LobsterDAO 🦞

02 Dec 2024 21:02

At this point, there are more efficient ways to offer higher LTVs while managing system risk without relying on trust like capital-efficient liquidation mechanisms or reducing other risks. Borrowers can access higher LTV loans, and protocols can provide them, without requiring users to build trust over time (which adds costs as you said) or relying on past behavior to predict future actions.

LobsterDAO 🦞

02 Dec 2024 20:56

One way to minimize the gaming of credit scores is by making the cost of gaming higher than the potential benefits.

This ensures that users can't extract value before providing real value to the protocol. And if someone still defaults, the effort they put into building their credit score would be wasted, discouraging such behavior in the first place.

LobsterDAO 🦞

02 Dec 2024 20:43

Simple action-based incentives are too easily gamed, and it's too dangerous for risk underwriting.

Users can game the system by borrowing and repaying small amounts or even large amounts multiple times to improve credit, then default on larger loans. They can exploit higher LTVs with multiple accounts or manipulate data to unfairly boost scores. All it might converge would be a situation where risks are inflated as more users gain higher LTVs without truly lowering risk.

LobsterDAO 🦞

02 Dec 2024 20:23

Don't think it will. Since there will be two distinct liquidation points, id argue that would make the protocol safer.
Plus give the protocol those sweet liquidation penalties.

LobsterDAO 🦞

02 Dec 2024 19:30

Would love to know more, sent dm!

LobsterDAO 🦞

02 Dec 2024 18:51

you could build a bank like loan facility and tap into Aave liquidity, and be mainly focused on borrower credit worthiness

LobsterDAO 🦞

02 Dec 2024 18:30

But what are the featurs derived from the composability? Use cases?

LobsterDAO 🦞

02 Dec 2024 18:27

Credit del was never about undercoll, it is about composibility

LobsterDAO 🦞

02 Dec 2024 18:26

Defi is kind of approval + credit delegation, and some extra logic around 😁

LobsterDAO 🦞

02 Dec 2024 18:23

Only project i know that did this and is still working is ethichub. It's small scale social impact lending though. @daveytea

LobsterDAO 🦞

02 Dec 2024 17:34

yes you could delegate to any address, hence why you'd become the middleman for managing 'credit worthiness'

LobsterDAO 🦞

02 Dec 2024 17:31

i wish more people would have built on credit delegation mechanisms. We tried to push that hard at some hackathons in the early v2 days but no one really nailed it. Essentially you become the middleman and need to manage the borrowers with incentives/disincentives, and manage trust with your delegators.

LobsterDAO 🦞

02 Dec 2024 16:58

https://forum.blast.io/t/blip-buyback-blast-token-with-yield/786

LobsterDAO 🦞

02 Dec 2024 16:04

You do need something when the position is under 100% collateral.

LobsterDAO 🦞

02 Dec 2024 15:55

The topic above was not about trust retail credit (what I initially thought was the question too) - but to the protocol. In which case you don't likely need any trust credit scores.

LobsterDAO 🦞

02 Dec 2024 20:56

And honestly I guess users getting liquidated has less to do with responsible / irresponsible behavior.

Liquidations in overcollateralized loans aren’t about trust or responsibility. Unlike undercollateralized loans, where repayment depends on the borrower’s reliability, overcollateralized systems act( liquidate ) long before positions become undercollateralized. These systems rely entirely on collateral, not promises same for the counterpart - borrowers repay and maintain loan health out of self interest, not trust. As you said, no one wants to get liquidated, not sure if encouraging users to build trust will improve the fundamental situation.

LobsterDAO 🦞

02 Dec 2024 20:55

Estimated buy pressure seems pretty high at $36M. I wonder how large farmers respond if part of their angle is the native yield. On the other hand, buybacks would increase the value of gold incentives and future airdrops

LobsterDAO 🦞

02 Dec 2024 20:40

I get your point about higher LTVs potentially generating more liquidation penalties, which can be profitable for the protocol. It’s an interesting perspective, and I can see how that could work if we assume liquidations are always timely and efficient.

But don’t you think frequent liquidations bring bad user experience? If users feel like the protocol is designed to push them toward liquidation, they might stop using it altogether. After all, no one wants to lose their collateral repeatedly or feel like they’re being penalized constantly.

So implementing a system that rewards responsible behavior like giving better terms to users with good credit scores could improve user retention and attract more users in the long term. Anyways this is what I feel :)

LobsterDAO 🦞

02 Dec 2024 19:44

pardon, but what’s innovation in the thing that you just allowed users to make unsecured revolving lines and saved them tx fees of 2 transactions ? It’s the same as I borrow myself and send tokens to the borrower, but they borrow directly

LobsterDAO 🦞

02 Dec 2024 19:29

Wouldn't you think that would keep the protocol at risk if liquidations didn't happen on time? If we have less LTVs for users who usually get liquidated we'd have more time/gap to liquidate so that the position will not go under collateralized too quickly but in your case it's risk protocols take like betting on their liquidations being on time right ?

LobsterDAO 🦞

02 Dec 2024 18:31

It is programmable credit

LobsterDAO 🦞

02 Dec 2024 18:28

Same as approve() is not about payments

LobsterDAO 🦞

02 Dec 2024 18:27

Why dyt delegation so important? We better find a new mechanism for under-collateral loans such as credit score and long term capital ownership.

LobsterDAO 🦞

02 Dec 2024 18:24

Aave credit delegation is most probably the most interesting feature we introduced on aave v2. Definitely the most elegant defi-wise imo

LobsterDAO 🦞

02 Dec 2024 18:18

I'd argue as a protocol, you want to give higher LTV loans to users who do get liquidated

LobsterDAO 🦞

02 Dec 2024 17:32

maybe these days with zk you could prove certain assumptions of the borrowers and let them have under collateralised loans, with progressively more borrowing power based on their reputation. Tellor experimented with this back in the day as well

LobsterDAO 🦞

02 Dec 2024 16:59

At least it will help us differentiate between users who get liquidated most of the time and users who doesn't right? then we can increase LTVs for good credit score users and decrease LTVs for bad credit score users.

At the end of the day this qualifies as it makes sure lenders funds are safe in the protocol. I think this also reduces liquidations.

LobsterDAO 🦞

02 Dec 2024 16:05

The only thing we have working rn is full/over collateral and liquidation

LobsterDAO 🦞

02 Dec 2024 15:57

Actually credit in a sense of "more capital for defi" was done by Gearbox in 2021 and is live today. Credit means leverage, loans, etc. If is credit in a sense of leverage, you can't just tale the money out. If you mean in a sense of overcoll loan (still a credit line end of the day) - then you can.

And it's getting elevated into "credit inside any wallet" where you do NOT need credit scores either. Both fully onchain, no documents required. So imagine turning your favorite wallet into a credit wallet, this (https://docs.gearbox.finance/overview/credit-account) concept much beyond... more than just leverage.

A bit of muh alpha, for real.
If anyone cares**

LobsterDAO 🦞

02 Dec 2024 15:51

We did an employer-based lending project. In which the employer used it as a way to do cash-advance that was offramped to m-pesa. So basically due diligence was done of the employees by the employer