New Ethereum talks, every Monday. The week's conference uploads by event, in your inbox.

Loading player…

Risk Management in DeFi Nikita Buzov

Ethereum DenverSun, Oct 9, 2022, 12:00 AM

Speaker

Nikita Buzov

Transcript

what is this i bet you all know that's a condom do you guys know what it's used for i bet you do right so what does sex and defy have in common it's risk right so today i'm going to tell you about d5 risk management and how to hedge that risk how to control that so you don't get wrecked not so much to talk about sucks i guess but all right so people use defy we all know we're all here for it we see the growth we like it the industry is growing but the toolbox is still half empty there are tons of tools we need to develop and regarding the risk management i think that's one of the spaces where we really lack some efficient tooling some ways to hatch the risk you know we give people access to all these financial tools which are awesome some of them use them for the first time some are experienced in the traditional finance but we need to give people efficient tools to hedge the risk so they don't have to worry about losing their money getting wrecked getting hacked getting liquidated and like solve all these problems so with that i want to kind of like i guess give you the list of all the risks i guess that exists kind of now i'm sure you're all familiar with most of these you know exploits being a huge one nobody really enjoys the impermanent loss they're rock pools and scams and hackers all that good stuff so what are the ways to hedge them these are the existing options you know they're not perfect uh but you know you can get a coverage on some of the platforms uh for your exploits for your hacks um stable coin to peg is pretty much in demand and within permanent loss you can kind of hatch that with options a little bit um i guess i mean permanent loss is kind of like a quadratic thing and then options you have like linear returns so you need multiple things kind of like to pack into this strategy and it still doesn't perfectly you know follow the curve and then i guess their market liquidation which is mostly just price race action so you can use derivatives if you know perpetuals futures you know all this stuff um and then yeah i mean the rockpool scams do your own research like aping in doesn't always work out um however like another big thing is the whole like infrastructure front-end attacks um and with that like it is a tough problem there's no really insurance out there uh you can't use options to hatch that risk so right now you can check the links check the data but until we really decentralize a lot of the infrastructure and have kind of like fallback mechanisms that's really tough uh because it pretty much just relies on the team on the people and individuals that are off chain right like we all have our laptops we can lose the key so you know just don't get wrecked but the vulnerabilities in the ways you know you get exploited are listed here um i try to kind of like bucket them into like a few categories um i'm sure you all heard of like flash loan attacks and how you can basically manipulate the market price or like the computation inside the smart contract uh to withdraw like some value extract some value their math errors we're all huge fans of copy and paste protocols right that just miss one line or try to tweak it and then they screw up the math uh forking might sound super easy and sometimes it's great that's what allowed the industry to go so fast but yeah like if you copy a protocol just just check if the math works out so with these you know i bet this slide has the most zeros out of all the presentations here but we lost tons of money i i bet nobody likes that and these only include just kind of like smart contract hacks i had to filter out oh like the bridges i'm sure you heard about poly network like over half a billion lost uh you guys heard about wormhole that just recently happened like i think it's 300 something um so yeah i mean the the magnitude of this problem is huge but the d5 insurance is not really catching up to the d5 tvl there are tons of challenges with it and they're all kind of like outlined here uh renat i think he presented on saturday i believe um he's uh the founder or co-founder of ether riggs so he also talked about all these issues right now that all the d5 insurance protocols risk management protocols face so one of them being underwriting which kind of like underpils the underpins the whole capacity the demand is not an issue supply is and so right now underwriting is kind of like this high risk low reward thing right nobody wants to [ __ ] do it like who wants to take all that risk and pay out that money anyone definitely not me um so yeah like we have the problems here like capital needs to be locked like if you look at the traditional insurance space you usually sign a contract for like a year or like 18 months and you get like some kind of like fixed return um so these things don't quite well transform into like the d5 paradigm and so then you want to do partial reserves you want to be leveraged one dollar backing multiple dollars of risk that's how the efficiency goes up but then the second point in efficiency it all kind of centered around this idea of risk pooling the more risk you pull the more diversified is the more you can underwrite with the same dollar so you really need to like bootstrap that pool of risk that capital back in the entire thing then i guess like there's an issue with access just you know kyc for some reason some citizens are not allowed to use some protocols which i think is dumb we don't want to do this but then there's also like expensive coverage sometimes you know you have to choose the term of your policy there's little flexibility and then another one is risk management so some protocols use voting staking um but i think there's lack of transparency and then you also can't really use wisdom of a crowd for managing risk like we human beings are really bad at evaluating this on top of our head you need quantitative modeling you need numbers you really need to kind of like dive deep for a long time to be able to do these things and so but as a user like what do you care if you have an insurance if you get screwed or your car breaks down your house burns down you want to make sure you get the money back right like that i mean that's the only reason you bought that policy in the first place right now in like d5 insurance you have long times you have committees there's no transparency um so you have like all these like hurdles to the point where you're not sure whether you're gonna get the money back even if you didn't do anything wrong but then like from just i guess the business side of things obviously like insurance is quite far from point of sale right now um and generally none of the projects can get insured you know we're a protocol i'm sure there's some founders here too um how can you get covered your entire protocol doesn't matter of the size like even if it's a small one and you don't need a lot of underwriting capital you can't and if you're like big or like you know six five billion worth how are you going to get insurance so with that i just want to introduce what we did um disclaimer we pretty much solved every single [ __ ] problem here so i'll start with the first one high risk low reward we've been thinking for a very long time about like tokenomics and like what it really means and how you can utilize this novel mechanics to help solve the problem instead of deploying the token to raise liquidity finance the ops or just create hype really wanted our token to serve that exact purpose of underwriting and so with that token with solas we're basically decoupling the risk of underwriting away from the reward solus does underwriting with its own pool of liquidity and so you as a holder get only exposure for the underwriting reward you're carrying no risk so if the claim comes in and you know we have to pay it none of your funds are at risk however you're still earning all the revenue from policy sales from investment activity on the pool uh from like any emissions of the token um so that's fundamentally how we approached uh this problem and then also like you can underwrite in a lot of acids basically you just bond them you get the solids back so if you want to start getting exposure and earning generating revenue uh you can either bond it by buy on the open market and that basically drives the whole underwriting but you don't carry any underwriting risk only price risk so just like with any token it's publicly traded so it's mostly i guess driven by the the sentiment and the whole thesis behind the project then the low risk pooling solus is the only protocol right now out there that pulls the risk inside each and individual policy on the chain across all the protocols in cross chain right now we're live on mainnet polygon aurora but we're quickly like scaling up and the whole architecture from the very start was developed for this compatibility so it doesn't matter which token sits on which chain it underwrites the entire thing so you can think of solas existing above every single chain like as a separate protocol and we have an interfaces on chains so yeah and then also i guess like on the policy product side we want to have a reality where you can have one policy on any chain you like that will be covering your entire liquidity everywhere no matter what chain you go i guess that works well with like evm because you have the same address so what we did with risk management is pretty fun i guess our models are the only public ones out there and we don't use voting it's pure quantitative qualitative analysis we build the entire risk pipeline for the protocols to feed it for users to feed it and we ourselves are working every day to pump data in and generate data so we can efficiently and intelligently underwrite and make sure that we always have the funds to back the risk but also the product and the users have access to the policies then i guess i mean on the claims i kind of said the problem um yeah i didn't update the slide a little bit okay um but yeah anyways this thing is simple uh the best way to do claims is no claims you know we don't want to put you guys in the position of adversary you know like when you broke the car and you file a claim now you're kind of playing against the in the insurance company we do not want to do this i mean i generally dislike insurance industry just because sometimes i don't even file a claim because my time worth much more than that payout with us it takes one week and you don't need to do anything because the data is public the data is available we track we build the system to evaluate your financial loss and we send you the money if you're not happy with that obviously you can come in and dispute but fundamentally we don't want you to have any pain filing out of the claim and then um here's some alpha so this product has not been publicly announced or released but projects can get insurance right but now you can so solas native allows every protocol no matter of the size age or whatever you do get covered with solace native so if you're the founder working for some just reach out we're on boarding the first cohort right now it's probably going to get released in about a month but you know i'm probably going to have another talk about this but i guess now on to the most exciting thing that happened today so if you followed i skipped one of the challenges it was access today we just released live on the mainnet the total wallet coverage takes literally just couple clicks to cover your entire defy activity with one single policy you don't need to choose the time it's perpetual uh the risk rating is dynamic so basically as you come in and out of the protocols the policy tracks you the policy sees where the risks are everything is being evaluated by the risk pipeline and you're just getting like billed on a weekly basis but then you always know what you're being charged for uh super easy you never overpay you literally only pay for what you use so with that i guess i can just show you how it works i still haven't bought the policy on my public account so let me just show you [Music] i can also take questions while i'm doing this but basically here's the dashboard um oh [ __ ] okay well here's an awkward silence um is it on all right awesome um but yeah so as you connect the wallet um we get the data on your portfolio basically all the protocols you're exposed to and here you can see your portfolio details what like the risk level is and we also give you kind of like the annual quote even though it's a projection and then you can just click get started basically we offer you to choose the limit basically like up to what amount you want to be paid we recommend to go like a little bit above your highest position because as i said you're pulling the risk inside your own portfolio now so you don't actually have to cover your entire portfolio balance only the highest position really because um you know if one protocol falls you're still going to get paid out the whole amount so that's where the efficiency is passed on to you but yeah let's choose 120 um and then yeah i mean okay well 400 days it's i'll just do a little bit okay um so yeah now we just need to approve to spend the die oh [ __ ] oh i gotta unlock this do you guys have any questions while i'm doing this what happens when the token balance inside your wallet evolves does your coverage premium also revolve over time what was that if your token balance so the max is set by the cover limit but the what you're paying for is based on your portfolio balance so let's say you like if you set cover limit 50 000 your portfolio is 10 000 you're not paying on a 50 000 balance you're only paying off of like your portfolio but then if it like goes above the limit um then you're paying off the limit and you will be paid in case of like in some short full event happens up to that limit so if your limit is like 50k you're gonna get uh up to 50k is that i guess the premium that i'm paying is based on the 10k that i was holding right but then what if that 10k becomes like 18k or like 8k yeah so like you're paying off of that balance yeah yeah yeah the rate rate is dynamic like i mean your risk rating is dynamic depending on what your exposure is and so the rate is also dynamic is there a risk so we can insure more than what we have that would be like the simple answer so if we have like 10 million in underwriting capital and we like underwrite 10 million dollars worth of policies um so basically there's always like a guarantee that the pool has enough liquidity to pay you out and also since we're pulling the risk inside all the policies um basically we have a way to track and see where are the policy holders liquidity sitting and so based on that we can go deeper into the leverage so let's say like if we have 100 million underwriting capital likely we can underwrite 200 million depending on how risk is spread across the ecosystem on the user side so you know you having let's say like five protocols uh and like some other user has five protocols it could be like ten in total uh then we can lever more but if there are like five same protocols then we don't leverage so much so it's all just like based on how much risk you're pulling more diversified risk more risk pooling allows you to you know basically do that uh to like lever in deeper do you do do you have any plan where's that i am over here here all right cool yeah i just want to see do you have any plan to launch a product which is just addressing the bridge exploits yeah so we've been talking about this with uh some bridges yes short answer is yes we are um so i guess like the high in demand um why is this pending okay yeah i mean like high-end demand coverage is our like d-bag for stable coins and bridges i think right now like everybody wants to bridge coverage and also like when we developed our cross-chain architecture we removed that risk we you know you have to do like an extra step basically after bridging you have to like unwrap the token with us but our token is native on every chain instead of being like a bridged asset because of that i recommend doing that

Automatic transcript — names and jargon may be misspelled.