 Okay, so what I'm going to do right now is I'm going to talk about the exchange for ERC20 tokens, which are ERC tokens on Ethereum that we developed at NOSIS. So we call it the NOSIS Dutch Exchange and it's a decentralized exchange for Ethereum tokens using the Dutch auction mechanism to determine a fair value of tokens to determine what that means in a second. So first of all, a small disclaimer, I did not make this presentation. I kind of read through it and I'm pretty sure that I understand everything. So I do understand everything on the technical side and the finance stuff. It shouldn't be very difficult. So let's start with an introduction of a traditional exchange that uses an order book and a continuous auction or a continuous exchange. So you can either do limit orders whereby they get placed on an order book. They can either be bids, which means that you will buy it at a specific price or they can be asks, which means that you will sell it at a specific price. And you have the spread, which is the difference between the bid and ask, and you can also do market orders, which are basically limit orders with the value being the current lowest or highest value respectively so that the order is immediately executed. So market order is as a result immediately executed. So that's why it's called a continuous exchange, is there is a serial processing of orders one by one. And also on a different note, I think it could be also termed a continuous exchange because the order can be fulfilled as soon as it comes through. And orders matched using pay as bid, actually. So that's that. So the problem we see with that is manyfold. So the biggest problem of all is the possibility to front run. And that means, so, okay, so for example, if you have a centralized exchange, then if you submit an order, then the exchange knows about this order before it broadcasts it, right? So it can actually act on that order, act on the information before it broadcasts it. So it gives it an unfair advantage. And in the case of decentralized exchanges, like other ones than ours, a lot of them use a centralized order book. So that still means that the exchange even theoretically has the possibility to act on the order before it fulfills it, right? So that's a problem. With our exchange, it's fully decentralized, so it's fully on chain. So nobody has any more information than other people. Yeah, anyways. So other disadvantages of order books include like you can have a high spread for like illiquid auctions, so they're not ideal for those either. And like for example, if you have a illiquid pair, then if the size of the order book on either side is not that large and if you make an order, it can push the price, right? So you will actually get a different price for every part of the order. Yeah, that's basically that. So like I guess it can be summarized by the paragraph or by the statement that applying a continuous mechanism to the blockchain doesn't make a lot of sense because in the blockchain, there's a discrete time obviously. And yeah, that creates all of these problems. Cool. So instead, what we decided to go for is something more appropriate for the blockchain, which is a batch auction system. And in particular, we basically collect all the orders at once and then we auction them off all at once, right? So we batch many orders together and we sell them all at once. Does that make sense? Cool. So in particular, we use a mechanism of the Dutch auction, which is like kind of the opposite of an English auction. So in an English auction, that's the regular auction people bid and every bid has to have a higher price than the previous one, right? For it to make any sense. So in a Dutch auction, it's the exact opposite. The price is like falling and the first person to like take that price means the thing. So that's like if you're auctioning off one item. For our purposes, we have to auction off tokens. So it's many, many, many. And hence, we adapted it for it to be usable for many, not just one thing at a time. And in particular, the way that works is, I think there's a graph that explains it the best. Yeah. So on the x-axis, you can view time. And on the y-axis, you can view the value, right? So we start off with a constant like sell volume, so constant amount of tokens to be sold. And we always multiply it by the price to get its value, right? So, yeah. So as a result, okay. So we have two expressions, right? We have the sell volume and we have the price. And what's crucial in this is that the price is decreasing. And if you have a constant sell volume and you have a decreasing price, then the value would be decreasing as well, right? So that's what we can see in the graph. Can you guys see it? It's the line in red. Is it visible? Somewhat? Okay. So the line in red is the value of the sell volume. And then what you have underneath is the kind of like the buy volume, right? So, right, right, okay. Okay, okay. So the price is like the sell volume in terms of the buy volume. So it's kind of like a conversion price between the two tokens, right? So as it's decreasing, the value of the sell volume is decreasing. And then you have the blue thingy, which is the buy volume. And that's increasing and at some point it will match, right? So a necessary and sufficient condition for that is that there's at least one buy order. And as long as there's at least one buy order, then, you know, at some point in the worst case in a really long time, those will intersect. In fact, even if there isn't any buy order, they will intersect because the way our price function works is that it actually reaches zero after 24 hours. Does that make sense? So basically, they will intersect at some point always. And that's when we close the auction. What happens then? The bidders, or in other words, the buyers, they can claim the tokens that are being auctioned off. And the sellers, they can claim the tokens that the buyers sent as well. So, yeah, does that make sense? I think that makes sense. So let's see. Right, so what's actually very crucial here is that the price is the same for all the bidders, right? So it actually doesn't matter if you participated, let me show you the graph again. Like at the very beginning or at the very end, say, the buy volume is always increasing and when they intersect, all the bidders get that price. Okay, so everybody gets the same price at the end. Okay, right, right. So actually, this is a very good illustration for understanding how it works. So you have a 2x2 diagram and you have, so prior to the auction, the sellers, they deposit the tokens and then the auction starts and then the bidders bid for them, right? So that's a very good summary of kind of the mechanism that I explained. Great. So, yeah, okay, I'm going to say it. So also notice that there's two possibilities. The buyer side can either intersect the seller side like buy with a buy order if it goes up or it can intersect it by the seller side going like for a long time and intersecting it from the top down, right? So basically there's like two possibilities. Either a buy order can clear the auction or it can be cleared what we say like with time. So yeah, that's that. So the way it works is we have like, for example, Ether and token A and there's always two symmetric auctions running at the same time. So there's going to be one where Ether is being the sell token and the token A is being like the token that's bidded and you have the opposite symmetric one where it's token A that's being the sell one and Ether is being bidded. So we always run these symmetric auctions and they always start at the same time and when the second one closes of the symmetric pair then after 10 minutes we initiate the next pair. So that's kind of how the mechanism works. And there's a condition which is that the next auction pair received a combined value of at least $1,000 worth of sell volume. So that could have, or sorry, it's not combined, it's like an OR. So either auction must have received at least $1,000 worth of sell volume. And I guess I haven't explained the price function so kind of hinted to it. So the way the price function works is we always start at two times the closing price of the last auctions and it decreases. And after six hours it reaches the past closing price, the previous closing price. And then after 24 hours it reaches zero. So I don't know, maybe it's better if I just really simply. So we have, it starts at like two times the previous closing price and then after maybe something like this. So it's a reciprocal function. Cool. Right, so yeah, I mean this diagram just summarizes what I explained except for multiple auctions, for multiple things after one another in sequence. Cool, so now I'm going to get into like the more interesting stuff apart from like the basic mechanism. So you don't need any tokens obviously to participate in the exchange and in particular, noses tokens are not required. And so at noses we have this thing where like you stake noses tokens and they generate these things called owl tokens and then those are used to pay fees. So the owl tokens you do not need them to participate in the exchange but if you have them then like it's worth it to use them basically. So then we have a token which is intrinsic to the Dutch exchange and that's what we call a magnolia token and it is used to decrease your fees. So there are fees in the exchange and they, let me just, I remember there was a slide, yep. So there's 0.5% for both the seller side and the buyer side but they do not go to us. So it's a fully decentralized platform. The fees, they are put back into the system. However, you can also decide to spend up to half of your fee in like the equivalent amount of like owl tokens. Does that make sense? I think so. So that's pretty much it. And how do you actually get the magnolia tokens for every ether worth of tokens that you traded on the exchange? You're going to get one magnolia token. So it's like a constantly inflating supply of tokens based on how much value was traded on the exchange and there is a specific function that determines the transaction fee that you have to pay. So when I say pay, if you also participate in the next auction you will very likely get some of it back because it does, since it goes back into the ecosystem, in particular, it always goes to the next auction. So that's kind of a way to incentivize people to participate in succeeding auctions. And yeah, the graph is a, you can see it on the thing hopefully. It's a step function and it's kind of very like social I would say because it works in like a log 10 manner. So yeah, as long as you have, what is this, 0.001%. So I guess that would be one, that would be like 100,000 of all the magnolia tokens in supply then your fee is reduced to 0.4%. It does say tulip, so there's been a rename and now they're called magnolia. Cool. So I'm just going to quickly finish off by, yeah, summarizing. So what are the multiple advantages? So there's, since everything is on chain there's no front running by any party apart from like, I mean this is with respect to the underlying technology that we're using, which is Ethereum. So Ethereum itself might have front running issues and we can't deal, we can't do anything about that but it doesn't introduce any new ones because there's no centralization involved at all. Right. So yeah, I mean, I think it's like kind of a summary of what I said. So we think this also will like make it easier to determine a fair price because it's an auction mechanism and auctions are very well known to be very good at determining prices of goods. So I mean, this is kind of like financial stuff but it's, yeah. And a huge advantage, we think we worked very hard on this is to make it very like gas efficient. So to make it very efficient. And yeah, it actually turns out that it's, that we did a good job. That it's very, very efficient. So one point to note is that, for example, Ether Delta. So if you or other order book exchanges if you post an order then it's very likely that multiple people will try to fulfill it at the same time but only one of them or there's a chance that only one of them will be able to fulfill it. Obviously it can be partially fulfilled but there is a chance that some of the orders or some of the transactions will not be able to fulfill the order they wanted to in which case there's a lot of gas wastage. So there's even a graph here, right here, which shows, I'm not sure which exchange this is but in any case it shows how many transactions are failing because they're trying to fulfill an order that isn't available anymore. It's 4 out of like 8 or something like that. Yeah, I mean as for the gas cost so it's pretty good that, for example, sellers they can choose a really low gas price because it doesn't matter when they submit it and the similar thing could be said for the buyers, the bidders as well. Oh yeah, and two more huge advantages that we see of this model is that it allows smart contracts to use the platform and the reason for that is that, for example, with EtherDelta you have to sign the order and then it goes into a centralized order book and smart contracts, they can't sign stuff because they don't have a private key. So in this model even smart contracts can participate because they just have to call a function of our exchange. So we think that's really powerful and it kind of changes the concept of an exchange. It doesn't have to be just an exchange, it can be a platform where some contract might want to sell or exchange tokens so they can do that using this as well. It means you can pay in one token and the person can get paid in another token if it is in the process exchanged on the Dutch exchange. And last of all, there's an on-chain price oracle which is actually one of the reasons, main reasons why we developed this is to have a reliable on-chain price oracle for all the tokens because with order books you can always sequence the transaction so that you post a really, for example, high limit order and then you fulfill it immediately yourself which means that for a very short time you will change the market price of the token and if you sequence your transactions in the right way then you can take advantage of that for your gain which means that the price oracle isn't reliable. Well in this case, you can do the thinking yourself, there's no way to do that. And yeah, so we have a front-end as well so it looks quite slick. So you choose, yeah, I think it's pretty straightforward, self-explanatory and you can use MetaMask or Gnosis Safe in the future and you submit your order and you can see the status. And let's see, so yeah, I mean, there's a lot of potential audiences for this right now we're actively looking for market makers which are professional traders so if anybody is into that business and you feel like you want to trade on this exchange please get in touch. So yeah, and last thing I'm going to kind of an offer so we developed this exchange we're also doing many other projects at Gnosis one of the other ones we're doing is GnosisX which is a competition for you to think and develop of prediction market use cases and yeah, so there's currently three categories and they all have a deadline in August and September and if you feel like you fit one of the categories I don't, yeah, I mean, you can access the information about the GnosisX which you can find for example on Medium and you can find the categories there and if you feel like you have an idea for one of the categories, you can apply and you can win up to $100,000 worth of Gino tokens and support with developing the application so that's kind of an offer that I wanted to make and with that, so yeah, GnosisX ID8 that's kind of the logo and oh, okay, so the three categories are Science and R&D Token Diligence and Blockchain Project Integration and we have an SDK that will allow you to do that in a better way cool and yeah, you can use this email address to get more information anyways, great, so that's it thank you very much yes, that's exactly what I'm here for use the mic this time when you have questions because it's much better finally for the live stream yeah thank you how do you get liquidity into the market? so you post the sell order okay, so the question is how do you get liquidity into the market and the answer is very simple you post the sell order that's that so the sell order will include the information of which auction you want to participate in and how many tokens and that's that you can either participate using the front end actually right now or in the first version there will only be a sell side front end so when you're participating through this you will always be selling and in the future you will also have a possibility to use a graphic user interface for the buyer side and yeah or you can just use the smart contract directly does that answer the question? cool yes, so the security audit should end next week I believe and it's deployed on test network on coven but mainly on a rinkeby and it should be live I guess within a month or a month and a half on mainnet yeah recently there's been a thanks for that yeah, so yeah, so recently there's been quite a few price fluctuation or large influences that people are selling off in large amounts influencing markets if you're selling in batches wouldn't there be a higher risk at that? would there be a would there be a higher risk of what? what's the question? higher volatility? yeah, exactly yeah I mean I don't see why so for example if there was stock involved if you did that markets would be quite effective if I'm not saying you have a batch mechanism in which she stops yeah so in fact I think the opposite is true so the reason is that let's say you have an order book I guess I'll just use a diagram it's the easiest so you have an order book and it looks something like like this whatever and so suddenly when you start selling a lot then this stuff always will move to the right it will move very fast you know this is very steep then it will move very fast and the price will change a lot right? well here if you sell a lot then it will not necessarily change the market price do you see that? so if you sell a lot in an order book it will necessarily change the market price but if you sell a lot in a batch then it will not necessarily change the market price because you will just have a larger sell volume this X will just be a much larger number and but there can still be sufficiently many buy volume so that the market price in theory could stay the same yeah cool yes I have two questions the first one is probably the most important is why you decided to rename from Tudip to Magnolia I mean sorry the second one is just out of interest how do you how do you model these things or do you model these things economically like how do you basically form an opinion of whether this model what effect this will have on a bigger scale if it becomes very popular okay so yeah I'm going to start with the first question I mean I don't know I guess the reason was that Tudip has a connotation of a bubble and we didn't want that the second question was how do you model this economically what effect would this have on the financial system if this takes off and if this model becomes really really prevalent well I think it will have huge advantages right one of them we just discussed yeah yeah I mean yeah you can just say at the comparison the disadvantages versus or the advantages they will be multiplied right if this is much more prevalent so yeah any other questions yeah thanks for the talk my question is if I was at the Dutch auction model correctly you choose this to reduce on-chain events and also to have avoid this concurrent transactions and waste of gas have you ever studied the possibility to use a lightning or raiden and completely go off-chain and do a matching engine off-chain and only when people go in and go out they sign on-chain so in this version it we did not use raiden but it is something that we are looking into so yeah this does have the advantage of even further reducing gas fees doesn't it yeah maybe totally eliminate those and it would allow a traditional exchange I think 99% of users now traditional exchanges much more so maybe they are more used to the conventional order books so I see a small risk for you that mainstream adoption will be hindered by this new kind of matching engine yeah well I mean so it's a new model and we'll see if it takes off we believe in that there's huge advantages and that it's more suited for blockchain than the traditional but you know you know the market has to decide on that yeah that's that cool any other questions yeah one more question by gentlemen in the third row as an auction maker as I understand if I want to sell I don't know my 100 ETH or what can I set the properties of the auction like for example the time no that's fixed that's fixed that's always 24 hours that's right so in order to aggregate all the liquidity we don't divide it by having everybody start their own auction everybody just all they do is they post the seller order and we aggregate all the liquidity and we have a uniform rules of the game that all the liquidity follows cool one more question over there yeah absolutely I mean yeah totally so there I mean there is a possibility for arbitrage right so in particular if the price ever goes below the market price in other exchanges then the auction will be closed because all the bidders will buy up that at that price because they can get it at a lower price than at other exchanges so that means that the sellers will actually never get a worse price than the market price at that point after the 6 hours so yeah I mean as a buyer or as a seller you can decide which exchange you want to use whether this one or different one so inherently there is always going to be arbitrage between the exchanges which is good because as a trader you should get the same price on all the exchanges and arbitrage works to achieve that not really here it doesn't matter what the market price is the world could be burning over here and that doesn't matter but what matters is once we get close to the closing I mean unless the so there is like two mechanisms going on there is time going on and there is price which reaches the market price that's very likely to happen after after 6 hours rather than sooner than that does that make sense or yeah so basically what I was trying to say is that the only time arbitrage can happen is at the closing point which obviously that varies where that is but as soon as the price drops to the market price elsewhere is when arbitrage happens before that it can happen because yeah I mean the bidders would just go buy it somewhere else yeah you said basically the price the option is the price is double the closing price that's right so the starting price of this auction is double of the closing price of the previous one what if that's already below market price what if that's what already below market price so that that would mean that the auction closes instantly or what so okay I'm going to see two things so yes in theory yes so if X is still smaller than or equal to the market price at this point then yeah it will close instantly and the second thing is it can't happen like from an economic point of view because previously if like as kind of my answer to the previous question if the price had gone below the market price in the previous auction then the economy the traders would close it then so prices will never go below the market price because otherwise bidders are incentivized to actually bid and buy them up right so or are you saying like the time difference will make the the price will change right okay that's a very good point so there's actually only 10 minutes usually between the two auctions so that is it's very unlikely that the price doubles or halves in 10 minutes what do you mean what do you mean the 10 minutes that's a very good question so the reason we put the 10 minutes there is like if you want to be if you want to claim the tokens like here right and then you want to participate in the next auction immediately so then we have time where you can like claim them and you can also sell them in the next auction immediately yeah yes other sellers like exposed to some kind of like volatility risk because they kind of lock that tokens up in this auction mechanism if the price moves a lot up or down then that could kind of be undesirable for them to have so they might want to not sell them anymore yes I mean obviously so there are many many advantages to this model one of the possible disadvantages is so let me start with the bidders so the bidders or the buyers they actually are like really well off right because they they come in like at the last whatever and they get the tokens very very soon so very very low lag and low volatility risk the sellers on the other hand you are right they will get their the tokens on average in 6 hours so yeah that's that we are aware of that like that is one of the disadvantage why should the sellers participate if they have this kind of risk or the incentives for the sellers so it's like it's all the other things together right so you have a smaller chance for the exchange to front run so basically this mechanism allows you to get the fair value of the tokens because there is no trust needed the miners yeah so there's no trust with respect to the protocol yeah any other questions no more questions cool so thank you very much Dominic thank you very much all for this great evening thank you very much the live peer team thanks many thanks to the Lavalra team many thanks to all our health groups for organizing this event next month there will be another meetup of course it will be more aim for beginners so a lower technical level I hope to see you next month thank you