDAO 🦞

24 Feb 2025 22:01

Also auctions doesn’t guarantee necessarily the best outcome. One key element with liquidations is to ensure there is sufficient incentives to establish a liquidator network. Thin spread auction margins might turn off liquidators elsewhere to other opportunities. Most recent market crash Aave liquidated over $200M few weeks ago during the market downturn, thats amazing accomplishment for DeFi.

LobsterDAO 🦞

24 Feb 2025 21:52

Pretty much the defaulting model in lending is to price as accordingly to the carried risk as possible, so let say that the less risk the user is bringing into the protocol, the less charge on that lending utility should count, whereas adding more risk that would mean it should be priced in, otherwise higher risk is carried by those borrowers who bring less risk to the protocol, so for example if a borrower keeps a position close to liquidation thats higher risk user than the one that keeps higher collateralization. Hence liquidation fees are following the same apporach where protocol earnings are carried from positions that create higher risk (i.e. creating a higher chance of liquidation).

LobsterDAO 🦞

24 Feb 2025 21:49

imo it will all be the same at the end and it’s a race to the bottom on fees. Where you take the fee doesn’t really matter - more established protocols will take higher fees as others will clearly be incentivised greater to accelerate the race

the most important thing I look at is liquidation limit not fees tbh

LobsterDAO 🦞

24 Feb 2025 21:46

Euler doesn’t leak OEV to the same extent because we use an auction on health score to determine penalty. Euler therefore has lower penalties and zero fees compared to Aave, and our goal is to attract borrowers by being much cheaper in that regard.

LobsterDAO 🦞

24 Feb 2025 21:45

But it’s an option for every lending protocol that is proven and exists today.

LobsterDAO 🦞

24 Feb 2025 21:44

By effective cost of borrowing I mean that borrowers who mostly go long bitcoin or ETH (the few that still exist) should add to their cost of borrowing the probability of liquidation multiplied by the liquidation fee. But I agree it’s not as simple as liquidation fees = bad. There are pros and cons for lenders and the protocol. Just not many pros for borrowers.

LobsterDAO 🦞

24 Feb 2025 21:41

this is not exactly accurate “ Liquidation fees make the effective cost of borrowing higher.” actually if there is not sufficient liquidation fees one way or another that would reflect into the interest rates and thus again borrowing costs itself. Also higher revenue from riskier positions is actually healthier for the protocol as it creates financial buffer for the protocol that can be used for rainy days as lending is not a zero risk game.

LobsterDAO 🦞

24 Feb 2025 21:07

Aave also takes fees on liquidations, which on days where there are big price moves, can be quite large. It’s a source of revenue most lending protocols have opted not to take for a variety of reasons. Obviously works well at making money for Aave, but for now our view is that borrowers will prefer to use platforms without them. Even a small fee can be quite punishing when multiplied for a leveraged position. Liquidation fees make the effective cost of borrowing higher.

LobsterDAO 🦞

24 Feb 2025 19:35

Because you're looking at Tvl, the real metrics is total loans