EthCC 2: Christiane Ernst
Ethereum Paris·Wed, Oct 9, 2019, 12:00 AM
EthCC 2 EthCC, the Ethereum Community Conference is created by the community, for the community. It brings together great minds and forethinkers, and allows the community to grow. Follow us : @AssethFR https://ethcc.io
Transcript
okay hi everyone it's great to be here I also presented last year ET sieve the Dutch axe a decentralized trading protocol and I'm really glad to be here and give you a few updates and I'm all I'm glad you have noticed that it's really nice weather out at lunchtime we were running late so I'm I'm glad the old state and so my name is Chris I work for granolas we develop infrastructure tools for aetherium under the umbrella trade trade hold and for today's presentation we'll be in the trade part the title is the DX bow governance of the Dutch X so I've split this into two parts on the first part I would like to tell you a little bit more about the motivation and the mechanism of the Dutch X and then why it needs a tower to govern and then what the dot the DX bow is about and how you might want to participate or could participate if you wanted to so it's basically split into who here has heard of the Dutch X before oh great so it's a little bit of a refresher and a few new things going at so what is the Dutch X it's a trading protocol a decentralized braiding protocol for ESC 20 tokens which is based on an auction mechanism and more specifically on the Dutch auction mechanism so what's different it's not on a book based most of centralized and decentralized exchanges out there our order book based so we've gone with a completely different model for this version and to understand why I will also really briefly go into what the order book model looks like so in an order book model you have you have your bids and you ask basically place as limit orders in an order book meaning what they're willing how much they're willing to buy for and how much they're willing to be bought from basically and it's populated into an order book so these limits autos get matched as orders come in so it's really a pay by bit model which is a continuous auction and leads to discriminatory pricing of those people taking part in it and this this means that an order book in our opinion has a few flaws one of them is front-running which is a practice where actors act on information that they have ahead of time very commonly loaned this is actually as insider trading one version of this is also based on upon speed where you have high-frequency traders sort of extracting value and acting that they have big computing machines close to centralized exchanges for example what can also happen is you can have something known as overlapping other books which is basically if these two sides of the of the other book kind of overlap and there's no automatic matching of orders this can lead for people to exploit arbitrage opportunities you can also have a big spread so you just have an order book which is maybe not matched at all with the highest slippage meaning if it's a thinly populated order book a big market order coming in can really spike and slip the order book price and a lot so we've designed this particular exchange differently it's based on first batching orders and then executing them based on the Dutch auction model so the are you guys familiar what the Dutch auction is you start with a very high price and then a price that nobody is willing to pay for and then the price actually drops until people are willing to buy and the whole volume is filled so why is this a good model in our opinion and the reason for this is that those bidders are actually incentivized to reveal the true willingness to pay so they're not really masking their bids in either direction to have a sort of payoff and this is because everybody pays the last final price when the auction clears and not the price that they are bidding at so if they were participating later they fear of missing out in the auction if they participate earlier on and the auction closes they might have paid too much than that they were actually willing to pay so it leads to a single price clearing it's more specifically how does it work on the Dutch X so it's actually happening in two states that are separate from one another in the first state we collect sell orders so everybody let's let's go through an example token a photo can be everybody on this side of the theater wants to sell token a for token B so we're collecting all of all your orders over a period of time we close this auction pot and then everybody on this side of the order wants to bid with token P for token A and they will have to be active why this auction is running so they're very two states of the auction so this is exactly what what I mentioned first we're collecting then we're setting off this Dutch auction mechanism so we always start the next there's only always one auction running for a token pair at a time and we only start the next auction once the prior has cleared and we started at twice the price and the prize function is designed such to reach the old closing price after six hours so after six hours only over time what does this look like we we match the SEL others we kick off the auction the auction finishes payout are made and the whole circuit starts over again in time and all token pairs run independently from one another so let's go a little bit deeper what are some of the incentives that we've designed inherently into the protocol so one really cool thing here is that we don't really extract value from the protocol so let's look at this specifically we levy something and that is called liquidity contribution and we haven't caught it fees though most people would compare it to fees on an exchange but it's really not so what we do is we take well we the protocol takes between 0.1 and 0.5 percent of every traders trade and allocates it to the next running auction as an extra sell volume okay so it's being redistributed within the ecosystem this means that there is at least some liquidity for the next auction to run and also it gives incentives for continuous use and as we see now it actually benefits high-volume traders so it's a little bit of a redistribution between different types of traders how you can reduce your liquidity contribution is by actually trading so we have an inherent token which is called the Magnolia token which is minted for every ether worth of trades executed on the Dutch x protocol the more you trade the more Magnolia you generate for your own use the lower your liquidity contribution so your liquidity contribution might be lower than somebody else's effectively redistributing to somebody who is a high-volume trader and we do this because an exchange sort of lives by having liquidity on it so we need to give benefit to some of the actors on the exchange also important to note it is actually only certain types of tokens that generate this Magnolia they are called the whitelist of tokens and this is just merely due to the fact that there shouldn't be just a random token added which is really not being used only for the purpose of generating Magnolias on a sidenote our owl is the token that is generated by having GNote Oken the grossest token which allows our holders to actually settle part of these liquidity contribution in our this is optional and the owl use don't go to any party they're actually burnt and taken out of the thing but this is an optional thing in the ecosystem so a common question is now is it better than other decentralized exchanges obviously we think so that it might have some advantages but for sure it's also very different so let's go into some of the advantages that this particular model has there is one clearing price at the end of the auction for everyone so everybody pays the same price and there's no scope for front-running also not buy any minors for example because there are there's a but the price function is decreasing so slowly that there are many blocks mind for each price point we also think that it's particularly suited for low liquid tokens because there's such a long patching period whereas if you have the order book based model you really have a thinly populated order book and which also leads to prices price spikes when the order book is sort of gone through with them with a much luck with a large market order so it is for sure it's a less volatile also due to the fact that we only reach a price point approximately every six hours when an auction clears one could also argue that it might be very suitable for some types of over-the-counter trades because many over-the-counter trades want to have large orders for low liquid tokens which is exactly what you argue that the Dutch X is a suitable mechanism for we also do believe that it is much fairer price finding mechanisms I explained before why it's really good for the for the bidders because they have an incentive to reveal the true willingness to pay which is sort of the only strategy that they need to go with and this has an upward price pressures for actually sellers also we don't really extract any anything from this ecosystem and it's being redistributed within the protocol we've been struggling with two things so far one of them is this mechanism is slow so it takes some time for the auction to clear and this is not what people expect when they use an exchange they really want instant liquidity which is given for a bidder a bidder can actually extract immediately and this is nice to note however the bidder needs to be active at the right moment in time right when when the price reflects their willingness to pay and then they get this at a minimum or more when it closes later on however they also need it's like time sensitive something else we've been struggling with is that it is sort of a new mechanism which requires a lot of Education so so far the most commonly used exchanges are actually centralized exchanges not even decentralized exchanges and now we're trying to introduce something yet entirely new so we've been sort of struggling a little bit with this and so what else is so cool about this mechanism for one it's completely on chain and there are a few advantages that lead to the to us calling it an open protocol and I just go into these few when writing the smart contracts that govern the Dutch X and we really kept in mind to tailor it to the integration of other smart contracts there are no nothing is kept off chain the matching happens on chain there's no need to sign any orders off chain and the only strategy really you need is also to deposit a token it also provides a reliable on chain price and Oracle and we've created another one which is a million of the past nine auctions which is also purely on chain another cool thing about this is there's no party in charge so it's fully open source and anyone can list any ASC 20 token based on the same rules that are in the smart contracts so this should incentivize developers to build on top knowing that there's no central party in charge of just Abid lee arbitrarily changing the rules so now the question is having explained why it's so cool and having also explained that it's a little bit difficult to get going and get in motion and who or what ensures the liquidity so obviously the theory is it is the market itself and shows this and you've seen that there are some incentives built into the protocol the the mechanisms of the token economics that should ensure this however it's not that easy because it's a two sided marketplace and you need these critical network effects to actually kick off the auction to work fully as intended so we've taken a few measures to kind of ensure that it's easy to take part one of them is we've built a few interfaces that can be used and also some graphical user interfaces I will mention in a minute we've been trying to onboard market makers to the new mechanism and trying to explain how the Dutch X works in comparison to other exchanges we've also set in motion some large volume dye actions so good noses as a company has funded so far one and will fund another nine large dioxins where we actually input a hundred thousand die as a cell volume to be auctioned off the first one happened two days ago the next one will start tonight and then it's every nine auctions until we've sold 1 million die we're doing this to for the one side demonstrate that it actually works and on the other hand also give a little bit of an incentive to look into the mechanism and start building on top and integrating we've designed some generic minimum liquidity pots which are really only there to ensure that the auctions run smoothly and without any interruptions they can be parameter the parameters can be changed by anyone as they're open source and just used for their own projects and tokens we've also been building open-source arbitrage BOTS currently one to uni swap and there will be a pooling contract for people to just deposit and then constantly take part as a seller in continuous auctions and generate Magnolias another sort of important point is that an auction only actually sets off once there's a critical sell volume to give enough incentives for bidders to actually take part so which ESC 20s well it's actually up to the community as I mentioned before anyone can add any tokens so currently there are only a few I think die maker Mesa go Raiden I think Kai was listed right so how does it work how does listing tokens work as I mentioned it's permissionless and open there is no central party deciding on what gets listed and there is no deal listing we've created some developer guides however we have noticed is still not that easy even though we say it's really easy anyone can do it it's not it's not that easy and so we really welcome any feedbacks regarding our documentation there is no cost involved in listing it's completely for free obviously you would incur gas costs the only requirement is to set off the very first auction with the rep tithi pair and to fund it with 10,000 worth of ether or rapt ether which is then exchanged in the first auction so it's just a liquidity it's not really actually spent and we do we do recommend token projects that listen themselves to sort of ensure that the markers markets at least initially run and smoothly so now you're wondering how do i trade so we've been live on may night for a while these are the interfaces I mentioned before anything that's to do what the protocol layer called Dutch X you can find on read the docs you can find the command line interface guide etcetera there is a cellar platform so as you know as I explained before there are two sort of types of participants the sellers and the bidders and there are unique graphical user interfaces for each of them one is called a slow dog trait because good trades take time and the other one is called Fair Dex where you can participate as a bidder okay so the Dutch X is meant to be a fully decentralized trading protocol so most there isn't just decentralized and centralized there's really a scale of decentralization so what most decentralized exchanges exit actually are their non-custodial exchanges the Dutch X is also one of them we don't hold any funds from the users it's also a pure unchain mechanism where anything from orders to matching them happens on chain and to really kind of go further in that in that goal to be fully decentralized the next step is to hand over the governance of the Dutch X to the D X tau which is a decentralized autonomous organization and this is actually the second part of the presentation but first let's go into detail why we why we need a doubt to actually govern something that is already on the blockchain so we think one of the core value propositions of the Dutch X is that it's a global liquidity pool meaning that whoever bids on top who creates interfaces who builds other smart contracts interacting with it anything actually taps into the same protocol so if you would like to do a protocol upgrade we've designed the smart contract architecture that that's possible with a proxy pattern where you can upgrade the master logic the logic of the contract without losing any of the action history or past data or the liquidity or hopefully anyone who has been on top you could have gone another way and said an alternative to this would be to redeploy the smart contracts every time you release a new protocol but this would then risk splitting the liquidity because everybody would have to integrate again with the new smart contracts and protocol upgrades are actually quite important they might happen for technology reasons but they could also just mean that you need to adapt to regulatory requirements another thing on the Dutch X particularly in comparison to others is that some parameters might need more frequent changing and just to name one it's the the whitelist so those tokens that generate the Magnolias token and there might be more that need to be added to this to this list and somebody needs to decide upon this which tokens get gets added and legitimate to generate Magnolia so who gets to decide this it shouldn't be a central party and it shouldn't be gnosis or anyone it should be somebody who can be trusted by the users of the exchange so this is really where another DX Lau enters and we call it a bold experiment in the future of human coordination with the goal of making better and decentralized decisions so when I go through the next part of the presentation please do keep in mind that it's at this point it is an experiment that we're talking about and so who of you has heard of the DX style before ok quite a few so what's a Dao again it's a decentralized autonomous organization for which the rule set is actually fully hard-coded into smart contracts so it's a new governance structure based of proposals and voting and then an execution of a result which is predefined in the smart contracts why are dolls so important were they enable decentralized coordination at large scale and they are really needed to have decentralization at every layer of the stack what's the DX Dao the DX talisman to be community-owned meaning that hopefully a lot of the community participates in it we are making it really easy to take part there's no pre allocation of the voting rights and we'll go into that into detail in a little bit we also think that it's potentially capable of growing beyond the Dutch X and I'll explain in a minute why but what could a doubt do for the Dutch X other than change parameters in the protocol it could for example change the incentive mechanisms it could fund further open-source development it could incentivize liquidity contribution so essentially it could do a lot of things it's all powered by dowsed X so it's using the Dow Starck holographic consensus framework and for this so the DX now beyond the Dutch X so we have the working theory that the DX doll would solve some problems that the Dutch axe faces which could also be really helpful for decentralized finance applications so what can it Dow in general do it can actually do a lot of things that can do can set out bounties it can take part in prediction markets it can trade on fully decentralized exchanges and anything but handle off chain permission assets really who governs the DX Dow the short answer to this is the dig star reputation holders and so who are these reputation holders who get to decide what is reputation reputation is not really a token it's associated to one etherium address and it's non-transferrable reputation grants voting rights in a specific dow so DX tower reputation grants voting rights in TX now it's not one voting right address but it's really the amount of reputation held by the address translates into that voting power and reputation holders then get to vote on proposals so how can you participate you can stake your vote and earn this initial reputation the reputation in the DX Dow and so a million reputation is generated at the beginning without pre allocation to anyone so everybody can take part in the same manner and I would go through what that exactly means so we've created a few categories trying to on the one hand side side make it really easy for interested parties to take part and be of a government role in the DX Dao but also to link it to the Dutch X to really make sure it's the users of this initial decentralized application that the DX governs who are also the ones deciding upon it and there will be a 30-day limited period in which it's possible to earn this reputation it hasn't commenced yet I will go into the time in a little bit but first there are five categories that you can see on there I will speak about the first four the last one really is more about spreading the word where the distribution of the reputation is then actually decided by by the DX now itself so let's go through the first four to see what that actually means within this 30-day limited vote staking period users are able to come to an interface and lock eath locking if means they are putting down a value of East and a duration of four how long they would lock it for up to a year this means they're paying in opportunity cost because it's locked and they will get it back at the end of this period we are distributing within this category a fixed absolute amount of eighty thousand reputation which everybody who takes part in this particular category will get a share of depending on how large the reputation score is which is calculated by the duration and the amount of ether lock locking an ESC 20 token actually works in exactly the same manner we've allocated three hundred thousand replication of the 1 million for this particular category and there's one more thing because to make it comparable how much of Omi's ago you've staked worth versus Gino we calculate the ether price of this for this we need a price feed we take the price feed actually from the Dutch X so you see we're going a little bit in a circle here and for this it needs to be whitelisted n traded on the Dutch X to be lock a lockable token the third category is registering Magnolias we are allocating half of all of this particular to the users of the Dutch X who would like to take part in this it's 500 thousand it's a little bit different than the other two categories because you don't need to move actually your token you're just registering your account to be then eligible to receive reputation based on the amount that you hold in Magnolia at the end of the thirty day period jen is the native dialect token and Dao to function actually needs some funding so the Dao itself will later need to decide on how it will earn money for example by levying something from the liquidity contribution of the Dutch X or whatever model they are going with but at first they also need some kind of funding to set them often to kick them off so there's a possibility to buy reputation by bidding jen token and there are ten auctions throughout this 30 day period each lasting three days and essentially within the the auction itself the reputation the ten thousand reputation allocated in this auction goes as a percentage to people who took part with the Jen the funding then goes to the DX tile so it's not returned to the people it's more a buying of reputation for Jen so this is a little picture on how a sort of the reputation is distributed in a larger scale we went through all of them so this is not a new it's not new at this point so when thinking about this period a few things to keep in mind each of the categories is separate in the sense of there's an absolute of rapid an absolute amount of reputation allocated to the category and you're in competition with people who take part in the same category you can take part as often as you'd like during the 30-day period obviously for registering Magnolias you only have to do this once the reputation is then accumulated across all the categories that you took part in this is currently a landing page but it will later link you to the interface that you can actually take part in and some people have asked me why how is then reputation distributed after this period so there's one mechanism inherent into the bow step framework which is for successful proposals the dau mints new reputation however the DX Dow is then free to decide it it will live on its own basically and is free to decide on how they would like to further distribute voting voting rights what happens after the votes taking period so this is a finite period of 30 days and the DX da goes live and this is where the full community governance will actually commence it's it's relatively easy because the dowse tech team has developed an interface called alchemy earth for which it's possible to take part in their framework which is submitting a proposal voting on a proposal boosting a proposal so the entire and it's easy for users to participate within this interface so I just want to go hold circle because I've split the presentation in two I just want to tell you again how it all fits together now it's a little bit overwhelming at first but let's go through this together so let's move from left to right on this one let's see do I have a pointer okay the Dutch X decentralized spread trading protocol you trade on the Dutch X you generate Magnolias the Dutch X also has a decentralized price feed which can be used which needs to be there in order to have a price feed for locking in ELC 20 token you can also register Magnolias the middle part is really what is the vote staking period which is only there for 30 days and after that it'll go away it generates the initial reputation which gives you the right to vote in the DX pal and then the governance phase has the entire set of the DAO framework and the holographic consensus where you have the in putting the proposals and the voting so the DX now later then can change the parameters and upgrade the Dutch X which goes the circle so when is it happening I want to say soon so we've delayed once before and as we'd run a bug bounty where something was found and we had to do another iteration on the dowel stack model and they're currently in the second audit which is going really well so we're waiting for this to finish but to then start and commence another bug bounty there was also a demo version which just ended so we probably do another test run and then the votes taking will commence probably sometime after mid-april so still relatively soon it's a 30-day limited period after that it's really really important to me to note that good gnosis and I'll stick with step back it's not ignore sis now it's not a Taoist Act ow it's really a community-owned Dao and we want to for the DX tower to be established as a neutral web 3 piece of infrastructure I have another slide maybe then at this point I just want to make one note that dowels are still in uncharted legal territories and so the legal framework is not really clear at this point so everybody who takes part actually takes part at their own risks so a dowel for truly decentralized apps it's more like a question than a board statement do we have time for questions I think we have five minutes thank you for the presentation about the Dutch auction model I've been looking actually forward to see teams using that model to run their token sales and for some reason that hasn't been happening a lot and the question is a little bit why and I guess part of the answer could be that it carries a bit of a reputational risk because taking for example the Rosie's auction some people said that the auction kind of played up to people's worst instincts of fear of missing out and it could be said that it ended up you know all the valuation I don't know if it did I'm making that argument but that could have been you know what one of the things that puts teams off from from earning these kind of auctions the second is also that there's also the risk of concentrating the ownership with so-called whales maybe reducing the potential number of you know of people that own the tokens and thus token all those you know they can become your developers your evangelists and so on and so forth and maybe I can answer a few and I'm done and I'm just gonna fish up its up which ways would a team that wishes to run a Dutch auction you know how could they mitigate this kind of potential pitfalls and downsides of this model thank you okay sure maybe I'll answer this in two parts the first one I'll apply to the Dutch X model so this is not really suitable for icos because the token would have been already created and then it helps decentralized listing and actually making it more diverse and spread by being on an exchange then rather being accumulated on maybe in one centralized exchange where it's listed so it actually kind of goes the other way from the motto the point that you were making and the other point that you pinpointed edie in the eye cos the fear of missing out for the Dutch X the two user groups you have the sellers and the bidders we actually argue that the sellers are typically people who don't know what price perception they should have and what they should expect and with the economics working of the model they should get a fair price so sellers are only required to deposit whereas I sort of agree with you as a bidder you need to have a certain price point in mind so you need to be educated at what point it makes sense to bid so what can happen if you're an uneducated bidder and you bid too early above the price point that you're actually willing to pay and the auction closes you will have overpaid so it is important to educate a bidder to not do that but only participate at the right moment in time but this is a risk definitely that an uneducated bidder might overpay so there was a team who built a better interface and we actually asked them to implement a warning a red sign so FedEx or net is life on Maenads you can try it out but if you are overpaying by I think 10% of the last closing price it pops up a red message saying are you sure you want to participate so this is kind of how we try to mitigate that risk does that answer your question all right we're running out of time okay let's let's let's go for lunch I think and we'll get bet there in 30 minutes cool thank you [Applause]
Automatic transcript — names and jargon may be misspelled.