Options & how to price them in DeFi / Mattéo Georges

Transcript

all good so hello everyone um I'm Matthew I work at pragma basically we're working on a decentralized oracles and especially we built like the the first provable Oracle uh leveraging validity proofs to basically prove the aggregation of all the the data and so today I'm going to talk to you a little bit about crypto options basically around the different uh so the the landscape today of options in crypto and what we can do better to attract more liquidity and to make like the user experience better um and so you know may I ask just before

um everyone is familiar with options yeah maybe you can raise your hand if you're familiar with options it will be easier okay great uh so most of the audience is familiar but anyway just a quick reminder um basic options give the right uh to a barrier to but not the obligation to buy or sell a specific asset at this specific expiry um and a specific price so we call it the strike price there's a precise expiry there's a few types of options but we're mainly going to talk about European options so you can you have the right to buy the option just at the expiry um and that's basically the the most

liquid options today in crypto so we're going to quickly just cover um the the landscape today uh and it's composed of mainly three parts I'm going to start with structural products so it's mainly defy option volts and basically it allows users to provide liquidity and the the defy option volts protocol is going to sell options to to sophistic sophisticated actors um and so it was started mainly with ribbon finance and there's this whole wave of protocol that took advantage of this so it's a basically the main problem with uh those

uh those uh those products uh is that uh they mainly they try all to sell the options at the same time so uh there was an article of Paradigm so the the liquidity layer uh that specifically specifically specifically um uh studied uh the fact that those option Protocols are selling all the options on Fridays because Dairy bit which is like the the most liquid market for options most of the liquidity is on expiries on Fridays and so it's easier for market makers to uh to hedge uh on Fridays and so you all those options Protocols are selling their options uh on Friday and so actors

of the ecosystem just took advantage of this and so you're losing a lot of yield from this fact so that's why it's not really optimized then there's amms which is basically the same Principle as uni swap but for options the main problem is the fact that you can always price options in a more efficient way of chain than all chain and so you have a lot of inefficiencies by pricing options on chain with amms and so there's a lot of Arbitrage with which leads to the fact that LPS and liquidity providers are losing money because of those inefficiencies and so Protocols are forced to have higher fees so that's definitely not

like the the best way to build on chain options uh the the last way is to have a classic order book um but the main problem was with the limitations of blockchains so it's really hard to have a scalable audiobook on chain but we are going to see that we're starting to have the technology that enables us to put order books fully on chain and so that is going to enable us to have more efficient options so I'm going to go through how we can price options in a efficient and decentralized way so that we can power those order books which are decentralized I'm just going to go through quickly uh

that the state of options today so this is the volume uh we're seeing a lot of volume on bitcoin but ethereum is uh quickly um gaining volume there's mainly those two assets because you need to have a very liquid future Market if you want to have options so that's why most of the liquidity is on bitcoin and ethereum to the encrypto let's talk now about the the differences in terms of maturities we have uh between crypto and traditional markets so we can see that there's a lot more maturities uh for for the the SNP and it's a good thing actually because having less maturities just concentrates the liquidity in those those maturities but we're still not at the same point as traditional markets in terms of

let's compare now the the log forward moniness and basically we can see that options are not used in the same way in traditional markets then in crypto basically the what you can see here uh to to the left is the the log for one moniness for um for traditional markets for the the SNP and so the difference is not really in the range of look for online s so you have pretty much the same range but it's biased to the to the left because uh in traditional markets uh hedge funds use this to uh hedge and uh to have uh to limit their exposure so that's why it's biased to the left but in crypto Market we have seen more

volatility and so people tend to buy options that are more uh out of the money or way like in the wings so really out of the money or in the money and so that's the the main difference we have in terms of log for one moniness and now I'm going to talk like the the main difference and why we need a specific framework to price options in crypto and it's about basically the spread we have uh and so the the low liquidity we have in crypto today so basically what we can see is that the the tick size for options in crypto is uh is like way larger than in traditional markets and this leads to a spread so be

that spread that is uh from like 2 to 50 50 times more than in traditional markets so you have like Australia which is up to 50 times more in crypto Market um so that's the the main differences we have and so it leads to a lot of problems in how we price options uh in crypto so I'm going to go through here quickly how to do we price options in crypto uh and how do we price uh European uh options uh so we are using uh of course black skulls and mainly black schools takes an input a few parameters so the asset price the strike price uh the the risk-free interest rate um but the the all those parameters are

parameters that are known by the actors but when you want to price on options what you don't know and you need to evaluate is the implied volatility so that's mainly the parameter that we're trying to evaluate in order to to price crypto options and so how to uh evaluate implied volatility yeah basically what we do is we look at uh the prices of options in the market and from here we we derived the implied volatility from the the option price and so here is a volatility smile basically for One X by one expiry you have different strikes and you can plot

basically with all the points uh the the different implied volatilities uh depending on the the the strike at the same expiry um and so once you have all those points um you you have to parameterize uh the the you have to do the parametrization so there's different type of parametrization but basically a parameterization is uh you have a curve and you need to minimize the error between the points and so the implied volatility points and the curve so you define an error here it's a mean square error and you have to minimize the uh to the left it's uh like the the the the points of the curve and

to the right it's like the the points You observe on the market and you need to minimize this error in order to have this parametrization uh in we're going to see that uh in traditional markets the the how you evaluate the efficient price of the market is by using mid prices uh but you do that because you have uh spread that are really tight but as I said before uh sprays are way larger so up to 50 times larger in crypto and so we need a file to find a way uh to compute the efficient price in order to to make the parametrization um so here is an example of a parametrization so you have observed volatilities and you try uh to to have the curve uh fitting those different

points that you observe in the market so that you can after that use it for different strikes at the same expiry um so now let's take a look at um how we can remove and how we can compute the efficient price in crypto Market and so the the first thing to uh to know is that there's a difference when you try to uh to to evaluate those efficient price if you're out of the money or in the money and basically if you're uh in the money as you can see like for instance uh uh on the the left the the real the sensitivity uh is way larger when you're in the money uh than when you're out of

the money uh and so this introduces uh basically noise when you're uh in the money and so what we need to do is remove all the uh the prices that are in the money in order to have something really robust so the first step to to compute a efficient implied volatility is to remove uh all the in the money uh options and prices uh the the second step and as you may see here uh so we have uh in uh red plots uh so this is a simulation to try to see like the the bias there is uh between like the when we use the mid price and when we use the the the the the real price of the the the real implied volatility uh and we can see that uh basically when we

use mid price we have an underestimation uh of the the implied volatility especially in the wings and so how to remove uh this problem um so there's mainly a big contribution that was made by uh kaiko which introduced uh two main techniques to remove the the noise we might have in the wings the first one is uh to introduce an anchor price which is basically derived from uh the implied volatility we have and which removes the the buyers we have in the wings so in in blue you have the the mid prices uh and in yellow you have the the anchor price and you can see that

especially in the wings we removed all the the the the the the underestimation we had uh and there's also uh better augmentation which is a data augmentation techniques basically there's um also uh because the market is not liquid you have in the wings a lot of problems because sometimes you use you miss some ask or bid prices and so you can't construct a mid price and so you can use a better augmentation to construct the the spread and to have the the data that is missing so that's the the main contributions here and that's basically how we compute from Market data uh all the the the the volatility

smiles and all the volatility points so now that uh we have all those algorithms what we need to do is to provide uh basically to provide uh uh here we go to provide the different D5 protocols with this data but as you might you might have seen um there's uh problems because the algorithm it's like heavily there's a lot of computation that you just cannot do on chain and also there's mainly one data sources right now which is dairy bit because it's like 80 of the market is on Dairy bit and so you need to diversify the the data sources if you want to have something that is really

decentralized for defy and so how do we from this centralized computation and decentralized implied volatilities how do we provide those fields to D5 protocols uh and so our answer is to like have multiple data sources uh so we Source uh from different exchanges uh but also we have different market makers that are providing their own proprietary data uh so that we have something really resilient in terms of data and then how to have this trustless computation because you want to make sure that all the computation that happens on Raw data is trustless and so we now have uh like

wonderful Technologies with validity proofs and the ER zero knowledge proofs and basically you can leverage um our Oracle which is built on on starknet to have this Stressless computation and to make sure that those algorithms are not introducing any more trust assumptions so that's basically how to uh efficiently uh compute and price options and that's our answer to how we can make D5 more efficient for the the option market so there's if you have any question feel free to ask you'll have a lot more documentation on the website Twitter and

everything but happy to take questions if you have right now hello uh on one of the slides you showed that uh on like in 2025 onwards there'll be no Bitcoin options won't have expiry is that like so there won't be any expiries for Bitcoin options is that like maybe going down the route of Perpetual options with like the likes of panoptic or what was that slide kind of implying so yeah the the fact is in crypto uh we don't have a lot of liquidity for options and so what exchanges uh did is they just removed a lot of expiries so you only have like I think 12 expiries uh

because you need to concentrate liquidity on those expiries and the market just don't want to buy options for 2025 because like the options that just expire even in 2024 they're really uh there's a lot of volatility and so just the the market participants don't have appetite for options that end later but that's for BTC for it it's just like uh that was like the options and expiries you had on Dairy bit um but yeah that's it just translates that the the market just don't have appetite for expiries that are you know in 2025

sorry I have a couple of questions the first is um you know you mentioned thereabit has 80 market share so the approach is to take other um market makers data Etc to make um to aggregate data but isn't there I mean possibility unless that price is tradable market makers can just provide wrong data right or incorrect data that's my first question and then the second uh question is this uh protocol actually is this a protocol that's actually live today providing prices and um if so can you talk a little bit about the cost is that borne by you guys or uh people who consume their prices and stuff like that thank you sure yeah uh so on the first question yeah uh of course so there's two parts the first

Dairy bit can manipulate data because like 80 of the market is there uh and so we wait uh by volume and so most of like the implied volatility comes from the Ribbit because all the market is there so they can also maybe manipulate the price but for the market makers we have uh so uh contracts with them so for now it's a trusted Service uh but at some point the goal is to uh have like some kind of proof of stake basically market makers have value at stake we compute a final parametrization and then we'll slash based on the final parametrization so that's the the goal but for now it's permissioned uh and so the product is uh is not live yet so we are working with options protocol right now to integrate

but so yeah in terms of timeline uh most of like the the older books are coming yeah like later this summer uh and end of the year so that's the target like sorry just just uh kind of related to that so uh like we saw with Futures Futures used to have expiries before kind of bitmex came out with a Perpetual Innovation you could say um and then Now on-chain options have faced the problem of kind of relying on like like they have like you said there's lots of expiries and it's hard to provide liquidity for all of them do you think on-chain options will go the down the

route of being Perpetual as well so not have an expiry but pay a constant funding rate similar to how Futures work um you showed on the slide it was like panoptic there so they're kind of going down that route I'm not sure if you've maybe looked into them to be honest I don't have an answer like uh yeah like right now I don't have an answer I'm sorry but that's an interesting question no problem okay yeah thanks everyone [Applause] [Music]