MakerDAO Know Your MIP, 11 Aug 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Know your MIP KYM#09 | Declaration of Intent - WBTC-B Collateral Type - August 11th, 2021

Description

For the 9th episode of Know Your MIP we’ll be talking about MIP13c3-SP11: Declaration of Intent - Onboarding a new collateral type backed by B.Protocol v2 2 with B.Protocol founder Yaron Velner.

Host:
Forum - juanjuan
Chat - juanjuan

Yaron Velner:
Forum - yaronvel

Original call with B.Protocol from December 8, 2020:
https://www.youtube.com/watch?v=CaDnKAlCpiM

Forum discussion about WBTC-B Collateral Type:
https://forum.makerdao.com/t/discussion-wbtc-b-collateral-type-backed-by-b-protocol-with-x10-leverage/5409

A

Hi, everyone welcome to know your mip number nine, uh where we get uh to know more about uh the mips, uh so the make improvement proposals that are actually proposed to the maker community and uh where we can discuss the mechanisms and and the team behind.

A

um In this case uh we have uh earlier on I'm part of his team. I'm sure he's going to do an introduction and we're going to be discussing how how big protocol would like to be integrated into maker and on the utility about that so yeah you're on night, and if you want to take it away.

B

Yeah yeah, thank you, john, and let me share my screen, uh so I also have some introduction slide at the end of the talk, but just quickly introduce myself. So I'm the founder and lead developer of the protocol.

B

Previous university of cairo network aitan is head of ecosystem.

B

And yeah, let me start.

B

Yeah, okay, so this meeps uh is for collateral onboarding, but it's basically still a declaration of intent, uh meaning you know this proposal is only to see if the dao really wish for us to move forward with this.

B

You know which is sold. So yes also. So what is the problem? You know not only in maker daw in all of defy learning platforms is that they have still very high liquidation ratios.

B

So you know you.

C

B

You want to encourage conservative users, and you know- maybe there is this misconception- that a low liquidation ratio you know is only like for the digions uh in reckless people.

B

However, the fact is that you know if the liquidation ratio is a lower than users will actually get liquidated, less and meaning the actual high liquidation ratio actually makes life harder for the user, because he he has to maintain. You know highly.

C

B

Collateralized positions, you know so just to you know like discuss numbers right so right now, most of the liquidation ratios in mecca are in the order of a hundred fifty percent, meaning that you know like on conservative user uh will typically be at the order of two hundred percent of collateral ratio, meaning that he has one dollar of debt. You have to have at least two dollar collateral.

B

uh Well, if you lower it down, then you know like a typical user will only have I don't know, like maybe 50 percent of the collateral and still not get liquidated when price changes.

B

uh So you know so you users in general, don't like higher liquidation ratios, and you know it has direct effect on die circulation. How hard it is to mean die, manage your die position.

B

It also has a direct effect on the fees that you know a plot for a platform not only make your dao, you know, can capture from the user because, like you know, the service or the product is less attractive and in general you know the result is low capital efficiency, both in maker- and you know, oliver b5.

B

Yeah, okay, so now why you know why liquidation ratios are still so high in d5 in general, so you know the original ratios were set somewhere in 2018 19, where you know d5 was highly e-liquid.

B

These days defect is very liquid. However, big part of the liquid, the liquidity is in a constant product, automated market makers such as sushi swap or unisop v2- and this is just a snapshot from today. So if you go to sushi swap they have 260 million dollars of liquidity, uh but practically you cannot even make a 10 million dollar trade uh with a decent slippage.

B

uh So we see here that this 10 million dollar trade will result in seven percent slippage. uh Okay, so social swap is not the only one. Of course you have other markets and you can go from usdc to die, etc, etc. But you know the the same. Assumption holds that automated market makers, you know, have very low capital efficiency, uh okay. So now what about unisof v3 and maybe also curve v2? You know so now there is a trend of offering a concentrated liquidity market makers.

B

So so, in this platform, such as unisoc v3, you know we see that if the market moves a bit, the liquidity will get depleted and you know whenever you have many liquidations, but you know like by definition almost this is when market moves or or maybe the other way around right. I mean market moves, so we have a lot of different liquidations and when this happens obviously liquid, you know liquidity obviously immediately gets depleted.

B

So you know here, for those of you are not familiar. You know we see the graph of liquidity in the uni swap is director, and you know we can see. You know like the high spikes around current price, but it's not clear what will happen or I actually know, based on the past experience. It is clear what will happen when price move.

B

You know at least temporarily it will get depleted and then you know it really depends uh on how quickly liquidity providers will set up the new orders and typically, when market moves fast, you know they're reluctant to put limit orders or put any liquidity there.

B

um Okay, so because you know so, you cannot take it directly from existing decentralized exchanges. uh So yeah, like the best taxes today, is to keep some network of boats and keepers.

B

You know, so you encourage them because you give them some incentive. Some even have you know some, you know maybe liquidated liquidators and keepers or on the project payroll, uh and you know like by products of these. uh Some of them are not committed. You know the one were just there. You know like to take the easy liquidations, but not the hard ones.

B

It gave rise to a lot of mev uh in in, in terms of you know, like transparency, you know. So if you have like a project like ave or others, you know that might have some liquidators, you know, might have some agreements with the liquidators etc. Typically, these things are non-transparent.

B

You know, if you go to the dao, and you know you, you ask the doubt to make some decision based on some risk analysis. While these agreements are non-transparent- and you know.

C

B

If they were transparent, I mean in this world we try to minimize dependency on. You know like sign agreements, so.

C

But but given this.

B

Non-Transparent, uh you know settlements. How exactly can a dow make you know like a transparent or educated decision? You know on risk parameters in the protocol.

B

Okay, so, but do people really need, or you know, users really want higher liquidation ratios? Then yeah I mean if you look at the traditional finance. You know traditional crypto finance uh in volatile days. You know traders are willing to pay over 100 percent. You know like yearly api or early stability fee, just too long ecusd, which is basically kind of the same service that maker offers.

B

While you know in mecca, the current interest is around three percent and in ec, for example, it's only half percent. uh So you know it's not like with the protocol tomorrow. Maker will be right, I mean as a popular by bit and others, and it's not like you know. We plan to immediately offer x, 100 leverage, uh but there is a utility, for uh you know, for a better quarter ratios.

B

Yeah, okay, so, okay, so like to recap, you know, like the the first part of the presentation, uh so you know liquidation ratios, I think, should be and can be improved in defi and there is a demand for it. There is, you know like a market for it, an ability to get a.

D

B

B

Okay, so what is the solution that we propose?

B

uh So we propose user based backstop amm, which is you know, so use the amm term, although it's not really automated market maker, it's more something that handles liquidations. So in this setup, users pull funds at all liquidations, and you know in in that example in mecca. It will be die, so there will be a pool of die aggregated by user deposits.

B

Now, whenever there is a liquidation, uh you know instead of using a uniswap or sushi like formulas, they will just be willing to absorb any amount.

B

You know, depending uh you know, up to their inventory size uh with a fixed slippage, for example with five percent discount, meaning that you know if you go back to the uni swap example. Sorry, the social swap example will uh right now sushi swap with.

C

200 million.

B

Dollars of liquidity cannot handle even 10 million dollars. You know like in one shop, so here with 200 million dollars, you know we will be able to absorb you know the entire 200 million dollars amount, uh so the 70 is much higher. Capital. Efficiency is much higher.

B

uh Okay, so like taking the liquidation is the easy part. But now you know users deposited the die and now they have ease, or you know whatever collateral we have. There is so wbtc, but now now the novel part from technological perspective is how to sell. This is back to die.

B

So here we present a new algorithm, uh so we have is now and we need to price it. uh You know, according to formula so here we built on top of a finance stable, swap invariant, although outside is not really stable to apply. It's, for example, die to ease but to use a price feed in order to normalize all the inventory values.

B

And then what is nice about kel formula is that you know it offers different slippage.

B

uh With respect to how imbalanced you are so here, you know if you have a lot of ease, meaning we are very anxious to get rid of this is uh so we offer a higher discount. You know over the market price we'll try to get rid of it faster, whereas if you know, if you have like only one or two or five percent imbalance, we will try to sell at market price.

B

uh So this is completely automated a formula uh governed by a smart contract and.

D

B

Is like the main novelty you know of our system.

B

uh Now, diving in a bit more, so the imbalance is one. You know like one concern of the user, how to automatically rebalance- and we are the first ones who offer solution here. uh Another concern is, you know how much liquidations are there anyway? You know what kind of apy I can get for depositing.

B

So you know in the real market, you know in binance, futures bitmex, there's what there's quite nice yield. You know there are many liquidations and this is big, but you know this is because they're still very much ahead uh of d5 right, but you know if I eventually would go there, so they would. There will be nice yields.

B

uh However, you know to start with, uh so what we offer is actually to put you know like the idle funds in some the idle funds in some protocol that offer interest rate- uh and I think you know like initially for maker- it would be easiest to put it in the dsl, which is you know, like a major compound module that offer very low, almost zero interest rate at this moment. But you know in the future, it might offer more.

B

You know, obviously, initially we will subsidize it a lot with usb protocol, so there will be some yield uh and then over time you know we can consider instead of just using the dslr to use, maybe wifi or other platforms uh which are might be a more risky, however, offer higher yield okay. So we have a pool of dye. You know obligated by or deposited by many users, uh and now uh what we offer is to change. uh You know for the collateral for this collateral. We try to onboard uh to change the liquidation process.

B

So right now a maker has a double component that whenever it bites or actually bulks, uh you just send it to some dutch option uh so in in in our proposal. Instead of sending it to an option, it will use the inventory uh user pool okay, so it will take the dive from the dslr, the guy, that users, you know only users who deposit it via b protocol, not any dslr deposit.

B

You will use it for byte, it will take ease according to a discount according to a price feed like with the medianizer price feed of maker uh and now, okay. Now there is instead of die, and you know the rebalance algorithm, the kell adapt we'll just sell. This is on the market and it will eventually get deposited back again to the idle module.

B

Everything is done automatically. You know without any human or even both intervention, so fully automated, fully decentralized.

A

uh Yeah, I'm guessing that the user can actually like see their their. I don't know the share of the pool remains the same, but then the the die would increase as the liquidity gets.

D

A

There's some liquidations, uh so you would see your your your die go up with time right.

B

Yeah yeah yeah, I mean like the expected or the desired behavior uh yeah. Is it like you'll die yeah, like you know, your shell will remain the same, but amount of diet will go higher, but actually I see there's a question from frank uh yeah. So okay, so you know there are risks here I mean so we have you know. So. Let's say we did the liquidation with 10 or 5 percent.

C

B

It could be that by the time we sell it market will continue uh going down. uh I will share some simulation results, so you know we did simulations. So sorry it could happen and there is risk here, uh but you know, there's no profit without risk. uh It is important to note that you know this. Risk is for the users who provided liquidity, not necessarily for the maker protocol right uh now. Okay, so so that's actually very different from yeah.

B

Okay, so that that's that's how we balance you know, there's no guarantee that price will not move. You know we have some simulation. That shows that in average or not only an average, we simulated on very volatile uh periods of time.

B

This is a profitable but yeah. There is risk, but again for the liquidity providers, not for a not for the protocol itself and how to answer paypal question.

B

So if you just like the liquidation and penalty, then again you arguably give more incentive to the keepers right. But still you don't know, I I mean take black thursday as an example.

B

Part of it was because you know the system was very complex but part of it. You know you had like 10 million dollars incentive to liquidate and people still liquidated with zero. uh You know someone liquidated with zero price.

B

uh So again, things are very different with the there's the liquidation to oh, but it's still to show that you know. If you have only keepers, uh you know which are eventually so we have few bots out there, so they may or may not. You know execute, uh but but it's even worse than that, because, okay, so uni swap you know if you take sushi swap so okay, so I showed seven percent slippage for 10 million dollars.

B

So you can say: okay, let's you know, let's put 10 10 percent penalty, but but this doesn't scale right I mean so. If you have now 50 million dollars for liquidations, it will roughly scale proportionally.

B

So this doesn't scale I mean define liquidity doesn't scale because it relies on automated market makers. So, even if you do offer, you know very high penalties, you know, eventually, socially social liquidity will get depleted arbitrage. Bots uh arbitrage bots will not be able to do arbitrage. You know if you have some.

B

I don't know very big traders with a lot of inventory that will be willing to absorb it. Then this could save the day right, but I don't know if you really have it.

E

Well, maybe just to mention that you know taking uh liquidation penalties. You know uh higher is a bad in a way user experience for uh for uh most maker uh users, so.

B

Yeah yeah also for the long run it will. Actually uh it will. Actually, uh you know it's kind of defeat. The purpose of having uh you know. Eventually, we really want to go with higher with better liquidation ratio right, so you know so uh making the liquidation penalty higher really gives. You know like upper bound on the liquidation ratio. You can put uh so okay next, instead of testing it, so I will show some simulation results. We.

C

B

Live with similar system outside maker dao, I will touch it later, yeah, so to answer primo's uh decision uh question. uh So, first of all, you know the 110. It's not final, you know it's not really the actual number. We propose it's declaration of intent and you know like the idea is to discuss it with the relevant team if the proposal goes slow uh but but yeah so like there was some discussion on the forum we said kerr and also talbanet from the protocol engineering team.

B

So it really depends. You know like, for example, if you start with 120 130, then yes, so the byte could still be according to the or the back would still be according to osm price, and only the rebalance can be according to the medianizer, uh for example. So you know this will so it will probably not allow us to go to 105 or 110, but definitely should be good enough for 120 or 130 for wbtc, for example.

B

uh Also, the idea okay, this might also uh answer partly to planet x. Question is, you know when to really start small as an experiment with very low ceiling, uh but yeah? Also again, I will touch and you know explain the simulations we did but, but eventually you know if you do have one million dollar buffer, for example, you know so up to maybe smart contact bugs there's. No doubt that you know we can liquidate one million dollars uh and then you know what should we put to?

B

The test is actually the rebalance process uh and there are ways you know so I also have a slide on how to mitigate okay. Let's say the imbalance process didn't went as well as we expected, uh so we can mitigate also that.

B

Yeah, let me read another question.

B

No so papers no.

C

B

Does not really give a monopoly, but rather a make it. Our protocol give a monopoly for a specific collateral type. You know like for a new collateral type.

A

Yeah, so people in birmingham is asking if, if the correct way to think about this is that the user gives me protocol a monopoly on any liquidation of their vault. That's not what I understood. What I got is that you're participating in open liquidations right.

B

No, no, no! No! No! So the idea is, you know so opposed to everyone in individual. So you know this collateral.

B

This collateral will have a amount so in this collateral protocol will have a monopoly, so yeah user that will use this collateral will give a monopoly for the liquidations.

B

uh Of course, you know like as a fail, safe measure. You know if the buffer gets depleted and then then it can go to the standard liquidation process, but but then you know, then it's also riskier. So you know we'll show some simulations and mitigation steps. Some. You know how to make it more secure, but basically yeah there will be or you know we do ask for a monopoly for this collateral type.

A

B

A

Know is there sorry, the last question is: is there any way of maybe instead of putting the dye into the dsr, putting it somewhere where it yields a bit more.

B

Yeah yeah yeah, so technically definitely one second. My kids are big.

A

Yeah so technically.

B

Definitely- and you know like in in the future, instead of the dslr, you know tend to use wi-fi or other protocols, uh but you know there's like a trade of fear on how uh you know like what is the trust level of you know, like maker dao has for these protocols, so really offer that you know like we start with the dslr, and you know we will subsidize it a bit uh but yeah like for the long run. You know put it in wi-fi. Maybe in compound could.

E

Make more sense, basically, it's easel funds. You know just waiting because liquidation doesn't happen all the time, definitely not with all the amount which is you know being deposited, so you know just keeping the this pool. You know just sitting idle with no and and currently the dsr is very low uh interest rates. So there are, of course, as don't mention other options, but you know if we want to test it uh kind of you know in a gradual way. So maybe this is something we can have. You know later, yeah.

B

uh Yeah, okay, so.

D

B

Asking if it's fair to say that the vote maxing that size uh related to the diet deposited.

B

Yeah, okay, so so we can, we can show you know so yeah. It is related and you know obviously not one to one but linearly related or maybe super linearly. uh No, but but we can.

B

We can actually guarantee that, because you know the idea is that user will lock the funds for some time period. uh We could have some, you know we could block or withheld withdrawals and then the only way that the dye in the buffer and the dye quantity will go down in the buffer is only upon liquidation. Events now by definitional liquidation event uh will also decrease the total debt for this collateral and.

A

Then you can also have.

B

That you know during this liquidation we will so so this is one way or one thing to consider that during this liquidation uh the ceiling will be automatically reduced for this collateral. So you know, let's say we had 10 million dollars in ceiling. Now we have one million dollar of collateral.

B

Sorry, so we have one million dollar of a debt.

B

It was liquidated, so we can uh put it in the system that you know that now the ceiling will be down by one million dollar er and until you know, until it gets rebalanced, no more restorals, no more borrows are allowed on the dive side.

F

I I actually have one question, because I'm not sure if I'm understanding this correctly, so your service basically consists of two what you should say. This backstop keeper basically consists of two types of services, so you kind of guarantee there are fonts available for liquidating this debt, but you also have this b protocol amm, which is uh you know good, because it has this high amplification coefficient and it makes uh minimum slippage when you liquidate it, but just to be sure, like these are two separate entities or ports within the protocol.

B

C

B

So yeah so like we have like four modules, so yeah. So what.

B

Holistic system, but yeah there's one component, which is the ideal component which pulls the user funds and then yeah. The another component will rebalance and will deposit it back to the idle found idle yeah. But.

F

But this uh the balance- this is this only rely on on liquidity on bmm, no.

B

No, no! No! No! No! No! No! No! So the balance is actually the bmm or no okay, so this might be confusing. No, no, not at all so so. Users deposit their die here. Okay and.

C

B

You know you can see the bam as this. You know all four components right so you know, let's say user- put 10 million of dye here and now you have one million dollar one million dollar liquidation.

B

uh So the die is taken. You know by the maker protocol from here and gives ease to the reverence component. Now this is will be solved in the open market, not in the amm. Okay. So like the high capital efficiencies, okay,.

F

Because you know.

B

We just give you as much dye as you want. You know, according to some fixed slippage,.

F

Now I see at first I thought you're building uh special liquidity, pools that we're you know uh reassured lowest and so on, uh not just from technical perspective, but also from liquidity.

F

That's why I asked you, you know: should the dev size be related to the liquidity of this of this bmm, but now we see it's only.

B

Never, but so so, but that size should be related to the size of you know this idle component right sure.

F

B

Okay, yeah yeah, so this we said on there I mean this will just offer it for sale in the open market, uh yeah. So, okay, so some simulations.

D

Yeah, sorry, I have a question. So basically you uh you guys are providing stability, pull as a service for maker right.

B

D

The thing you, uh the integration you had with liquidy, where you use the stability pool to get ether when uh through liquidations and then you're using, let's say, carve or unisoft v3 uh to swap it back to more dye and feed it back to your stability pool. So instead of getting eat you get more dye. Is that kind of? What's going on.

B

Yeah yeah definitely so I was not sure I was allowed to mention that you know this name here, but yeah definitely I mean so it's something very similar to what we do with celiquity uh on the liquidity already has this notion of stability pool uh also, you know liquidity originally or you know. Without us, it's not really end-to-end.

B

You know like a closed system, because so but yeah what we offer iniquity is a closed system, and this is exactly the same concept. How we try to offer. You know to other platforms, yeah exactly.

C

And you incentivize uh folks to deposit lsu l, usd or dye uh by giving them, I guess p. Protocol tokens is that right.

B

No, so so, technically now in lusd, the incentive actually comes from liquidity, but yeah like for for maker. You know like initially to what bootstrap we will give. You know token incentive over time. You know we hope to we. You know over time they will have to be incentivized somehow.

B

So you know we imagine that you know as defy scale. There will be actual profits from liquidations. You know as diff, eventually the either the dslr will give higher yield, or we just put it in wi-fi or compound. So so yeah I mean over time the sorry. Initially we will incentivize it over time. You know we see they will make actual extra profit. You know by depositing. So this will be the over the long term incentive but yeah. Initially we will incentivize.

E

It's worth reminding, you know that the heat that is being seized during liquidation is discounted. You know you get more heat than what you know was used with the with the dye, so eventually again, if sold, market price or even lower still will provide. um You know benefits for uh for the depositors.

B

Yeah yeah, okay, so pepper, there's. Another question: is this dice float guaranteed to be used, uh yeah yeah, so so this you know like yeah yeah, so whatever is in the ideal component is available always for liquidations, uh but now regarding the price. Okay. So how do we sell it? You know, so we can dive in uh into the selling algorithm.

B

So so, first of all you know we always we don't buy it, we don't buy it anywhere. Don't you know we don't trigger any transactions at all. We simply offer you know. Okay, so we have is for sale in return to die. This is our price, which is quantity based.

B

uh It is encoded in the smart contract and now everyone can buy at this price, uh but practically you know in liquidity. We also aggregated index aggregators. So this is. This is just you know like another order in the big defy order book.

B

uh However, you know, like the properties we try to guarantee. uh Is that you know so we we typically will sell it below market by site. So if someone wants to buy die, you know we will typically offer the best price, uh but now to encourage you know like a big imbalance, so you know the bigger the imbalance we have uh the higher discount. We will offer okay, so up to some cap, you know also to prevent some attacks, etc.

B

uh So you know like initially we just selected market price. Now, if the imbalance is very big, we'll offer it. You know like, for example, right it's one percent discount on market price, and it could go all the way up to four percent discount if the imbalance is very big technically again, there's no guarantee for profit here, but simulations were very promising, uh and you know also- and this is also why we require you know like a monopoly- is that.

C

You know many of the liquidations.

B

Will not be you know like in extreme market conditions, so eventually the good liquidations you know, should compensate on the heart. You know on those who that were harder to do.

B

Yes, so so- and this is explained- you know in more details in a white paper, so in order to simulate it, you know so because eventually we aim for higher leverage here you know it was 105 liquidation ratio, which is not what we offer to start with. uh But but you know it's this amount of liquidations.

B

You know like we didn't really simulate it on compound or a maker dow, because you know like the vision is that eventually it will be able to support much more than the current status quo, uh so we simulated on binance futures uh which had that month. So it was on months of april, uh which had around 1 billion dollars of liquidations.

B

So april was less volatile than may, but may actually add fewer liquidations, because everyone got wrecked on april. uh You can see on the right hand, side that liquidations, you know very unevenly distributed.

B

You have a few hours here with 200 million dollars of liquidations, and then you have days with very few liquidations uh and now on the left-hand side. You know we deficit, two scenarios. One is that you know you have certain liquidity, which is only in constant product amnes, uh and now we can. You know we can use only that for liquidations uh and then the other other bars okay, what if we would have used the bam and we defined some module model?

B

uh You know for the rebalance, so we had very conservative model, we had very conservative model and namely that rebalance will happen only when there is an arbitrage uh on uniswork, where unicef has only 100 million dollars and the trade can happen. Okay, so here a trade can happen on uni, swap only once a minute, uh and here a tray can happen only once every 30 minutes on when we swap which kind of represents.

B

You know the time the default market will take to you know bounce back from the liquidity devices, uh and basically so you know we examined the couple types. You know three types of buffer sizes from 100 million to 200 millions.

B

uh So this show you know like the apple upper half. The upper line shows what we, what would have happened without us and unislope would have missed almost 50 percent uh 50 percent of the liquidations in binance futures, uh while usb protocol it was very close to zero uh to certain uh you know so, depending on the buffer size.

B

uh Also, here you know if you take one billion dollar of liquidations with five percent premium, so you know like the potential profits is in the order of you know like 100 million dollars right. So so so again, it's not like. We start by preparing ourselves to one billion dollar of liquidations right, but it is to show that you know if the defy at certain point would want to support it and there will be actual need for it. Then you know making 100 million dollar profit for 200 million dollar inventory per month.

B

Again, of course, it's kind of ideal now you're right, but there's a lot of a lot of profits there. You know conceptually, uh yes, another question about the monopoly yeah, so the monopoly is, you know, so. Monopoly exclusivity reality.

E

Might be the right one yeah.

B

Yeah, I mean maybe priority, so you know. Eventually we want to incentivize people to deposit in the ideal pool right now. So now, if you give them only the hard liquidations you know and market is really crashing, so they don't have much of an incentive. You know to contribute liquidity, but so by this monopoly or priority you know, so it's a bigger incentive for the liquidity providers.

B

You know and we can think of other incentives, but you know like I think, from political point of view. This will make the most sense- and you know this is like the most honest and fair scheme and of course needed to say that you know this architecture is open for everyone, any user can deposit and the user can take part of the liquidity providing.

E

Maybe also wolff mentioned, you know that um getting this priority is what we are aiming for is getting a vault. You know and either in it wbtc um volt, you know, with with lower liquidation ratio, which is kind of uh you know a unique feature uh in in this vault and is being enabled because of this um backstop that is being proposed by by the funds that the users are um providing. So it's kind of you know an um on a risk taking benefits. You know, rewards uh kind of an equation, so yeah.

B

Yeah, okay, so so so what is it? You know in the mip? Exactly so? First of all, you know it's declaration of intent. uh Also, you know declaring that you know like the final details will be eventually discussed with the team you know. Needless to say, you know the final implementation will also have to be approved by the by a dow vote, etc.

B

uh So, right now, what we are looking is mostly like for a green light to push it forward. uh So what we offer is a new collateral, and you know, like the exact character, will be decided.

B

uh You know by the community, yes, but like, for example, it could be another kind of bitcoin wbtcb, another kind of ease or any other. You know the collateral the community may wish.

B

Okay, so so it's a new collateral with better liquidation ratio, uh it will have higher stability fee right, so it can be seen, as you know, like a you know, so, allegedly there will be extra risk because of the liquidation ratio, although you know we believe it's quite secure, but the stability should compensate for it and, for the long run will be another source of revenue.

B

uh You know effort to start with a loaded ceiling uh in this. You know just to experiment with it see how it goes uh and yeah the bam. The protocol will be the exclusive liquidated or will get a priority. If it fails. You know it can always fall back to the current mechanism, which is quite easy to implement.

B

However, there is the assumption: is the current mechanism might not be able to handle such leverage such collateral ratio? uh So it's not something. You know like like a selling point, but more of like another another. I don't know like an another line of defense.

B

Yeah, okay, so there is a question on how important the yield is.

B

So so no so we don't believe you know no okay, so I think you know like in the very long run. I think the dslr is viable not because of the dslr is great, but because you know there will be liquidation profits.

B

uh However, for the short run, the ddsr is probably not desirable, uh but you know we'll subsidize it, and you know, let's just put to a test. You know the imbalance component and all the others uh and then maybe consider having you know, putting more risk by using y field compound instead of the dsl.

B

But this is something you know. We definitely want to improve there and also to get the community feedback and now they could they. You know like the care, the concrete uh proposal.

B

I think we explicitly mentioned this, but again, details are very flexible depending on what the community will want.

D

B

Terms of the fallback how the fallback will work uh so yeah, so we have the dog. You know we have like a another dog module for this collateral.

B

Now the check is very simple: let's say you know you have a half a million dollar a liquidation, so you can simply check. You know whether the ideal component has half a million dollar.

B

If yes take it, if no, you know send it to the clipper or send it to the dutch auction.

B

D

This is the fallback mechanism.

B

A and again it's not a scenario, you know we want to get to, but yeah it's like another line of defense.

B

Yeah one second.

B

Yeah, yes, okay, so that you know nothing is risk-free in life, especially you know when there's potential profit to be made, uh but but it is important you know like to under to understand- uh and you know like to properly categorize what is a like: a user like liquidity provider risk and what is like a maker dow protocol risk right, so the the user. Might you know the liquidity provider who deposits and then it goes to the dslr, might not might have trading loss, which is fine.

B

You know the oracle uh might go crazy, which will result yeah with another.

B

You know trading lost, uh but you.

D

B

The face of it, these people, they get comfort and they are compensated, and so so you know it's a user risk.

B

However, it's not really, you know like it's not fully uh a quarantine or separated from the protocol risk, because eventually you know like the risk from the maker point of view. Is it because of trading losses? The ideal component, the liquidity pool, will get smaller.

B

uh So so here we offer some mitigations, and you know this is what we proposed in the discussion in the forum, because the protocol engineering team had some concerns uh so yeah a mitigation could be to auto, adjust the debt ceiling according to buffer size. So in order so let's say initially people deposited five million dollars.

B

Now there was, I don't know, trading loss, yes and now there's only two million dollars for some reason, so uh so the ceiling for this collateral type could uh be automatically in lowell and yeah. So these are the potential risks uh yeah, okay, so about to be protocol. So we are live from october 2020.

B

You know we actually had maker integration, which I didn't mention, because it will mostly confuse you guys, but this is still live.

B

Okay, but so sorry, so initially in the protocol in everyone, we tried to encourage professional liquidators and got around 270 million dollars tvl at peak uh and back then we also proposed some kind of professional liquidators. uh I know union or protocol for b for maker dao.

C

B

We withdraw, we withdraw this proposal and now we actually aim. You know like to make everything more democratic. You know give more power to the people also, because we believe we can get. You know like more liquidity. You know like from the people rather than professional liquidators.

B

So this is our second version uh where we offer user-based liquidity.

B

And yeah and as it was mentioned before so similar settings, but you know not identical, is already deployed, you know so with liquidity. You know the project which I'm not allowed to say the name.

B

So it's actually live for only a couple of weeks, but although there's 30 million in deposits, I didn't put it in the slides, but there are already some rebalance transactions. I can show you uh you know like for small amounts, but still in this rebalance. What actually happened is that you know there was some user just he wanted to buy lusd.

B

So we offer the best price so palace walk. Aggregato, uh sent him to our ba.

B

Yes, so that's that's us uh and yeah. If there's any more questions, we'll be happy to answer.

C

A

Yeah pretty much! Oh sorry, god.

C

Oh yeah, just a quick question. uh Speaking of liquidity, um I saw that uh in one of the write-ups they say that uh uh you guys are using the chain link um price feed.

B

Yeah yeah, so so that's yeah, so that that's for liquidity and because also liquid is using the chaining price feed for maker. You know, eventually, it will be decided by consultation with the oracle team, but probably would make more sense to use the medianizer not changing, but.

C

B

We do use chaining.

C

And I think, as a backstop, uh real quick point, uh you said you used gelato and chain link right as a backup. uh How would it work here.

B

Yeah yeah yeah, okay, so so this this is, like you know, like the the core protocol, uh and now there's you know like the periphery right. So so you know like the code protocol, we just offer it for sale, and you know we offer a certain discount uh and this this discount uh become, you know, gets bigger as the you know, as the balance gets bigger. So what we do we actually, you know deploying keepers by gelato and by training.

B

They just say that you know the discount is so big that you know there is an arbitrage already with unisop or whatever, so they just execute the arbitrage. So you know it's an either safety measure. It's very likely that you know arbitrage boards will do it much sooner than them right, but it's another safety measure.

B

You know that's more like in the periphery, not like kind like part of the core protocol, but yeah. It's definitely something you know. We will uh do also here.

E

It's another service that uh chaining is offering it's not in a rocker, it's their new keeper services that they just launched.

B

Yeah yeah yeah and again it's certainly the core of everything, and you know it's very likely that arbitrage bots will take it much before they do. But you know like.

C

It's another safety.

B

Measure so we also used it.

B

Yeah yeah, okay cool, so I I know I can, if you can quickly, maybe paste a link to the form discussion.

C

In the chatbot.

A

You're on regarding the the next steps, are you guys planning on going into this uh next cycle, or are you planning on waiting a bit more for comments? No.

B

No, no so the vote is already live. I think I got already 7000 mkr in favor. No, no, so we are in the current cycle. uh Okay, some other thing not sure I mentioned, but you know. Obviously all the development is on outside.

B

But you know some consultation with the core units will be needed, but you know development and we will push the development and and implement.

C

B

A

B

Yeah and and the vote is live you know like for, I think, there's still 12 more days for you to go.

A

Good are there any any questions we're almost at the top of the hour, so I want to be conscious of everyone's time. There's also another uh yeah. Another call in three minutes with the nip sticks for real world lasted, so I'm not sure how many of the people here want to attend. The other.

F

A

F

uh I just want to mention so from risk side. um I know the original intention was to lower liquidation ratio of these vaults. I think that's still the whole idea, apart from improving the you know, liquid efficiency or auction throughput, um but it really depends on the oracles when it comes to two barks. So not the not the medianizer.

F

That's used settling auctions, but when uh you know triggering them- and uh this is something you guys would probably need to discuss with oracle team, because if oracle prefers to use uh so the delayed uh price, you know the one-hour delay to asm.

F

um This is like a problem if you want to start with- and I think you you actually addressed this earlier in the call- uh if you want to start with 120 percent liquidation ratio and having one hour delay, this could be already an issue, maybe not as much for bitcoin, because it's less volatile but uh yeah. Ideally, if, if we use medianizer even for you know for triggering options, this would be much more smoother, much more efficient uh but because maker, I believe, somehow prefers you- know oracle safety.

F

um This becomes then a bit of a friction point. If you, if you want to go.

B

Lower on the liquidation.

F

B

Yeah yeah yeah, so that that's a good point- and you know like in the declaration of intent yeah, so we do specify that it will be determined uh with the oracle team, uh but you know like even just to start with you know, even hundred and thirty collateral ratio, for example, for up bitcoin and wbtcb. For example. It's still something uh that I think it's not available at this point right. So.

D

C

B

Start, like very very you know like flexible, you know.

F

I mean for sure, because, because the whole- it's not just about uh you, know how frictionless auctions are but about committed liquidators. So that's you know it's always a good point. It increases our auction throughput and reliance and liquidity. So you know this is definitely good.

A

Nice um yeah just for for anyone watching this offline uh you're on. What's the best way to contact you, I guess the the forum you're quite active there or do you prefer, I don't know telegram or discord or.

B

Yeah, so you know we have a community not.

D

B

All yeah I'm available at the forum. We have community discord. I know okay for actually for offline. It wouldn't really help in the future, but yeah.

C

B

We have a telegram hey, we have different twitter channel discord and yeah available everywhere.

E

But for any specific questions, um specifically on this meep, you know there is the forum, a proposal that that we posted- and you know any comments over. There will be replied.

A

Sounds great, so thank you guys for presenting it was really good and informative and thanks everyone else for for coming.

E

Thanks juan for having us here thanks, everyone.