DAPPCON 2019: Multi-Conditional Tokens as a New Crypto Primitive - Friederike Ernst (Gnosis)

Transcript

it's a pleasure to be here so let's get cracking so I want to speak about conditional token standard that we just released recently and I would like to introduce a product called site in the end um so can I can maybe get a quick show of hands who here knows what a prediction market is okay so maybe about half so let me quickly go through this so a prediction market basically sent us around a question that is not known yet but will be known at some point in the future so for example here the question is where the AirTrain in Manhattan be repaired by the end of July 2020 it's

impossible this is the current prediction but as public infrastructure projects go you never know right so what you do is you actually set up this market you formulate a question you specify the possible answer so in this case it's yes and no and then you set the initial probability so say for instance here it's set at 50/50 but you could actually have it started an arbitrary probability and these in the end of the market so maybe you you set a date at which this will be this will be resolved and what happens then is if you

hold the yes token and this actually turns out to be true and then you will be given the collateral token that is backing and this token and if it's if you hold a yes' token and it doesn't turn out to be true you don't get anything so you don't get a payout what that means in essence is that these tokens the yes and the no token they should in principle trade for the further probability for the the problem the price the price that reflects the probability of this happening or not so basically if you think the probability that the mta's go delivers 80% and we should trade for 80

cents on the dollar um and because it's a mark I mean it's it's a liquidly traded market so the idea is that this probability actually updates in real time so as soon as some new piece of information comes I would say for instance the MTA is having troubles with one of the tunnels under the Hudson and then this would be incorporated into the market and the probability of this would drop hence the price of the yes token would drop so far so good quit fantastic oh so this is this is the idea behind prediction market markets

what you usually have is you have different kind of metrics that are actually interrelated so what you actually want is you want to have conditional tokens where the tokens themselves I mean the where the tokens themselves you can see the correlation between the actual events that pack them so this is a little bit abstract so I give you very I give you a very blockchain example oops No okay so this this is an example that sent us around the make at our ecosystem so who knows how make it all works small s okay so 2/3 so basically um make

it I was going to switch from single collateral die to multi collateral die sometime later this year probably and they would have the magic collateral die together we're seeing it in in addition to the single collateral die for at least six months so basically these two systems will be running in parallel for at least six months so these questions and kind of reflect this so the first question is will multi collateral die go live on main net by November first this is the first friction and you could have a prediction market as to this as where them the l-train ii market is where mostly collateral die

over take single collateral die by a certain date how do you measure this you simply see how much die is being created with one contract and how much with the other right clearly these two are interrelated right so basically if if multi collateral dies there's no way for multi collateral guy to actually overtake single collateral dies and then we've thrown in a third market here where the stability fee for single collateral die go up from today's twenty point five percent and if you hear the correlation is not so clear right so basically with with the where the stability fee go up from today's from

today's civility see if you throw in another another system that can be a way you can actually create CDP's with other collateral backing other than either or not we don't know there's some correlation but it's not clear what exactly the correlation is so clearly you want to be able to trade these these markets in a correlated fashion and I'll show you I'd show you a demo in the end so bear with me while I explain to you how would used to work and how it works now okay this is little bit here we go okay so how it actually used to work is you took

one collateral token here and you put it into the event contract so that actually bought you a complete set of outcome tokens so this is an event that actually has three outcomes outcomes a B and C so you would be given a complete set of outcome tokens and then you can sell them off individually and trade them or you can actually use an entire set and trade it back for one collateral token if you don't write it back eventually when the when the event is deliberated one of these will be worth the collateral token and the other two will become worthless so um how does this work with nested markets

now principle you can have them and they would look like this okay so now you have two events so for the first event you have outcomes a B and C and just because having another one with them three outcomes got kind of messy I made one with two outcomes so this is has the event one has outcomes a B and C so you put in your collateral token you get these three tokens they are each an ER C 20 token again so you use them as collateral you put them into the event contracts that ask iolo and then you end up with six nested tokens so we're basically a and high and low B and high

being low C and high and low you get the idea so this is this is somewhat arbitrary and that I could have done this the other way around let me show you doesn't like me okay here we go so this is this is the same this is more less of the same market it's just turned around so basically now the first question MIUI you're asking is is it high or is it low and then in the second round you ask is it a B or C so you end up with six tokens as well but if you compare them okay you compare them they're not the same see they actually match up one-to-one so

basically this one is a in high and this one is high in a but they're not fungible so basically they are worth 1 in the exact same conditions namely if a and high are true but they're not fungible you can't just trade one for the other so and this is just with two conditions this becomes even worse if you add more conditions so basically what we did is the following oh no please work here you go so this is the conditional token framework and this is this is the idea so basically you specify a collateral

token that you actually need to agree on and then you can specify an arbitrary number of conditions so here it's the same two conditions that you had earlier high or low and a B or C and it'll give you are the crossover token so you have the same six tokens that you have had earlier but now they're actually fungible so now you only have one a in high token and it's exactly the same as earlier so what does what does this actually give you it gives you a couple of really nice properties so as I said in high-end high-end a now I the actual same token so you can actually they're fungible you can also have partial

Redemption so you can actually imagine if they you know that say you have this market and now you know now now this part X resolves so you know a turns out to be true what you actually do is you know other ones that actually have B and C in them these four so B and I be a low C and high C and low they become worthless and what happens to a in high and a low Sarah no they just become high and low because a is true right so basically it's its high-end a but a is true so it's just high-end oh so this this in effect it allows you to resolve

markets partially so basically if you have different conditions and they don't resolve at the same time you don't you don't actually have to wait for the longest one you can just do this at any time okay what else you have you can buy baskets so you could for instance say I would like to buy a basket of everything where a does not come true you can split positions so this is actually a super nice one so basically if you look at this you could just add a another condition that say X or Y and it would just double the number of tokens you have the atomic tokens so it would you would get an X talkin for this

and the y token in the next token the white hope in and so on um and you can do this after the fact so you can do this when you've actually when you already have all of these tokens you can split them after the fact and you can also merge them after the fact oops okay so how does this actually work this works in a fashion that we're typically not used to and that seems like it should be a bad form we call it the monolith so it's something that people typically want to move away from but it here it's actually deliberate smart contract design choice so how does

it actually work so basically our conditions I handed I handed in one single smart contract so easy there's one smart contract that lives on the Ethier young chain and that has other conditions listed in in them that you can that you can subject your outcome token to we there's one restraint so in principle you can have as many conditions as you want but you can only have 256 outcomes that's per condition which should probably do it for most cases yeah so this is just for guest reasons and this this monolith smart

contract is permissionless there are no fees when you add a condition you have to add the Oracle as well and anything can be certified as an Oracle so basically if you if you go back to Edmonds Edmonds talk from just before now he lists at a couple of Oracle solutions and basically this this thing is completely completely agnostic towards your Oracle solution so it's plug-and-play really so let me show you a so I was actually going to show you a live demo of this but the Internet in this race is actually forbids that

because there's nothing live about the internet here so let me so I actually have a video that I'm going to pretend the live demo so bear with me okay I hope I can pretend this is a live demo okay so this is the market from earlier so you have with my equilateral Daigo life on Maine net by November first with my equilateral die over take singer collateral die by March 31st where the civility fee for singer collateral die go up or down from today's value and I will show you okay so basically here we

select yes for one of them and we buy this so basically this is this is the case that we had earlier with the l-train this is just buying one uncorrelated market so chain vanilla let me show you how it goes through so you place the order for 0.1 of an ether okay okay so that went through so now you actually have a sake that you can't see because basically it says transaction processing which is a lie because it's been done so now let me show you something else one more thing

you can do is you can say what is the probability of something happening given that something else is it has happened or will happen so basically say for instance what is the probability that my T collateral die over take single collateral die given that might equilateral die goes live on may net by November first so what this would actually mean is if magical etre died doesn't go live on main net by November first this would void and he would just be returned your collateral token and if it got if it does go live the condition is met so you've made the try a trade as you wanted to so let me show you how

this looks in the interface so here you click this should be conditional so you can see that now the interface changed so basically you're assuming that this is going to happen and then you say if this is going to happen then the that then you have nested probabilities what is what is the probability that the other two are going to happen you can place trades on these conditional trades so basically in the case that these are completely independent markets there will be no correlation obviously and the probability of a time's the probability of B will just be the joint probability

but in case there are correlations or cosy occultations actually you will be able to tell from the way that the probabilities look so let me quickly see show you how this works so basically now we're assuming that market true market true is going to be true and then you can buy whatever out count open you want and this was actually this was recorded earlier I really wanted to do this late life but yeah okay okay let me go back to the presentation okay so um I told

you earlier that I wanted to introduce and I wanted to talk about the conditional token framework and in fancifully this this is something that is open its infrastructure anyone can use it anyone can build upon it anyone can add conditions to the monolith smart contract um but we have actually built an application that runs on top of these on top of this conditional token framework work it's called a site and basically it's information markets on anything but sports so we're not allowed to do sports betting it's it's a regulated application as you are no prediction markets are touchy from a

regular regulatory point of view so because they kind of straddle the divide between gaming which is regulated or forbidden in most countries and financial markets which is also heavily regulated there are things in between that are probably not regulated in most places but it is a touchy subject so we actually obtained a license it's currently still in the in principle stage it's we'll get the full license later this year it will be it'll be for offering prediction markets to retail

customers in Europe it's non-custodial so you you you always you always have your tokens in your own wallet it's fully collateralized so it's it's it's different from other products which give you higher leverage it will be nominated and died just simply because nominating something in a volatile acid often doesn't make sense or you want something that that is somewhat stable because it's regulated we will also have to do a mln kyc we're trying to make that as painless as

possible eventually we will also take up trading features because we also have to cover a.m. and kyc costs entities will be able to request markets so we will actually start with blockchain based markets so things like Willy theorem 2.0 beacon chain go live what will the maker stability fee be we'll die be reasonably stable well Project X launched their offering and what's going to happen to the price of ether we will in the future branch out into other into other domains one of them being politics because that's

that's interesting to many people but also other things and companies will also be able to request markets off of us so basically for instance say BMW would be able to come to us and ask us for market how many electric cars will be registered in Germany next year and we would be able to offer that to them and would I would have um would allow anyone who actually passed the a mln kyc standards to actually trade on that market um what's particular about the side product is that there's an automated market maker that is integrated into it so we think you have at every point in time you have a

probability against which you can trade up or down and it's the same no matter whether you actually trade amadon it moves so basically say the probability is 50% and you say I go long then the probability is going to move up with it so you can only get it for to go in either direction for an increment bit and how much it moves depends on the seed funding of this market but in principle there will always be one price no matter no matter whether you actually buy it up or down and we think that this is super important for the user experience for the user because I mean what what you often see you on prediction markets is

that they're fairly illiquid so that basically if you if you buy a position and immediately sell it after that you will actually lose a lot of money and we think this is this is something that really trans users off okay so what's the road map for this so what you can do is you can sign up so we think we're still testing interfaces because in principle we can I mean we can support super simple markets things like where the AirTrain be whether you're air train repairs be done by July 2020 but we can also do the correlated markets that I

showed you earlier and as you can imagine doing a good user interface for the correlated market is really difficult because it's it's non-trivial what what positions you actually hold and in which cases they actually resolve to one and in which cases they don't and creating a good interface for that actually takes some time so we will launch a closed beta in early October of this year if you want to participate in this closed beta sign up and we will send you an invite because it'll be invite only and you can you can sign up at site dot p.m. I think it just requires the email address you

have to specify how much crypto knowledge you have okay this is a somewhat related announcement so we have we at Knossos have the gecko fund then the gnosis ecosystem fund and if you there's there's a new wave of grants coming if you want to find out more about this I go to gnosis dot gnosis / gecko and we will be awarding grants of up to $100,000 in September and we're also looking for a couple of colleagues so if

you're interested in working at a really nice company go to go to who puts that greenhouse that io / noses we're currently looking for a number of Engineers cryptographer but also an HR manager marketing manager and an office manager fantastic thank you guys [Applause] questions it was all crystal clear oh I think there's a question there in the

middle in the middle great oh and where's the mind here we go banana here at the front okay so I'm I hope I didn't miss the point in the presentation but how are you gonna resolve the markets in the site product sorry I had trouble understanding so I hope it didn't was wasn't mentioned in the presentation I missed it but how are you going to resolve the markets and the site product this is so basically the market and the question is how are we going to resolve the markets on-site and those are going to be resolved by gnosis so gnosis is the Oracle next one somebody I have a question yeah why did you decide to GU

regulator can you introduce yourself as well that started doing doing that stuff so we know oh my name is Gonzalo Ferrara's and I like prediction markets so the question is why did you decide to be regulated and the other question is what's up with the 1% fee do you really need it is it too high okay so let me answer the first question first so it seems so basically if you look at prediction market projects out there and there are a couple of projects that actually when the fully decentralized route so Olga they build a fantastic

piece of software you have to download it you have to run it on your own machine and then you can then you can use it to interact with the etherium blockchain I don't know whether have you used it it's difficult to use and the reason it's difficult to use it's it's because it's it's incredibly hard to build because you actually have to build it as one single app that that you actually run on your own computer updates as hard and it seems that if you if you don't actually manage to be fully decentralized and being fully decentralized and still having a market mechanically there two things here right so basically you need a you need a valid market mechanism so you

need something that that actually provides enough liquidity for people to actually trade either way and you need for that that's typically somewhat centralized so actually doing that in a completely decentralized fashion as difficult and if you don't if you don't manage to be completely decentralized the experience has shown that you will be shut down so you will be shut down as a business I mean this has happened to other projects in the space so it seems like there's nothing in between being completely decentralized and being centralized / regulated so basically being in between is impossible as seems to be impossible and being being

completely decentralized is technologically really challenging so we've decided to go with the regulated route for now and I mean in principle this doesn't so we see the conditional standard the condition spoken standard is out there anyone can build upon it I mean that's that's the way that is is that is intended and basically the way that I see it is in you time people will be able to interact with it through our interface or through other interfaces and it's their call so um and the one percent rating fee so we won't be loving it in the beginning we we do actually incur some costs so basically if you kyc in a ml screen someone it'll cost you about five euros and then per customer

you also need to repeat it periodically you you have certain other obligations that you are put under as a regulated entity so yes being regulated just increased costs and this is why I mean basically the one percent is not set in stone if we find that a lower trading fee also allows that allows us to not lose money with it we'd be happy to go with it too and but yeah so we do need a trading fear