Whiteboard Series with NEAR | Ep: 21 Alex Chapman from Augur |

Captions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.

1. 0:03

   hi everybody this is Ilya from near and we're here in New York on the rainy day discussing was Alex not Inauguration myself Alex Kevin lead Evan auger at the forecast foundation we're a group of software developers right now working on a prediction market that runs on aetherium so I guess high-level overview of how augur works yes yeah sure and I'm not great at drawing too much in a whiteboard but I can kind of give like an idea what the pieces are just a

2. 0:38

   really high level view I actually never spell augur like a failure you type every day literally never spell it every no because it autocompletes to a you know if I've I've misspelled it I should get forever but you know if not done nothing so I guess the two biggest components of this stuff we make are a trading component and a oracle so trading component of augur kind of you'd expect in a prediction market people can make markets by themselves you know but we don't actually make any markets

3. 1:17

   totally this head applies people just come in and make their own you can make a market on something like political like who's going to in a particular election sports games or something more sophisticated like some sort of a specific business hedge you can make with scalar markets which is something we have prediction markets to track like a particular stock price so there's a lot of different functionality you can get out of the trading park portion there's also not it's not that difficult

4. 1:42

   to make people can make markets and exchanges all the time using smart contracts what actually sets order apart is the fact that we have a decentralized oracle so decentralized oracle means when someone makes one of those markets like an election market for example and it passes the deadline where the election results have come in the block team needs to know and the contract needs to know how to pay things out so people who bet on the winner need to get

5. 2:07

   that the money that they deserve and the people who didn't bet on the winner they don't get to get anything a lot of the ways that people solve this without using specialized Oracle typically revolve around just a trusted party going in and punching in who won and then the results come out the problem with that is having a centralized party in prediction market has a lot of risks associated with it pretty much every prediction market in

6. 2:32

   history has been shut down attacked or has ended up being some sort of like an exit scam situation so the decentralized oracle is really the super-complicated t curse or a part of augur so the way that our work works somebody's calling we have a token called rep like I said I'm bad at white board by the way so I'm just switched to talking I don't know your guy anything yet we have a token called rep when a market is supposed to

7. 3:02

   end the market creator had designated when they created the market someone to report so someone to go in and do that manual process I subscribed where they go in themselves and say yeah this guy won the difference is that once those results are up there's a time period where anybody who owns rep can go in and say that's not true and there's a preceding sequence of disputes that happened where the amount that they had

8. 3:25

   to put up gets higher and higher and higher which if that was the only system we had which is the way a lot of other decentralized Oracle's work would be a Keynesian beauty contest which just means if someone just find more people - exactly right there right and whoever ends up having the most just wins and as a rational game theory participant you should just go with whoever has the most rep yeah which means we wouldn't actually be a good decentralized oracle it'd just be a Keynesian beauty contests

9. 3:50

   system so the interesting thing about our descendant Oracle is at a certain point we actually fork auger within the ethereum blockchain so not like an aetherium fork like an actual system where our contracts are capable of forking into two different auger systems what happens is if in the election market for example if that reached a fork there'd be a fork let's say with the 2016 election there'd

10. 4:13

    be a fork where Trump wins and a fork where Hillary won or maybe Bernie I don't know and actually there could be all three because any one can make it was a categorical market for example anyone can can say I would like for that child universe to become its own auger once the pork occurs the intern with that is rep holders have to choose where the RIP goes right right so even though someone could potentially

11. 4:36

    have a lot of rep say I have 50% of the rep I'm gonna I'm gonna say Hillary one because I was betting a lot on that they can do that and they will end up getting the money that was in the market so it will pay out for the Hillary winners but they have to spend a lot of rep to do that and because their rep has to go to Hillary universe which presumably no rational trader is ever gonna want to trade on they have to burn a lot of rep to do that and that's kind of the the crux of our whole security model the

12. 5:00

    fact that they they have to make their rep go to a universe where it is probably worth nothing hmm so all the security calculations we do everything in the white paper is basically based on this notion of them having to burn more in rep than the markets open interest is so even if they got all of the money from the market they've been overall lost money by performing that attack that's yeah that's sort of a super high arrow and

13. 5:27

    I've drawn absolutely nothing I'm sure there's a million questions yeah sure yeah I mean let's let's start me this training just to kind of your basic understanding houses all set up so usually there's just a template of a contract that anybody can deploy just spin up a new prediction market yeah yes all those are called markets market it's actually a market contract yeah okay I think it's creative and then yeah so pretty much people just deposit like if it's you know binary decision they

14. 5:59

    deposit on either yes or no or right writes it so I'll actually draw the different types that we have so the distinct types we have on a contract sir yes no markets you can think of is like the binary sort of thing where it's you know will X let's do everything in terms of the 2016 11 whatever any sense always talk about so it's will will Trump win yeah so it's either yes or no in the in the UI that we have the reference UI

15. 6:23

    right now for example you just say yes and then you have to actually sell yes - I see you sure - yes play resume yeah and we have categoricals which we call cats and these can have up to eight outcomes and the reason for that limitation is not a fun one it's just a cost sort of thing so on the etherium network if we like you'll need to like fork into aids actually it's not even that that that part technically with scalars you could fork into like hundreds of thousands of

16. 7:00

    universes but but someone has to pay for that if someone was like yeah no I think this universe was correct they're free to make it basically the way that the forking works the categorical is it's that the number of shares that we have to generate and when you're trading the number of shares we have to send to people including like emitting logs for all those transfers and just all the associated costs with having that many shares makes anything beyond eight kind of get a little wild

17. 7:25

    I think it was there was actually the cost of just creating a market was too high originally we could actually probably increase that but there would probably be negative side effects they would provide such a bad user experience that it would essentially seem broken to anybody like their their worst case scenario trading situations where an eighth outcome categorical market if if both people are reversing their

18. 7:48

    positions across all like right look if you're reversing your positions because somebody had to have been short which means they so this kind of gets into synthetic yeah like like for categorical yes no okay I buy yes or I'm selling yes for category Co you either buying one decision you can buy multiple yeah how does that work so buying multiple you can go long multiple which means make multiple trades the key thing though is if you short so the way that other thing works in the underlying contract system

19. 8:18

    is all of our sorts or synthetic shorts so it's easy to understand with yes no there's a yes share in a no share if you short yes if you say sell that you'll get no shares yeah it makes more sense categorical if there's eight candidates and you say I want a short Trump yeah that actually what the system has to do is give you shares of all the other candidates so if sort what does get into with the synthetic sort problem with reversing positions is if you're reversing that means it has to Train

20. 8:42

    for eight chairs per person twice and has to admit logs for all those and do all of the accounting for all those individual chairs so what is the problem happening is a single trade if they're both reversing positions costs like 3.5 million gasps so we could increase it like that's still under the block cast limit but when you get that high-end estimate for you need to like pay for that and get into the block getting in

21. 9:06

    the block is the weird part because then you enter the part where paradoxical a very toxic Lee you have to pay the miners more more high yeah it's sort of a hard sell to be able to fit anybody else so we could we could have that but it's one of those things where I suspect people using it would be like what this is broken what is going on this can't be right yeah categorical is a data type you can have eight outcomes max on those

22. 9:29

    and in scalars scalars are not used very much in v1 I hope they're used more in v2 because these are things are the neatest type of markets so scalar markets it's like a margin Coles kind of yeah I mean it's so there's no actual margin things are sort of leveraged II but not not actually like you don't actually borrow any money or anything but the way that my work is bet on the

23. 9:53

    price of Apple and then you could set a range between it's like you don't trade yeah I actually don't know what Apple's worth right now but let's say it was somewhere between one and two hundred sorry 100 and 200 you basically tracked that price so the way we have everything set up in the UI and it works differently underlying on the chain but everything in the UI for example if you thought it was gonna be 150 you could buy buy it anywhere on a hundred fifty you would actually pay that amount so if

24. 10:19

    you someone's going to sell it a hundred forty-five you could buy a hundred forty five if at the end of market resolution identity 150 you would make five dollars mmm and that's how the payouts actually works the payoff system for everything is actually actually works as a scalar market interestingly enough yeah it's and we could go through all that all that how that works on contract but yeah the the scalar payout there's actually two shares along in a short and there

25. 10:44

    payout is kind of on a range and that's how that helps yeah yeah like settles that shirt right at some point yes it isn't those are the three classifications and market and again on chain we have three different creation functions but they're actually all the same thing and they all work the exact same all the math is the exact same on all of them oh yeah I mean this is just it right it's much simpler so you can imagine I'll actually go into it so there's like a thing called payout we

26. 11:09

    call them payout numerators which is a misnomer right now actually in the contracts it has to do with how we payout invalid markets but in yes-no you can measure in there's there's a I think the first one of the yes share and the second one's a no share don't quote me on that I always forget which one it is but so there's two errors in computer science off by one yeah exactly so so

27. 11:32

    you can mention this is the payout for the share and is the payout for the no share if the market resolved no no Trump didn't win and this tells kind of the chain how to pay out the different chairs in a scalar market it would look something more like you know the reason by the way I'm making everything add up to ten thousand is sort of a artifact of how the contracts work for figuring out how to pay everything out and what

28. 11:59

    prices people can bet on chairs do usually like ten thousand of something so there's this concept in the contracts called num ticks which everyone kind of dislikes cuz it's complicated but we have not thought of a better way to conceptualize and program this notion but the basic idea is when people are trading and we will get chairs paid out there has to be some notion of how many

29. 12:23

    ticks they trade in seek imagine if you were buying something like on an exchange and you want to set a price for a share you might think it's worth eighty nine point five cents but and that wouldn't be possible if that exchange supported 1000 ticks I think where's the hundred texts yeah 1000 days because they're busily when a thousand units have increased you could have you grew from eighty nine point five to eighty nine point six day and

30. 12:47

    point seven if they didn't want to deal with fractions of a cent you'd only have a hundred ticks you can only you only deal in that many ticks so our our underlying system for yes/no in categoricals has 10,000 ticks right now it's an eighth so that actually is fairly close to like trading assent with current prices yeah so the issue is this is the payout is also based on those gnomes and that makes it work out really really nicely with making sure that all of the

31. 13:12

    the funds are actually like covered and we don't have to do any weird rounding or division or anything like that we just limit what you could have actually traded and what you get paid out scalars are more complicated scalars we let people do custom amounts because ten thousand takes does not always apply to every scalar market you could have something very complicated if you what

32. 13:28

    your trading Berkshire Hathaway shares or something you know like it doesn't yeah it has to be custom categoricals worked the same way but obviously there's more there's like for each category yeah right so this is Trump Hillary it's Bernie did it guys all right yeah I guess Andrew yang is the the burning of this election oh there's Bernie mysteries is burning this lucky dude so what it what is a car in the current market an auger what was the currently well Karen market

33. 14:02

    an auger for the Democrats like you know this doesn't fit at all right I think people have pointed that out too so interesting Lee I mean right now envy one of our people shouldn't create markets that far out because the b2 is gonna be coming out before that would actually resolve and if if the price of rap and v1 gets it's too low because the fact that v2 has come out and everyone's kind of migrating the platform to v2

34. 14:26

    that affects the security of v1 because as I mention before yeah yeah you can imagine if rap is essentially worthless they don't care if they have to burn it if there was a lot of oh I for example in in the Democrat which is actually possible so like as a very specific example if the if the open interest and that got up to let's say three million and the market cap of v1 rep was less than three million yeah then just by all

35. 14:50

    the rep and resolve the market how you want to write it's a hypothetical though so it's one of the things where we don't actually know exactly how it'll play out it's something that we just yeah where's you work out messaging right now this you know this kind of decided they don't do that yeah or this will have this may happen it's like it's like you can do that but just know like dad's how the game theories gonna work out there it's yeah it's everything we do we try and make completely game theoretical secure

36. 15:16

    but there's also the weird the humans are weird yeah actually it's also like I mean you do have underlying like security right right so yeah and a lot of that is based on you know multiple assumptions and if you read the white paper for example there's a whole issue of parasitic markets a familiar with that know that concept so parasitic markets are kind of not only the thing that augur sort of

37. 15:39

    can't solve it's actually probably provable that it's an unsolvable problem the idea of a parasitic market is imagine the draw down actually so imagine there is an election market for 2020 and people bet two million on Trump people bet four million and this is all on chain gang jerk edges Trump and yang I'm thinking dude he said uh he's come to the conference he's talking on was it I got a I got a Johnny hanging so so the

38. 16:19

    total Alliance market is six million dollars so say you and I off chain just I have a bag of cash say I'm I'll give you I'll bet you a billion dollars the yang wins problem is there's nothing the chain can do about that there's nothing our our security can do about that we can never know that so there's no way that any systems we create or any mechanism that we create could account for that basically the best we can do is and it's what the white paper does is it says we don't

39. 16:44

    think they'll ever be more than this much parasitic oh I based on the oh I within our system and that's kind of that's the best we can do so so reading the white paper and with that amount I believe the amount of we we limit that to is something like 50 percent of existing oh I so we're basically saying is if there was six minutes market our white paper is assuming there's no more

40. 17:08

    than three million you and me bets like agreeing to use our our Oracle to resolve as if you believe that there's more oh I than that than the system is it can't be completely totally secure on that there's a lot of things that go into that obviously it's purely kind of pedantic game theory sort of stuff but but dude I mean there's you know multiple prediction markets yeah there can be multiple prediction markets that

41. 17:31

    right so yeah and agreeing to use our Oracle for that winners other options too yeah like I said it kind of gets into a thing where it's purely pedantic game theory is I believe it's entirely possible there could be a lot more personally and things would not actually break again like that there's more reasoning on that what if they have other markets and they want those to resolve correctly yeah and you know presumably showing that the Oracle resultant badly would have other

42. 17:55

    negative side effects for them and potentially their other markets yeah that's true yeah who knows yeah interesting all right yeah okay so let's say I go I create a new yes/no let's say Tran versus yang so while I'm creating I like I don't need to take a position in any right I just kind of offers this to somebody else to take positions as a creator do I take any cotton on this

43. 18:25

    always it's great for the market yeah yeah so every market has sort of fees associated with all of our trading stuff too what did I hear see you like trees all right yeah I'm a programmer with everything all right so every market has a market creator fee this is something that's set by the market creator of yeah right now I think fees kind of are around one or two percent I think

44. 18:57

    typically the expectation is that over time they'll go to zero because people should be yeah you work it off and yeah but but v1 v2 and Marcus there's not a lot of competition not a lot of high quality market craters so people can charge higher fees and people seem to be okay with that don't mean for trading on things that have you but even as high as like four percent fees which i think is probably way too high but but that just goes to

45. 19:21

    the market painter so when two people the way that we do fee fee settlement is whenever a trade would make people sell both their shares so if I have gang chairs and you have Trump chairs and we both want to close in our positions and we get matched in a trade it'll sell that back into the system and extract these I see so we take these in that case and we take fees in the case when the market is over and you're just getting your payout from the system those are the two case if you take these if you're just trading

46. 19:46

    back and forth shares in tokens we actually don't they're the auger platform doesn't actually take any fees so you go entirely to market crater in v2 actually this is a v2 feature pre yeah the market crater can specify an affiliate fee so the way that the affiliate fee actually works is they specify a portion of their market crater fees they can go to affiliate I think people just like yeah yes okay they can give 25 percent of whatever they get from market crater fees to anyone who

47. 20:17

    presumably from yes right right and the way that's implemented because of smart contracts is you just specify an address yeah um so really savvy people are very quickly going to learn that they can just specify their own address and yeah but you know for a lot of people that is the case I mean arguably you can sign up for like the Amazon affiliate program thing but nobody does so in practice it'll be actually use that way yeah the

48. 20:43

    other fee is the reporting fee this is kind of essential to oh wait affiliate is defined not by market creator but but like when you're actually doing the when you're doing the trade so the percentage of their creator fee is specified by the market crater but they don't actually specify an affiliate there's no no whitelist for that or anything it's just something you you enter in when you make

49. 21:08

    a trade so it's like the front end will put it in intersection pretty much win right people give links and it'll happen if you like in our UI will know how to to interpret that I know presumably other you eyes will know how to interpret that as well like it would be smart for anyone building an auger to also build into that obviously yeah so the the reporting fee is actually more important it's kind of central to secure you the system but it is a function of

50. 21:33

    the price of rep and the öin the system and this is how we make secure that whole system where we try to guarantee assuming a certain amount of peracetic interest that if someone does incorrectly resolve a market they will overall burn money and we do that by adjusting reporting fee so if you imagine the open interest on the platform is see I'll try and write this out so if the

51. 22:01

    open interest the platform is 10 million that's across all uh right across all markets yeah okay and imagine that the rep market cap why it's called drats for reputation suppose we saw relation yeah in the first iterations of system the way that I worked was you you have a reputation it was more sort of reputation base like you would by reporting tokens for the right to report on markets maybe we get assigned to you and there were a lot of iterations where kind of the game theory

52. 22:40

    folks were progressively more and more cutting everything I think this is not necessary this is not necessary so it's sort of it's less reputation took it uh reputation based now I'm sure there's a little reputation what next a cannery yeah yeah I mean right now we actually assume basically for Game three purposes that represent by two people like it should still work it doesn't actually matter the reputation of the holder now you mention rep market capillary mean it anything that's transferable is by

53. 23:04

    definition kind of that's her petition exactly right so let's actually make that a hundred nine if ret market cap was 1 million then obviously something's wrong with the security model right now rep is not worth enough to secure the system and if there's actually there's a multiplier that we have I think for v2 we're changing it to the rep market cap should be sorry I do I should be five times I believe the rep market cap that might be adjusted but we have a formula

54. 23:33

    basically for for adjusting that and what happens is when the rep market cap is too low reporting fees go up reporting fees are taken from settlement and given to reporters for the people who hold rep so just by holding rep you can participate in the system and then get those fees right so so presumably the higher this is the more the rep market cap should go up and the more secure the system is against someone trying to buy up all the rep and then take all the money so this is kind of how we attempt to in an automated fashion our mechanism

55. 24:05

    for adjusting the security of the system based on the rep market cap and oh I so right now in v1 the reporting fee is the base the minimum which is 0-1 percent right so I order your feet so look just for my understand for my visual understanding so you have let's say market so as part of it there's some amount of eath locked it locked up and this issued some yes tokens and no tokens we can freely trade this it's like a 20 you can before this finished

56. 24:45

    you can also just like pretty much burn it and take out ease thank you like both both people yeah sure show yes now like you close the position you take out these right and then at the end you have like when the Oracle comes in right there's a way to pretty much close this up and then the the reporting fee is taken out of the fool oh I of this market yeah but based on the amount that they catch that right so if they stay together let's say for simplicity sake there's two people yeah yeah so it's like two million here four

57. 25:25

    million here right right so then in total I mean so that means the total of the market were to be six million oh well yeah so threat reporting fee was 1% then it would be 1% of 6 million how else like you don't have one-to-one matching right of yes and no talking so it is it is a lever like what is it called actually don't know host go look for to one kind of betting thing so the prices that they pay for their shares are not actually something the system ever takes into account so whatever they

58. 25:58

    paid for them it's it's something that MV one we don't actually record we we we have a middleware component that reads logs and tries to kind of show them what they paid for it and in v2 we actually do we record that on chain not for purposes of payouts but just purposes of showing them what their profit losses and everything but the way the system works is someone somehow acquires a yes share after you dig into

59. 26:22

    our trading contracts it's you know they did it by specifying a price they want to pay and somebody else was willing to match that either sell their own yes shares or take the other position is no but once they pay that money to the system think about it they've they've bought purchased a complete set basically of yes and no so they bought all the outcomes that exist in that market together then they must pay whatever a complete set is worth but it's irrelevant what each of them paid

60. 26:47

    it all the matters that the system got enough to pay out enough basically this amount so interestingly the amount of complete set costs is always the number of ticks by the way to which again one of those nice properties that works on the chain so together they came let's say they just bought one share of each so they collectively had to pay ten thousand but the person about the yes may have only paid one the person who bought no maybe one who paid $9.99 or

61. 27:09

    whatever but then when they cash out you know whoever one gets the 10,000 and we know that we have that because that's what they had to pay to get there is that too now the the the fees are taken out of it as well obviously so the way the fees are taken out is also just split evenly not split evenly sorry personally yeah proportionally based on whatever they what if they are getting out as well well I'm just like I mean there's a underlying price at the end of

62. 27:32

    yes and no like if somebody wants to go and trade it on like secondary market the price is defined by this time well so the price is defined by whatever people are trading so there's actually an order book on on chain and people going in say what they're willing to trade for so just like any other exchange it's it's not actual something we're like it's not like a actual betting market where somebody defines the odds or anything

63. 27:56

    like that it's just a open market yeah yeah although Fordyce worth they even with with an underlying mechanism like that there could be you eyes built on top of it that kind of translate that and the betting odds like that yeah it's kind of a display thing yeah right okay yeah I see so pretty much yeah take out the yes if I want to take no position I take out the yes share I go to some

64. 28:20

    exchange I sell it somebody buys it right and then that defines the odds right like whatever position I'd like sell it here that's kind of odds and frankly to take yeah and then if for example it doesn't close I can go close this up and say hey they work great okay cool so by increasing reporting fees this is pretty much more reward to this to wrap holders right so presumably this increases market cap that's the idea yeah and in theory in practice like I

65. 28:59

    mean I we have a behind present because right now the actual rep market cap is I think like gosh I remember but let's I think it's something like 200 million in the the oh I in our system is something like 2 million so we're wildly over secure okay great there is zero risk of that happening well that's good yep hey okay so yeah so we did this election 2020 and then things are happening how does this pretty much Oracle feed sirs into the market right so what happens is

66. 29:34

    when the market expires let's see what's the right way to draw this so a but yeah we figured out okay so the market ends let's say you know in June 19th so what occurs is at that time a designated report window starts do a box what's the dr that's what we shorted to the designated reporter is the person that the market creator designates understand aetherium address i see this is the this address is allowed to go in

67. 30:12

    and say say what happened say whether Trump were you anyone I see so the D are currently in v1 has three days to do that in v2 they're gonna have one day because we found that if they do show up at all they show up within one day if they they don't show up within one day they don't show up at all so so there's a few actually interesting if they like if they lost their key to I don't think

68. 30:37

    in v1 there's a way to transfer that so they would the shadow look basically if that ever occurred so there's two things that could actually happen here let's say they show mm-hmm or they don't know show some decision tree happening so in this case the dr shows and they say yang and then in the case where there's no show so the three a limited past or in be to the one day limit it goes into something called open reporting so are an open

69. 31:18

    reporting the no Shobana which is something that the market crater has to put up at market creation time which is rap gets used for the open report mmm it's just like a extra reward yeah for somebody else do you exactly and the open reporter there is Li entitled to it once the market finalizes presuming they they report it correctly so the initial bond goes up and up and up until it reaches kind of equilibrium that that's the idea of it and that's based on that

70. 31:48

    something the system records the system records how much people don't show up and the more people don't show up relative to some target lives or people the more new markets need to put yeah yeah so so you can imagine if five percent of markets have no-shows and we're targeting one percent the bond will just keep on increasing we're presumably once it gets high enough people will show up and then I can start decreasing again but shouldn't be like one per market basis just like

71. 32:12

    increase the reporting fee like over time like let's say you know one day passed nobody reported increasing little bit so the reporting fee is are different that's right reporting fee or Awards all the rap Oracle's or yeah it's not market specifics for everyone but yeah and it's sort of a mechanism just to adjust the the market cap that the the no-show is very specifically just people aren't showing which is making

72. 32:36

    markets resolve slower have me adjust that a little bit so there are no show the o.r gets report let's say the o.r reports for trunk I don't know let's say in this in this world yang wins so so regardless of whether the O R or the D are reported at this point things are gonna be the same all that actually matters is who ends up entitled to the rep that was put in this report so dr gets the the bond is like

73. 33:07

    yeah this is really so very typically I think in pretty much all cases that ever occurred the dr is the market crater just like i'm yeah yeah um for just to be technically accurate what ends up happening is when the dr does show up the market here gets the no Chabon back and then the designate reporter has to put in a new bond but in practice that's just them putting the notion bond and

74. 33:29

    just like what happens with the the open border I see so at this point yang one so a few things can happen the market can resolve with this outcome for that to occur this has to go at least 7 days with nobody putting up a Rhett bond to challenge this and say the Trump one it's actually a little longer than seven days there's a cadence of Windows that go by called in v1 fee windows and in v2 they're called the spiit windows and there's run a seven-day cadence so it's possible that if this occurred at the

75. 34:02

    beginning it has to go a full one of those windows so if this was reported at the beginning of a previous window it could be almost 14 days for it to actually resolve and there's ways in v2 that that's going to give people opportunity to dispute it for it to give people who have rep and who are looking at markets time to actually find it if the the window is too short we risk I guess 7 days what do I do like why

76. 34:25

    didn't junk it because it makes it so there's a few reasons one it makes more predictable so if we chunk it in like a regular cadence like that there's no segment of time where someone could have missed out on a market people can can make schedules you're saying pretty much like every week I go exactly I go every Wednesday Wednesday is my my day to look at markets and pick out what I want to I want to do it's gonna be work a little

77. 34:48

    different of you to there's still a cadence like that but there's other mechanisms we have to speed up reporting because you know time is adding up here we have three days and up to 14 days and additional other time it all makes for really slow resolution time but let's say that goes seven days and nobody nobody reports on it or goes the period of time has to you to resolve then the market results okay so what the

78. 35:23

    market is final and yang one and then pretty much anybody can get just go and right so city boys it's there yes exactly so so now the market is finalized anyone can there is a actual contract of claim trading proceeds you can pass it in a market it'll look and see what shares you have see what to pay you for them destroy the shares and give you the money from the market and then pay feeds for you the other thing that could happen though

79. 35:53

    let's say somebody just beats this so we're in let's see this was the initial report so this is second round and they say no yang one when you dispute what you have to do is put up there's actually a formula for how much rep it's for purposes of this we can just say it's twice the previous bond it's actually different than that and it especially when it goes back and forth it ends up being a different amount than

80. 36:26

    that but just for simple to guess it's twice the amount of rep that they put up so judo are put up some well it says so the are acted the o.r or they are rather it's always gonna be the envy one it's a little complicated it's the initial report size but for simplification purposes and because in v2 this is how it work it's just the notion bond so if the no Shabana was one rep then the

81. 36:51

    amount staked on Trump right now is one so to dispute that you'd have to put up to rep nicely then this can continue and continue and continue if it doesn't go this period nobody disputing it until so so if you booktube and nobody disputes for seven days this is for like finalized right this guy loses years right so will actually open is if that this is the speed that one let's say it actually goes in seven days yeah and

82. 37:18

    what happens is let's say this is the trump rep mm-hmm this gets sent it doesn't actually get sent directly to this person but in effect it basically goes to whoever put up the rep for this for this dispute so I'll call these the yang holders so what the contracts actually do is when the market is finalized it looks at all of the losing bonds all the ones that weren't for they were arguing basically for the outcome that did not end up being the finalized

83. 37:56

    outcome and just takes that rep and then puts it into the bonds for the for the winning outcome so when someone comes in and redeems their bond with the system for a sample this yang bond they'll just receive that rep so it liquidates all of the losing rep and gives it to all of the winning bond actually and that actually the way that we have the formula set up that I was mentioning which it's actually more complicated in again and then just two times the previous bond but it

84. 38:21

    guarantees a 50% ROI so if if you think you if you think you're correct right if you think you're correct it's actually it's extremely profitable to all right if you think that's the truth really not not you guys are like yeah the reporting is actually it's it's weirdly nuanced actually just the concept of what's true yeah there's some interesting markets we should talk about actually but yeah if you ended up disputing before you wrap

85. 38:51

    your name and label it right you end up getting 60 um yeah it admits very very interesting markets one of the one of the first markets and actually I think the the biggest dispute that ever occurred I think it got up to either some around like 80 thousand dollars with a rep before actually a little finalizing was a will the weather be nice for the Bastille Day Parade tomorrow and I was that was a description and

86. 39:17

    there was a link to a weather site that did not use the language good or nice yeah it it was odd actually seeing a good disputed so much the the marking crater was very very adamant in our discord that it was a valid market and that people should be able to find a shelling point on what is nice or good but it end up resolving to in it invalid which is arguably what it should have resolved to invalid is interesting to know saying that is like a separate

87. 39:42

    thing right is yes so in bowel disease is a special case it's handled very very differently in v2 than it is in v1 so in in v1 the way that invalid works so for an invalid payout it wouldn't look something like this in v1 so if you mention is yes no market will just be an even payout mm-hmm so this is nice in some ways and easy to understand in some ways but it has a lot of really negative weird side effects and not every market that doesn't even pay out obviously is

88. 40:30

    invalid there's also like a flag you have to provide that says I intend for this to be invalid you can imagine a scalar market there's actually a fairly decent chance it will end up just being paying everything yeah yeah yeah so in this case you know you can imagine it's not it's not actually the same as like paying people back for their bets which i think is a lot of you'll expect for invalid instead it just says okay we have to pay all outcomes equally because

89. 40:54

    we don't know how to how to resolve this so as there's only happens at the end at the resolution time yes like you can't adjust flags with this valid from the beginning and be one no right which is actually my problems and it you can't even really glean it easily from what had the order books trading because this is sort of weirdly baked into the price of things like if I pay 0.44 yes that a signal to the market is invalid maybe

90. 41:17

    because if it is invalid then I'll actually make money on that I as I'm paying less than the amount that I'm going to get from payout but it's really difficult to factor that in based on this payout so so pretty much people betting on yes/no and invalid in reality exactly but the bets on invalid in v1 are very difficult this yeah it's like it's it says this like bicep is the price you're actually betting that it'll

91. 41:43

    be valid and it's it's it's tricky and it's especially tricky because invalid markets are sort of a conundrum in v1 these people enterprising individuals again everything is decentralized so people are gonna find everything eventually and figure it out I figured out that they can create markets with really tiny subtleties that mean that they should resolve to invalid surfers who might make a market like

92. 42:05

    what will the price of a theory and be at the end of April 15th and they'll make the market and the beginning of April 15th the very start of the day reporters what they have to do is to be consistent they have to say what was it at the time market ended if they can't know then they have to resolve it to invalid um you know which is unfortunate because people see these and it's so subtle that they just see April 15 to

93. 42:30

    April 15 yeah oh yeah yeah yeah well so the thing is even though there's three days they they the kind of ethos is you have to know right then because the DR could have reported right then and do we you can't go back in time sort of iced know if it's like a market price yeah you won the market price it's that time not like market price in the day right and if you if you imagine reporters kind

94. 42:54

    of as a whole whole body if they don't want to get in situations where they're hurting themselves um just because of when people report it or something like they're very intolerant for example of api's that go down or provide different results to different viewers there's a problem with ESPN because we've other people who report in Europe and their Google report in the United States and they see different ESPN's and yeah there's they do not display in a way

95. 43:19

    that is easy people actually discover this this is one of the weird subtle invalid markets was they referenced ESPN um but this is kind of an issue for v1 because like you said there's no there's no curation obviously the thing is completely decentralized we actually run ball all we just make the software the best thing we can do is add things to to a reference UI basically for viewing the on chain data that can help people there's a few of those that can be added

96. 43:43

    in v1 like looking at the spread in a particular market like is it oddly gearing towards this for example we can look at that and if so try and make some estimates sort of on whether it's invalid we can look at things like market fees we can filter out anything with the market for you like we could have a default filter basically in our UI of anything that has fees above if I 5% there were some markets with a 50% fee for a while which is not gonna be

97. 44:08

    possible in v2 but obviously that was like a way people tried to manipulate the system for the invalid pay out though in v2 we're not actually gonna do invalid pass this way instead invalid it will be a tradable outcome on every market regardless of whether you specify that or not so you'll always be able to trade invalid the the neat thing about that is you can you can obviously trade

98. 44:32

    invalid with that I'm going to do weird calculations but also it gives us sort of a it gives everyone an insight into whether the market is invalid or not yeah and it's like you can hedge kind of as you taking a position yeah it's a the cool part you so sue so you could use the invalid outcome to hedge you could also like one of the things we've discussed is building filters into the reference UI that look at what in Val is

99. 44:58

    trading at like if there are Ani any bids for invalid say above 0.1 or something you can imagine that we could have to filter have a default or hey it was like like this there's warning this is probably invalid if people are trying to buy the invalid outcome for for that much consistently obviously there's a lot of questions about how the filter like that would work because it's gamble obviously if

100. 45:21

     someone wanted to just be a troll to a market yeah but the issue is making sure that they like everything else that they have to burn enough money in time but it wouldn't be you know actually economically advantageous to them to troll or to try and mask something as valid well yeah you can say what is a like position of invalid pretty much compared to for actual decision right than there if it's bigger or if it's like if it's oh yeah maybe meaningfully yeah if it's

101. 45:46

     trading at point nine like that and there's gonna be some formula obviously there yeah because like we don't want like a single dust order four point nine to hide a market but you know if there's significant volume whatever that is maybe some portion of the overall volume on the market trading for prices above a certain level then yeah you can say with some confidence actually the market is invalid that's something everyone's looking forward to a lot in v2 because there are

102. 46:08

     a lot of very silly markets yeah env2 actually another really interesting thing about reporting is the reporters you know they they're their goal was to essentially to report truth and to distinguish these invalid markets and handle things like that but really at the end of the day their goal is to serve traders really to make traders confident that markets will resolve how the traders expect there's actually a more subtle weird notion we weren't really sure how was going to play out in practice but there was market that

103. 46:43

     actually kind of exercised this notion which was made by someone who famously makes a lot of invalid markets to I think this was actually an accident but it was on the the 2008 2018 house election mhm it was whether Republicans would maintain a majority and the market ended I believe December December 18th or something somewhere in the middle of December the interesting was even though the the elections happened before that they don't actually take a majority until the next year and the way that the

104. 47:15

     work market was worded it was will they take a majority adds a point of time right and it ends at December 19th so technically Republicans still had a majority even though they they actually lost seats the Democrats and the Democrats took the majority from the elections yeah yeah but not really right right and traders had actually put I think 1.5 million Noi into this market so it was getting heavy traction and

105. 47:40

     there was some debate about how it would resolve actually because it was sort of the spirit of the market obviously what traders had intended it for to be it should resolve that Democrats won yeah but if it was sort of too to the letter of the market in the description and the end time then it should be either Republicans it was actually even some legal argument should be invalid because they're saying because this was known before before the

106. 48:03

     market got made this shouldn't actually qualify as a valid prediction market and should be made invalid yeah you cannot pretty much right so there was there was vying camps actually within sort of the decentralize Rep reporter world on how this resolved I mean there's also like you know all kinds of things that can happen before yeah it actually takes a seeds right that's right and change that yeah there's I mean there's plenty of other valid reasons things could be

107. 48:26

     invalid yeah but it instantly it did it resolved sort of where people had hoped it would for traders with Democrats I'm so resolved in the spirit of the election who won the election um which was interesting to see and nice that the the reporters could kind of find a way to amongst themselves figure out what would be best for traders and for the future of the platform which is what they kind of have to do yeah yeah because they want more people to write

108. 48:51

     and this is also what makes this the Oracle problem so interesting because it can't it's not something that can really easily be solved with like presentable just simple programming it may be a machine learning and extremely fancy AI or something but a lot of these questions are sort of weirdly human-like what what did humans intend by this market yeah I was one of the more interesting yeah hi so can you explain

109. 49:18

     more what what is this like the idea of to split universes right decisions yeah right so so let's say that's where the home was getting uh if the dispute just continues and continues and continues it'll eventually be some round let's say it's the twentieth round and this varies depending on you know what the new shabanam was originally on how people dispute but once it reaches a certain

110. 49:41

     threshold it will cause a fork so when a fork occurs for this forking market anyone who disputed who has any any rep locked up their rep is forced to go to those universes so they they may call a function because redeem my bonds and put the the correct universe so all the so when you say is a senior's what this means is the rap smart contract which is it was just an auger yeah yeah it's a rap contract right you just fork it so it's actually so

111. 50:21

     it'll be a universe contract and that universe contract has its own version of rap okay so yeah so pretty much originally there's one universe and then you like fork it into let's say because of this need for it into a new one so it's just a clone event where rap holders pretty much here it's rep holders off like yang the pretty much got everything cuz take on Trump and here's a Trump holders got everything from Yang sort of so so what actually

112. 51:02

     ends up happening is I mean this other there will be a trump and yang universe in the in the yang universe when they when all the people who staked on Yang move their rep to that universe it will mint right for them in that universe as if though they had won the dispute yeah and similarly in the front peanut verse it'll mint rep for them as though they had won this view in those respective universes but it does actually take it from the other universes it just sort of

113. 51:26

     says you guys both want it on you know I mean the question is is there original universe or it kind of like is original universe that just got destroyed so the original universe is essentially frozen forever nicely so the the original rep for example still exists and anybody can move their rep into one of those universes and they just cannot like trade trade anymore organs or so so trick trading continues forever but nobody can create any new markets and no markets can find lies so

114. 51:55

     any existing markets that haven't finalized also to mike resides a also need to decide so the markets actually don't decide the markets just go to the winning universe so the markets are forced to go to those winning universes which is again why this security calculation considers the entire open interest of the whole system instead of adjust the open interest of like the particular market that is working because if someone wins a fork all UI and the system moves into that universe

115. 52:19

     they've even though they don't win that market they've effectively want all those which which in a way makes sense as well because if they have the if that force whatever it is has the ability to win that fork then they have the ability to win it in future forms as well yeah exactly so how is the winning decided so so the winning is decided when the fork occurs there's a 60-day window mm-hmm

116. 52:49

     there's a 60-day fork window uh and by the end of that window whichever universe has had the most rep migrated to it is considered the winner or or if before that fork window ends over 50 percent of rep has moved to one of the universe's meshes how fork fork winner is determined so but the people who are like forced into za Zola and they pretty much just burned like in reality like because this has kind of got frozen or like you cannot create new markets when it rains Oh so in the trial universes you can create

117. 53:29

     markets it's just the genesis universe the universe that had the 40 market originally that's frozen forever so so in the Trump and yang child universes those just those keep those go in their own so in theory like this people can't like you know keep going crazy market trade if if they're and that's an important part of why the system works the way it has to because if if there is legitimately a group of people who

118. 53:51

     believe that there's another truth to the universe they can go on they can it is actually perfectly viable it's hard to think of the exact same happen so it'll be like around - pretty much on the market yeah right like it's own thing and there would be traders that it's really ultimately a decision of traders whether they believe that or not like there might be traders who wanted a market it'll probably something more subtle than like whether Trump or Ganga

119. 54:16

     yeah something like there's no actual true sits right or if it was like a sports game with a very very odd call to it or something and they you know they're people sports traders are like no we're trading in this universe where it was evaluated this way instead because that's the way we like to bet on things so I could imagine some sort of a scenario like that where maybe the people who bet on a specific subsection of things and they

120. 54:38

     want the self selection pretty much inches of space where all the validators believe into Zeus right like it in the bastille day market actually there was a just you know one in the best deal day will it whether it be nice for the best deal they parade you can imagine maybe they're people who like to bet on weird hard to define things maybe they're people who want such a market for example to resolve is yes when it was sunny outside and they could have their

121. 55:04

     own universe and then people who don't want to bet on strange in difficult to word markets would go no I'm go to avoid that universe but that's perfectly valid if there were two universes just what people want to bet have not happened I mean have you had fork at all no no I I sort of suspect there's a chance that a fork will never occur but the important thing is that the threat of fork always exists some way oh I mean it's like a

122. 55:28

     slashing yes yeah yeah you'll get slashed don't do stuff right exact people don't do stuff it don't get slashed right it's it would be extremely disruptive to the system obviously because there's you know a 60-day window where it's quite likely no new markets will be able to be created people are kind of potentially panicked wondering which where to put their rep I suspect well I mean it's like even when the four created you can you can start creating markets here right right

123. 55:53

     right so like if you believe this should resolve that you'll go there and strike creating markets but if it was something you shouldn't settle for example it might be less obvious which which more than the other rep - right it's there's not really a good hedging opportunity it's basically well you can't market in those that's the thing if you create a market in both though I mean you have to split rep into both you have to split your open interest into both you have to bless the quiddity because you're splitting so you could if you had to

124. 56:19

     create immediately but it costs money and and presumably you are burning half of your money unless both of those universes end up being legitimate yeah yeah so it is disruptive and it would be good to avoid it but it's important that it exists in functions for the game theory to actually hold interesting now all this time lines are pretty much were deciding mostly around like how

125. 56:42

     long like for people to check I'm kind of made decision yeah yeah the 60-day window is essentially a thought process of how long would it take someone who has who is a rep whale to get their rep out of cold storage so if you assume you know it's gonna take them a month to travel around the world and final dig up a place where their Hardware wallet is and then give them another extra month

126. 57:05

     to actually use it and their secure underground bunker yes so transactions from 60 places in the world random coffee shops yeah but it is I mean it is kind of just a arbitrary estimate but sort of based on that that mindset yeah I see yeah it's pretty cool so what happens was scalar decisions how does this work for scalar decision so it's sort of in the same vein when people report on a scalar for example let's say it's the price of Apple what dispute for

127. 57:36

     that could get particularly messy if there's not some very specific instructions on how to actually resolve it yeah the the argued method of doing that is point to a very specific singular source like like a API feed or something that will give you a historical time period so that there can be no debate about what the actual number is if you did for example like a general resolution what's the price of Apple it's very likely

128. 58:01

     there'd be a dispute because it's all mental sensors but like let's say there's you know price which is like 118 yeah 83 and I believe it's one 1885 and then somebody believe it's one 1981 so like I pretty much go in and say it's one 1885 and then you want to go in and say can you like bet on something out like yeah yeah you could say one 1884 and then somebody else could say 117 I

129. 58:26

     mean so they do like fork off and all those universes and so if there was a fork yes after you redeem your your money they would fork in those universes there's and that would be bad obviously I guess alien like that's extremely unclear there's I mean there's some just sort of uh rational processes that would probably go into that to prevent something from happening so for example if you're really disputing that obviously there's probably a shelling point somewhere in here where people can kind of agree like

130. 58:56

     obviously if both of you had money on this or even if you're just reporters it's doubtful people would argue about amounts that small and if you actually had a why to protect in that market it's doubtful you would make more by disputing that small amount then you would just going okay that's close enough own make enough money it I just feel like that would have to be driven by something kind of irrational and it's difficult to prevent people from doing irrational things because they're

131. 59:21

     irrational so some like my guess is in such a situation there'd be some sort of selling point like okay we'll just agree it's 0.8 e 4 it would end up resolving to that but theoretically if there was such a dispute yeah any any possible amount that could have been bet on in that market would be a potential fork universe as you yeah so in theory can actually have like 20 Forks oh yeah I mean it could have thousands thousands

132. 59:48

     of Forks the people are willing to pay them because you only have 20 steps so it's not actually just based on what people report on though too so any possible outcome that could have one you can make a child universe even if nobody actually got somebody can just go and like yes when it forks anybody can like hey here's a another fork it has to work that way too because otherwise there's sort of a attack in the forking procedure where you could imagine

133. 1:00:11

     somebody let's say this was a categorical for example and interestingly Biden and Murray are also running they're actually like five own candidates they decide to be independent and let's say yang wins somebody want to attack the system could just go by and Bernie by demurred anybody burning and not let anybody else do any disputes and if that was it was actually yang that one there could be no true child

134. 1:00:35

     universe so we have to assume that attack could have occurred and let people make in a universe that could have possibly been disputed or could have been the truth there has to be there has to be a true universe basically so to do that we have to a lot of people to make any universe that could have been alright that's magic to display yeah so the display is a whole whole er notion there's there's how those actually played in the UI it's because it's something that we have to show traders and reporters to basically

135. 1:01:08

     there has to be some sort of notion of choosing the truth that you live in - right yeah try spring into it and it gets even more complicated if you imagine a world where there's been multiple Forks so say you know the the election market Forks and then in the yang universe they have an argument over how many Twitter followers yang has or something of that sort then is a sub fork there so I live you know if you're using augur you would be using I believe in the truth where yang wins and had a

136. 1:01:33

     million Twitter followers you know but it gets it gets more more cop it could get more or more complicated obviously if there's multiple Forks yeah I get you're getting some of those things that exists and it in theory works but if you for the best if it never actually okay all right cool yeah no it's super interesting anything else we doing the cover go on all day if you go there stop stalking better yeah unless pops wasn't my head at the moment hey yeah I can close it up

137. 1:02:07

     oh you're great cool well thanks a lot have a super interesting check it out on I guess white paper and the website itself I mean so life yeah Garnett hang our discord if people ever interested or have questions we're very much always there yeah all right cool thanks