A dex with reactive liquidity

Transcript

so i'm presenting mangrove now that's a decentralized exchange with new features so um it's based on an older book that's not new that's kind of the traditional model in traditional finance and it's been the dominant model uh not so on the blockchain as you know and from the take care point of view or from the point of view of price aggregators the really this is just a price curve price volume curve it offers nothing special the the dax is aimed at the makers the people actually providing liquidity so here we actually provide the distinct unique features that we think are

breakthroughs in the in the field of dexes so the very first one is simply explain it's uh that we do not like capital so differently to usual order book based or actually amms based the taxes in define people as a maker ask you to actually provide liquidity which means concretely you have to lock it in right and people talk about total value lacks we don't and the second thing if you're ready to juggle with credit lines is that you may actually don't have to provide capital at all so you can have zero cash market making strategies i'll explain that now the third thing and and that's super important for people who track price

prices accurately is that we actually have a mechanism to defend your prices against price motion uh sudden hikes and jumps in prices so you don't have price exposure or inventory risk as people call it sometimes and this is uh so we'll call it here in between us defensive last looks it may not be the final work for this and this is particularly targeting market makers because they really need to be careful about prices and then the fourth feature is really another way to refresh the first one is that there is no exclusivity meaning uh since you're not actually providing the liquidity you might as well publish it multiple times right so

and that's a kind of a liquidity amplifier mechanism so the uh by the way there is a if you can as you can see uh at the bottom right there is a telecom channel that we just created and if you want to chat get a you know some kind of discussion technical or not later on please just connect to this we'll be watching it and we would be delighted to talk to people and get feedback all right so these are the features uh now the next thing i want to say and i think it's important or maybe it's not depending on your point of view but if you're a builder type this this is interesting actually everything here followed so first of all it's not a closed list

we expect many more features to emerge actually we didn't start with at least our feature uh it's some of them we actually found it on the way and actually everything follows from one single simple technical idea that we have implemented in the code and that's that idea so the idea is that differently to normal decks uh you have you're not providing liquidity so what do you provide instead well so you provide code or rather a callback address to some code and since and second idea we do not want to know what that code is you provide whatever you want complete black box so

whatever strategy you can come up with we don't ask you to follow any kind of requirements any kind of structure it's a bit like a flash load mechanism you take the money you have to return it by the end of your transaction if you don't then we revert right so we it's very much in the same spirit and uh since you're uh not we've seen not asking any guarantee that your code is actually correct it may not be correct it may just be clumsy it may be buggy or it may just be malicious so we ask you to provide in f some provision that will cover for the gas expenses that you may incur to the taker so that if you do nothing of interest to the ticker in the book then at least

he's not paying for your gas expenses right i'll see we the maker else would leave like a parasitic being that's actually pumping gas from the ticker for to no avail so an offer is as usual a volume that you're ready to give and then a volume that you're ready to that you're asking in return plus a gas requirement which is some upper bound on the amount of gas that you're projecting to use it may be a you know very large upper bound and then code or address to code plus a provision that's covering your gas requirement at what gas price well at the mean gas prices dictated by the current platform so this is the same and now

i have to be careful so this is how we see people could actually use this platform on the maker side but we have to be to agree that this is not prescribed by us so you if you want to do other things then you're welcome this is how we see a reasonable way to sequence the way the matching engine wake ups will be processed by the script so uh to not be confused i'm going to call the code that to which you give a callback address a script but it has to be also very clear that this code is on chain right and this actually is key for the rest so you have a dedicated piece of code that's managing a bunch of

offers on the decks on multiple pairs of the decks and that that i'm going to call the script it's a bit of a loose denomination it's just to be clear between us so here's how we see a reasonable way that the that script could be organized in three easy steps so the very first event is that the matching engine is actually waking you up and says okay you said you wanted that much i have that much or maybe you know some partial feel and so what do you do so the first thing the script does after receiving this match from the mangrove engine is you know asking itself the script is wondering deliberating about whether

that's still a good price because the last time the script was alive was maybe at time t zero and now it would time t one and maybe the picture of the world that the script had that time t0 was good enough and it could formulate some kind of offer but now at t1 it has drifted so maybe it's not willing to actually go with the deal anymore so the question here the script is trying to answer is should we go or not go am i going to renege on my offer or not so to make a good decision about this you're going to collect signals from the book and signals from the on-chain environment where you are to see whether the price is unchained now i've drifted

and potentially have changed signals are returned to this of course they are more expensive and so there's a trade-off here and at that stage so your defensive price module has decided whether to go or not go and if it goes then we go to the second stage which is to so for example you know little piece of code markup piece of code uh you're actually going to compare the price you're offering or you were offering at t0 with the current uni swap twap for example and if you find that you're under selling by more than one percent you'll say no i'm not doing this i'd rather pay the little compensation fine to the ticket and actually run my order but suppose you do actually want to do it

then you go into the second stage of the rocket which is to source the liquidity and for example so this is where the no capital thing or the no capital locked thing gets into play you don't actually have the money inside the decks you're gonna have to source it and so one simple way to do it is for example to have it in some kind of savings account you know it's by compound it's earning yield and it's just now that the matching engine has woken you up that you say okay give me my liquidity your redeemer on compound there are more clever strategies more you know you know that there's an entirely uh new novel uh strategic design space but this is about the simplest thing you can think of

you can also redeem liquidity from an amm everything everywhere where your savings actually earning yield you can bring back at that time and not before and then the third stage i'm simplifying a little bit the mechanics here but the third stage of the rocket is to actually update your offers because you have taken new information at the first step you know more about the prices as they are now so maybe you should update all the offers that are on the management by your script and have a better more accurate picture of the current price state so this is the refresh offers and i'll also comment on this so for example you can you know you've just read at the first stage the current

t-wipe unisop price so maybe you can just update it in all your offers on the other side of the book to begin with so if you're under selling on one side you're overselling on the other side which means that you have offers that are waiting resting offers as people said that are waiting on the book and they're not going to be taken because they are just too expensive now so bring them back a little all right and of course there is this uh possibility of filling and so there's this preconception that failing is a bad thing but actually if you look at traditional finance failing can be a very efficient way to handle capital and you know there's a lot of literature about strategic fails in traditional finance at some point it was eight percent of the orders we are

failing on new york stock exchange so we don't consider feeling as a bad thing you consider feeling as a as a something you have to control you have to sandbox and that's what we do here but filling is actually a very efficient way to if you're never gonna fail then maybe you know you're a bit uh too too inefficient all right so there is this strategic fail option or last look or protection against price exposure all right so uh this is this so every single of these three stages connects to the features i was explaining at the beginning so let's label them so this third one is liquidity amplification or ghost liquidity and here i'm using the bad name or the name that

traditional finance people have given it to which is ghost liquidity and which is something they'd love to do but they're a bit afraid of doing here no problem it's under code control so we can do both liquidity and that's what i'm going to exemplify third then the second thing is defensive pricing defensive pricing is actually very subtle because there's a difference between the price processes you measure them on chain and the ones that you measure of chain they don't have the same cost or price purist so the price of on-chain is not the same as of chain and the time scales are different as well so this this makes it a a subtle strategic space to navigate so i'll give you examples and then the third thing is maybe the

most obvious is the liquidity sourcing so this is the first example we'll look at okay so let's start with no capital locked or how we can source liquidity so here i'm using so this is not a strategy per se this is a piece of a strategy that you can use as a module that's going to source liquidity but with a lender-only behavior all right so it's just a savings account and essentially what the strategy can do is move things from your local money to the savings and back to the to your checks account so we start with and i'm not telling you how this is this is part of the really strategic strategic module this is just how you move capital around so here you start with 100

dies on hand this is your local money that's in your pocket right so real cash and then you have 900 here which is deposited as collateral uh on compensate and it's collateral that you'll never borrow from right it's you're purely putting it there to get returns and now you have a a mangrove event and here there's a ticker that comes and wants 800 of your uh your dice and 800 is carefully calculated so that it's a volume that you have because you have a thousand you're here you're not doing a liquidity amplification so you have you don't have to draw it's fine but you have only 100 in your packet so you have to go to your savings account and that's what the transition does here it's going to go in the savings account

and at the end you get 200 left on your savings and in return you have your f which you also as you can as you can see here you also deposit here uh in your savings account right so we drew a bit of dice and i put back a bit of f and the price is a little optimistic i think it's now cheaper than that so this is the pure lending strategy one simple instance of generating returns and having your money that's in your resting offer not luck but earning now something more ambitious oh no sorry so this is okay this is different this is a bit mathematical i think it's interesting

i try to be quick but clear so we actually did a little simplified mathematical model of this there's obviously a contentious uh thing here which is that uh you know if it cost nothing to go between the savings and the and the checks accounts then you'll do it all the time and you can have everything hundred percent of your capital just earning right but then the local money uh is never enough so every time you have a match then you have to rent your savings account and that costs so on the x axis here uh for both rows you have uh the relative price of the interest rate and

the cost of moving between your two positions so it's the it's how much your capital is earning in the savings divided by uh how much it costs to move from the savings back and forth so the rebalancing cost right so moving capital has as a price especially on ethereum it can be quite costly so you have to be careful and so the surprising fact so this these are this is extracted from two weeks of uni swap data uh hundred thousand actual orders uh on the those three respective pairs and the question we ask to the model is okay you get this time series what is the percentage of local cash that you need to keep

after which you are going to push to the savings what's the optimal one it can be it's it's easy to show that you know it's a convex problem so there's the best solution and here it is and this diagram actually gives you the solution have i pressed something about here it still works here someone is doing something okay cool so anyway i can continue with this line

so the surprising discovery is that uh no matter how expensive the rebalancing gas costs are you can actually get eighty percent of the return just by uh having a five max so max fraction of thank you a max fraction of four uh eighty 80 40 sorry so the the optimal policy for that time series is that you will never let the amount of local cash that you have in die or at in both it's actually quite symmetric exceed 40 so when you're in an excess of this you're going to push your savings account and uh when you're winning when you're under zero then like we saw in the transition you're

going to pull from that uh savings account and so this in a way uh you know if uh you you could say this is a strategy that's gonna beat uni swap no matter what the gas costs are uh because uh you know and there's no magic actually the reason is that the uh distribution of size orders on uni swap because of the penalty for sleep age is extremely uh narrow and and will hardly ever in relative size exceed one percent um okay so okay well i'll do it without slides unless we get slate so we finished that first uh uh exploring that that first set of

features which which are uh related to um you know managing liquidity so now we can move over to the first stage which is price defense so price defense so the principle is very simple you're going to collect uh on-chain information and try for example you need to have two apps because there are public variables that you can get for cheap and you're going to compare what you thought was the spread which you posted which you based your offer on at time t0 with time t1 which is now right and if you find that you're under selling then you'll say no if you're over selling you're always happy right and then so the the interesting question here is the following um so that two there are two beautiful

slides that talk about this but let me do about slides it may actually be better even so first thing um the script the script is not managing just one offer it's managing maybe a multi-band offer on both sides of a pair but actually maybe managing an entire set of pairs right typically an a b b c c a triangle so this means i'm managing actually six different uh half books and in each one of them i have maybe three offers right because i'm discretizing a price curve so maybe i have up to 18 offers that i need to manage so now so this is this is simple but actually shows in the maps and it's

fun each time you wake me up on one of these 18 different offers i have an opportunity to take the information and bring it back to all the other 17 right so the uh in my sleeping pattern as a script the random time between two instances when i'm running is going to be roughly scaling as uh the typical interval which depends on the spread complicated the question but divided by the number of offers i have under management so there's a and so it's so in traditional market making optimal control problems you manage your spread and you know you

you kind of find some kind of trade-off between your inventory risk and you know how often you want to be taken here there is an additional element in this trade-off which is the amount of information you can collect so you have enough intuitively you have another reason to be closer to the top of the book which is to be taken more often but now you can also mitigate this or you know you can control that typical inter wake up times or that typical delay uh between the t0 and the t1 by multiplying your offers so this is this is kind of fun so that you know that's one thing now there's another

discussion to be had which is really interesting and i had the cool data to show you uh so we've been tracing back okay so let me take a step back so this is pure this is purely about on-chain collection of information and of course we want on chain collection because it's cheap right relatively cheap it's just reading a viable now the price mechanism or the the price discovery process doesn't only run on ethereum of course it runs on the the the kind of the crown of centralized exchanges around and they are operating on a very different time scale and so and actually unicef is essentially a zombie compared to i don't know kraken or binance not the

binance chain right the binance exchange and so maybe they operate on the you know millisecond uh time scale and uh you know by construction the the physical inter block time on ethereum being what it is this is really super fast compared to the evolution of the price process on chain so what does that mean concretely it means that it may be and and this is what we measured by looking at the data it may be that what the chain believes or what unisa believes the price is is actually not a very good approximation of what the rest of the world believes is the price so you have this two time scale mechanics and uh what we did is we calibrate so we downloaded the data for i think for the

entirety of june so the execution price of binance on on one one of the three main pairs that we've looked at i can remember which i think at usdc and then we looked at how how often there is a significant discrepancy in both directions because on both sides of the book between the price as read by binance and the right the the one that's actually read by uni swap and this is important because when i'm collecting on-chain data then maybe it's not accurate so how often as a market maker with a little infrastructure and a but that can talk to my script how often do i have to press the emergency button of my script

sorry of my butt and tell the butt to tell the script that the unicorn price is not actually something he should read anymore and the answer so we calibrated so there's a standard model for financial data that's not predictive but that's measuring key characteristic of financial time series which is called the germ diffusion process and so we calibrated and these are you know well understood as you can imagine techniques to do this we calibrated that and the answer is the statistics is not very robust but one robust observation which we could also do just with our eyes if we had the slides is that uh the the jumps or the poissonian component of the germ diffusion is

actually has very low intensity what does that mean is that you have to press the button very rarely maybe once a week and then it's a completely different ball game right because i'm ready to do something expensive if i'm you know trading large volumes once a week that's okay what i don't want to do is to do it once a block right so this is the second question and now i'll go to the so so i don't know if you following the outline of the talk anymore so we talked about sourcing liquidity first then we talked about collecting information about price and making a good price decision including off chain information and now we're going to talk about ghost liquidity ghost liquidity is a very simple say you're trading on a triangle or abc and you have a bunch of ace you can post

your a's on both a b and ac and the sequential nature of the engine will make it so that once if you're taken let's say on a b and you're throwing a hundred something ace i guess uh then and that's the total amount you have then uh you you can volume adjust the other offer so that you're never in the red right so the next offer will not find what people interact will find and again it exists in traditional finance it's called an overfeel you don't want an over fear right the funny thing is that so we've talked not just to defy people but we've talked a lot with traditional finance people and you can find uh correlates of all that the mangrove is actually allowing

you to do it's just that it's a much more uh level uh um space of exercising all those ideas that have been around for a long time and of course you have composibility for example you you see a concrete manifestation of composability as the liquidity sourcing strategies that i've just shown all right so i'm going to conclude so um let's let's uh summarize right so first thing um we have a new dex it's really breaking the rules uh you don't have to provide liquidity it's uh you can use define

the d5 blocks as a back end you could say and that's something you can do in traditional finance right you can't actually or you can do it in a limited way by borrowing from your own exchange but you can't borrow somewhere else right so here you can so this is the uh no locking of the capital or the you know very low requirement on capital if you're ready to juggle with credit lines playing with compound and ave constantly to just finance your strategies so that's the first point of course this is how so if you think of it what i've explained here all the trade-offs of the strategic space the one the ones we've seen uh those

trade-offs will play out very differently depending on what the evm compatible substrate chain is so if we deploy an ethereum then you know you'll have harsh conditions it's slow and expensive and so the type of strategies that you want to deploy the market making you can do is uh you know restricted uh although i've shown you the example where you can be tuney swap so this is already something now if you move to something which is more like a an l2 and is a both faster and cheaper then the strategic space unfolds and you have many more opportunities and at in the limit where gas costs are zero then there's a lot of those trade-offs that just disappear and you you you have full strategic command

so uh okay so uh so in particular you could do this zero cash uh market making which is something like the people market makers are doing international finance or in centralized exchanges except that they're exposed to a really heterogeneous bunch of interest you know credit line constraints whereas it's far more homogeneous and competitive number one uh number two both liquidity which is this liquidity amplifier thing and number three and we spend a lot of time on this the the you know how accurately do you have how much you have to pay in the end to have an accurate tracking of the

price processes so these are the three things so uh now what comes next uh so the the first axiom we pursued was uh you know the evident actions like you know best execution things like this that the uh book structure gives you for free but the first thing we said we want black boxes uh so that's something we may come to actually relax because this is so mangrove is just for spot markets right what does that mean conceptually it means that whatever the matching engine is trying to create a match and some kind of transaction at the time it's uh executed then it's finished there is no time extended relationship

between the maker and the taker but if you think of credit markets or credit line markets or options that's exactly what it is right there is a time extended relationship between the two parties but that so and so in and we are very interested in moving to those kind of markets like interest rate markets we've been talking extensively with another french d5 project called morpho or another one called ap wine where they are getting into this time extended thing for that we have to change a little bit something in the design we have to not uh adhere to the completely uh okay cool all right let's see let's see let's see

what the talk actually was okay this is the lender borrower one and uh this is actually funny defensive pricing this this is what you didn't have to look at i think i think that this is these are the this is the two time scales uh data look i'm i'm diving on this kind of anomaly and here i see that binance is severely uh you know shifted compared to the to the red uh price process which is almost constant right because we're at a time scale where it's where it moves very slowly and so this

this is some stuff i said this is the technique we use this is the ghost liquidity diagram this is the the dependency of that this is my last slide all right so um so black box is something we actually want to uh modify there's nothing which we're really ready to modify which is that at the moment we have a mangrove doesn't talk to your script unless it actually runs it so on so you never run a script unless it's actually you know set to motion by the matching engine but why not and also other questions of the same type what's the type of information that's

available from the book not from the chain but from the book to the script when you wake him up so that will depend on we are going to organize the market maker workshops try to understand how they think what kind of strategy they would like to deploy and if that if there is additional piece of info that they like to have maybe it will enrich the engine that doesn't really change much to the most of the code these are important modifications that we can [Music] make to the code easily and then the final thing is that the prices do not include the gas costs so we uh we are i'm not sure that's the that's directly translated from french so i think we're showing gross prices i

don't know if that's what you say not net prices and so that means that the uh this seems a reasonable model if maker pays the gas cost not the mangrovian cost of matching but the idiosyncratic specific gas cuts to finance its price defense is liquidity sourcing but okay this is to discuss and so i'm finished uh please you know direct any question you'll have after those i don't know how much time we have now but to this telegram and it would be really good to get a return reactions feedbacks you know suggestions and so on thank you for your [Applause] attention

you don't have to ask questions for a long time but you make a complaint yeah online the whole transaction goes from like the user sending the call to like the router and how that goes like what exactly when a user so he knows more about the code than i do

so when the taker comes in and let's say they do a market order you just have the offers that are already sorted and at this point the offers are just these volumes and this this pointer to a piece of code and then one by one all the offers are executed if they succeed in providing the liquidity at the required price so this is black box okay the black box code is executed now if they don't succeed then they are penalized so we take from their provision if they succeed then there is still some volume left to fulfill we go to the next offer so this is the first pass and then there's a second pass because there's at some point offers want to reinsert the stem cells in the order book and in this first pass it has to be reinferency locked

but once the order is complete then you recall them again and you say okay now the order book is accessible you can reinsert yourself and you have a memory of the offer you just executed we sent some data about what just happened in the previous call is this clear okay cool yeah i think so i simplified a little bit when i described like the processing order on that very first diagram so that refresh our first thing actually comes so when the matching engine climbs up the book it will push every order that succeeded or not actually on a stack and then when the volume is

filled then it will go back down and call every single offer and say do you want to update something so there is a little bit between this and that there's actually a disconnect and you go back to sleep for the time uh to the for the matching engine to actually fulfill the make the ticker's market order and then you'll come back and actually so okay and the little game theoretic uh oddity here is that if you have the best price then you'll be called first which means that you'll be the last to update which means you can look at what the other players have uh have made which is kind of fun because it means you know if you're the best then you have a little

kind of a self accumulating uh advantage but you know just be better than the best and and you're good thank you i like the idea how you bring order book but as a defy building block but i have questions maybe stupid sorry if you have a lot of orders with equal price for example but blockchain has like limited capacity so how are you going to order this transaction because uh like your bot will be order this execution and what happened with the transactions that

not able to go to this block so how how you decide this like it's like mev inside the mov or like i don't know because uh it depends so that there are two issues here this is how much gas you're allowed to spend within a block no no no no no imagine you have like 100 000 transactions that much like you like you you see that it might be executed but there is no chance to include this into the blockchain how you will order uh what transaction should go and what should like postpone okay so so we are not doing anything about scaling and we take the substrate as it is right so if it's slow then the miners are solving or

you know you you could dispute whether they're solving it in a fair way maybe but they're solving that problem for us we're only concerned of finding a transaction that's a market order and then we just serve it and so the there as well you have a restriction which is the size of the block but then it's up to the transaction to not ask for an amount to not uh you know exceed that size anyway it will not run so so i think what you're asking is a question about the substrate on which the mangrove is running you know and if it's very restricted and you know we'll you will have to they will it will take the you will take the takers what you're saying is there may be a very large demand for actually sending others

to a given pair and you know it's it's not going to be it's going to be congested but that's not our mechanism that's the underpinning mechanism of the chain do you don't say something else don't you all right so so i think another way to put it is uh the point you're making maybe is we'd better move to something which is faster like cardano no i'm kidding thank you so uh um yeah probably a stupid question too uh at at which point uh the matching happens so uh where do you have like like some bot that does the matching or uh no no or the maker pays uh

for the guests that uh thank you so so the so it's a bit like the talk from uh masha musavi yesterday so the book is and so it's a frequent batch auction that she's doing but so it's slightly different because yeah it's a continuous action but it's similar in the sense that everything is unchained there's nothing of chain maybe the strategists will like to be partly of chain and that's what we discussed a bit like the hummingbit stock if you've seen it right but here you differently to hummingbird strategist you can actually uh delegate some of your iq and strategic intelligence on the chain and that you know gives you a lot of the advantages that we discussed so uh so the entire everything is on

chain your strategy and the other book everything is on chain but but but but uh like someone has to send the transaction that actually uh does the matching so the question is uh yeah who who who it is is it is a taker who submits yes so it's taker or price aggregator you know like this is so like i said so an exchange is a two-sided market right so if you go from the taker side nothing new it's just the ordinary mechanism except that it's composable so you can you can be a contract and call the and behave like a taker that's fine you know it may be useful if not maybe uh maybe a certain pair of the mangrove will actually call another pair

and so the the maker of a pair will be a taker of another one that's entirely possible uh but uh so it's possible uh but but in a way it's nothing novel on this side it's on the other side that we provide the the breakthroughs i see thank you i um you said all the orders are unchained so what happens um or how do you prevent it from becoming very gas and efficient if i change orders or cancel orders etc because everything will be very good question so uh so there's this kind of preconception

that comes from traditional finance really that uh the only way you can interact with an exchange is by you know cancel and updates so you you have a change signal you're observing everything you can you aggregate signals in a kind of a proprietary way on your servers you have a big infrastructure maybe you pay very a very hefty sum every month to be collocated with the exchange and you try to make that loop as short as possible but the only way you talk to the exchange is cancel update cancel update cancel update that's essentially what hummingbird is uh is doing right and unsurprisingly because the people are coming from traditional finance here the unique thing about the mangrove is that you don't have to be close to

the exchange you are inside the exchange we let your script run at the time it is matched you cannot be front run right for example if you have good price sources that are not manipulated by uh the taker then you can just reneg on the order so it's a very special way to think and so it has far-reaching consequences and one of them is the one that i explained here which is that you do not have to bother about having accurate prices from the outside world which would be update update update more than once a week and that's not expensive even on ethereum right now having said that all the

models we've we've had they're instructed by data we try to we are not ourselves strategists right so we don't really know whether the market maker actually will hit the sweet spot and be able to extract value in this way yeah so it's hard to make money in three easy steps right you know you may wonder whether gas prices actually are going to be too uh punishing on ethereum for that scheme to work but we're not totally married to uh ethereum we'd like to work with other blockchains that that i'm not going to mention here no no i'm kidding no evm substrates are totally uh interesting and right now there are evm substrates

that are very cheap on gas and the fair targets for us although they have they are less defy rich and since at the middle step we use the define the back and if there is nothing there we are limited for for another reason we can still do both liquidity for example which is the reason i was showing it to you that that that's a strategy that exists even if you have no defy thank you very much i can comment quickly so yeah i said that i talked to traditional finance people so and uh so i was telling them in a way trying to you know capture their

benevolence i said you know all these ideas here they're not really completely uh entangled with decentralized finance if it's uh uh maybe we could we could go and talk to a standard exchange and you know try to euro next i don't know new york stock exchange chicago mcintyre exchange whatever they have for example the chicago mercantile exchange has something called a non-firm quad court with a price improvement that's something you can do with the mangrove and it's kind of fun so you can you can say i did this but that no you change your mind and then you change the price it's very so the way i would phrase it is that traditional finance has been trying to do this but they didn't have a define

so they couldn't do it and there are other reasons obviously so but the people so i was talking to a big trade fund here in paris uh on monday the these guys are the scientists the scientific traders they're really good and they were really totally enthusiastic about this traditional finance thing so a lot of the ideas here are not just uh things that you that have correlate in traditional finance but the new ones would be something we could actually bring back there and then you know if we believe julian butler and his uh you know view of from the future that uh try track finance and decentralized finance will kind of converge in some way then it can be useful to have this kind of a double-sided

discourse to to make it to where we want to be thank you thank you very much you