pull down to refresh

I had heard of Ducat before, but didn't ever bother looking in to it. Then I saw this post on X claiming

For the first time, liquidations on Bitcoin are not only possible. They are public, enforceable, and profitable.

This made me curious. Especially because they say they are using Taproot to do it. One big question I have with the bitcoin loan companies is how their oracle works. Even if the liquidation process is 100% automated by some on-chain script, the oracle is the weak link: if someone manipulates it they can steal your bitcoin...which isn't so different from just giving your bitcoin to someone in return for a loan. I was curious if they have a solution to this problem.
Here is how they say it works:

a way to handle liquidations directly on Bitcoin Layer 1 by:

  • encoding a liquidation path in a tap tree that requires the pre-image of a hash
  • an oracle holds this pre-image but only reveals it when a threshold price has been hit
  • the FROST signature from the Multi-party computation (MPC) network acts as a backstop against any malicious activity from the oracle and provides a second layer of authentication

The basic flow:

  1. A user opens a vault, deposits BTC, and mints UNIT
  2. If the value of BTC falls and their vault drops below 135 percent collateralisation, it becomes eligible for liquidation.
  3. Any liquidator can step in. No whitelist. No backroom deals.
  4. The liquidator recapitalises the vault with fresh BTC, takes ownership, and later repays the UNIT to unlock the collateral and collect their profit.
When the liquidator restores solvency and repays the debt, they receive discounted BTC in return.
Okay, so you send some of your bitcoin to an address and then they send you some stablecoin called UNIT, which I couldn't find much info about. Their docs for it are limited. There's this philosophy page. And this supply control page. It seems that UNIT tokens are runes?
The address you send your bitcoin to is kind of like the addresses used by Spark: it's a shared utxo that relies on FROST to give stakeholders shares of they key. I think this is why they claim that it's self-custodial a la this from their website:
Deposit as much BTC as you'd like to leverage within one of our vaults, and borrow UNIT against the vaulted BTC collateral.
  • Non-Custodial
  • Full Control of Funds
It may be non-custodial, but calling it full control makes no sense. If you can't unilaterally exit, then it isn't full control. And who in their right mind would give you full control over the collateral you used to get a loan? It isn't collateral if you can go spend it.
Lol so what do you do with UNITS, there's no liquidity? Even the biggest on-chain protocol that took this concept to the max makerDAO still had to hold USDC and build out massive liquidity pools to make it work adding more custodial risk on it
I honestly don't see the point of these borrowing protocols, just spend your sats or go to regulated lender and take that custodial risk on with a company you can sue for a partial repayment if it goes sideways
reply
510 sats \ 1 reply \ @cmd 15h
It may be non-custodial, but calling it full control makes no sense.
The definition of custody has been bastardized by marketing. True non-custody means you and only you can withdraw funds at any time. Anything beyond that, there is some shared custodial relationship.
And who in their right mind would give you full control over the collateral you used to get a loan? It isn't collateral if you can go spend it.
I think people are afraid to call their project anything other than self-custody. Nobody wants to hear the boring details of your contract. So you just say the magic words of "self-custody" and people nod their heads (hello ark).
Okay, so you send some of your bitcoin to an address and then they send you some stablecoin called UNIT, which I couldn't find much info about.
It's essentially a rune. The protocol for runes is actually quite simple and elegant for alternative assets. They are still a shit-coin though.
The address you send your bitcoin to is kind of like the addresses used by Spark: it's a shared utxo that relies on FROST to give stakeholders shares of they key.
The loan contract is a 2-of-2 multi-sig between the borrower and the FROST key, with an extra spending path (for liquidation) that is the FROST key + pre-image from a price oracle. AFAIK spark setup is more like a FROST key wrapped in a 2-of-2 musig key.
One big question I have with the bitcoin loan companies is how their oracle works. Even if the liquidation process is 100% automated by some on-chain script, the oracle is the weak link.
All lending protocols rely on the multi-sig and price oracle not colluding. There's really no way around this. You can try to bury the problem under complex multi-sigs or staking / slashing mechanisms, but you end up with a cluster-fuck solution.
Unfortunately there is a lot about the protocol that is not in the docs (yet). I know all the dirty secrets though.
reply
Thanks for your clarifications/corrections.
I do think that hodlers will eventually decide that don't want to leave all their capital in cold storage. Especially if tax structures remain the way they are, so the lending protocols make sense.
reply
102 sats \ 0 replies \ @patrick1 18h
This is cool. It's basically an implementation of the Liquity protocol from the EVM world but on bitcoin. Sovryn has something similar (they call it Zero) on the rootstock sidechain.
reply