How to Create a Thriving Prediction App System On-Chain, with Dan Kaizer @ Azuro

[00:00:00] Hello everybody and welcome to the Crypto Hipster podcast. This is your host

[00:00:05] Jamil Hasan where I bring you the I am the crypto hipster and I bring you founders entrepreneurs

[00:00:11] executives thought leaders

[00:00:13] artists amazing people in the world of crypto and blockchain all around the world and today I have another amazing guest

[00:00:21] I'm looking forward to this interview and I assure you it's gonna be a really good one

[00:00:26] I have the chief technology officer of a Zorro

[00:00:31] Dan Kaiser has joined me today then welcome

[00:00:34] Hi. Hi, Jamil. Hi guys. I'm very glad to be here. Thank you. Jamil. You're very welcome. I'm happy you're here

[00:00:43] So yes, we're gonna kick we're gonna we're gonna talk about some exciting topics today

[00:00:46] And I'm gonna kick things off by asking first about your background

[00:00:50] You know, what is your background and is it a logical background for what you're doing now?

[00:00:56] Mmm. Yeah, so absolutely to be honest Azura is

[00:01:00] Is really a project that combines all my lifelong interest and experience

[00:01:07] All together really perfectly. So because before getting in crypto

[00:01:13] when I was student, I was deeply interested in the betting and

[00:01:18] Probability theory etc. And during my university years, I even developed some

[00:01:23] Some strategies to make sports betting profitable for me. So actually it was

[00:01:31] So I can say it was the first money that they are

[00:01:34] First real money that I earned by my own it was

[00:01:39] through professional betting I

[00:01:41] Had some special systems that I developed myself is my friend

[00:01:45] So actually this friend working in Azura as well, so he is responsible for risk management in our team

[00:01:51] So

[00:01:52] however bookies often blocked my accounts

[00:01:56] So I had to because if you are profitable books web to book is usually a block you so I had to use

[00:02:03] accounts under my friends

[00:02:06] Relatives names my grandmas names, you know, etc. So and they create the numerous of accounts

[00:02:13] It was quite annoying to be honest

[00:02:16] But then then I discovered cryptocurrency

[00:02:19] First of all, it was BTC obviously it was

[00:02:23] 2016 I believe

[00:02:25] then I

[00:02:27] Attended that some hackathons at some conferences about his area

[00:02:34] So and I got completely

[00:02:36] excited about it so I forgot for a while about betting about everything else so I was I

[00:02:43] Was working on the encryptor?

[00:02:45] since this day so and

[00:02:48] once

[00:02:49] It was

[00:02:51] 2021st year

[00:02:53] When my future asura partners came came up to came to me with this idea of Azura of creating some

[00:03:01] decentralized betting platform

[00:03:03] for predictions

[00:03:06] So the compass so I just understood that it's really a perfect match because I love

[00:03:12] I

[00:03:14] Love probability theory I love betting and I love crypt. It's like all together

[00:03:18] so because of that without any doubts I I agreed to

[00:03:23] to continue this project so and

[00:03:27] It was like clear for me is that we had the right mix of skills and vision

[00:03:32] To create something that can really

[00:03:35] be a

[00:03:37] Game changer to balance. Yeah, what country you went again?

[00:03:43] Hmm, what country are you in?

[00:03:46] It was in it was in the Russian so now I'm in Portugal

[00:03:50] Portugal okay, that's right. You're in Portugal

[00:03:52] my wife and I went to our honeymoon in Greece and

[00:03:57] In Santorini, there's this bakery shop the guy makes leuka modest and he makes chocolate and flu use

[00:04:05] Lucca modest which are donut balls, but what it is the bakery is a front for

[00:04:10] Soccer betting practice the guy has very profitable

[00:04:16] So it was just made an interesting what you said to me the web two bookies block you if you win

[00:04:23] Yeah, if you do it like on a regular base

[00:04:27] So bookies can easily understand if you just a better who try to guess

[00:04:33] Just because you I don't know because you really know a lot about football for example

[00:04:38] And sometimes you can guess the results. Sometimes you can guess outcome

[00:04:44] Or if you just make bets like a robot like a bot

[00:04:49] It's usually pretty clear because if you have some special system

[00:04:54] it's completely obvious for for bookies so and

[00:04:59] They can block you

[00:05:03] So I want to find a little bit more about a zero and then I want to get into the into the probability

[00:05:08] Theory in the prediction markets, but what is Azura all about and you guys want some awards?

[00:05:14] So what makes you an award-winning?

[00:05:16] in the

[00:05:17] winner in the space

[00:05:20] Yeah, good question. So I think the main secret of Azura success

[00:05:25] so I

[00:05:27] Think it's it's our team first of all because

[00:05:30] as I mentioned previously, it's a perfect match of different skills and expertises and

[00:05:38] In our team it consists not only of crypto experts live

[00:05:42] I mean web three experts with three builders, etc. But also

[00:05:46] some professionals in betting and also just some

[00:05:52] Interpreters with a very impressive backgrounds. So we have pretty grown-up team. It's not

[00:05:58] Young as a lot of other web three startups

[00:06:03] And they see I think yeah, I think this combination allow us to find the perfect formula for our project

[00:06:10] When we have designed Azura as I mean Azura

[00:06:15] Structures or architecture

[00:06:17] We designed it to be fully compatible first of all this web two analogs

[00:06:25] And even more advanced at once because we are using with three advantages

[00:06:31] So if we will dive into

[00:06:33] the details

[00:06:34] There are two main features that set us apart a part of all others. I mean first of all

[00:06:40] It's our unique peer-to-pool approach

[00:06:43] What does it mean previously most of other

[00:06:49] Decentralized prediction markets was based on

[00:06:54] On approach where every where each

[00:07:01] Separate

[00:07:03] Event let's say some football match or something like this some football game

[00:07:09] It's a separate pool

[00:07:11] Separate liquidity pool where liquidity providers provide liquidity

[00:07:16] In our case, we have general liquidity pool for all events all together

[00:07:22] So and it's very important

[00:07:26] This become possible because of our self-made solution

[00:07:32] Liquidity tree so I think we can talk about

[00:07:36] Liquidity tree a bit later in more details

[00:07:39] But

[00:07:42] But

[00:07:45] Long long story short

[00:07:47] this model liquidity tree allows us

[00:07:50] for a more efficient and scalable liquidity pool, so because it's

[00:07:55] I will explain later. But yeah, it

[00:07:59] Allowed to make some calculation

[00:08:02] Incredibly efficient like for example you can make one billion you can make a calculation under

[00:08:11] Under a range of 1 billion elements

[00:08:14] Using only 20 directions something like this. It's a logarithm complexity

[00:08:20] So yeah, first of all, it's our peer-to-pool approach as I mentioned and the second part that

[00:08:28] Is really unique now now team. I think it's our risk management approach

[00:08:34] It's mentioned approach I mean

[00:08:37] The way how we work with odds with probabilities, etc. So yeah, I think

[00:08:45] This

[00:08:46] Something that really set us apart of all other

[00:08:50] Sounds good. So when again, I want to get into a little bit more into the risk management approach

[00:08:57] First actually what I want to talk about this. I want to find out

[00:09:00] you know how prediction markets work and

[00:09:02] How can people partner with you to create create a thriving prediction app system on chain

[00:09:11] Which includes the risk management discussion, right?

[00:09:15] Yeah, so

[00:09:18] Let's start with risk management

[00:09:22] first of all

[00:09:25] It's managed the main part of risk management in predictions

[00:09:29] Obviously it's an odds odds odds is

[00:09:34] Is some kind of indicators of probability

[00:09:38] Lower odds mean a higher probability of the outcome so higher probability that it will happen

[00:09:45] While higher odds mean lower probability. So pretty easy

[00:09:51] For example, if we are speaking about let's say

[00:09:54] The most simple game ever it's head or tail right with this coin

[00:10:00] in this case

[00:10:02] Normal probability is 50 50 50 right 50 percent that it will be head and 50 percent that it will be tail

[00:10:10] And if we are speaking about odds

[00:10:13] Clear odds clear odds. It means odds without margin

[00:10:17] Clear odds will be 2.0

[00:10:20] Clear odds will be 2.0 on the one side and 2.0 on another side so but

[00:10:27] obviously

[00:10:29] If we are creating prediction markets, we need to have some margin inside the odds and

[00:10:35] Our odds will be let's say

[00:10:38] 1.95 1.95

[00:10:41] And it's very easy to understand that

[00:10:44] maths on our side and if we will

[00:10:47] Play this game for

[00:10:49] For a long period

[00:10:54] So house will be in the profit for sure because of this

[00:10:59] Lack of probabilities lack of these odds

[00:11:02] Great. So

[00:11:11] That's the risk piece so predict

[00:11:13] Prediction markets, you know

[00:11:16] in crypto

[00:11:18] Have some really good benefits have some really wonderful and amazing

[00:11:22] Opportunities after right?

[00:11:24] So what are some of the amazing and we're gonna get is dark side too, but we'll get into the positive side first

[00:11:31] What are the what are the benefits of prediction markets in crypto?

[00:11:36] So I think benefits is pretty obvious for me because

[00:11:42] One of the main issues in web2 predictions in web2 betting

[00:11:48] always worth

[00:11:50] The

[00:11:52] lack of trust between users and

[00:11:56] bookies house

[00:11:58] Whatever we can call it

[00:12:01] So and blockchain can solve this issue perfectly because it was designed to

[00:12:08] Solve the issue of trust last right because it's a trustless system

[00:12:14] Yeah, so because of that I think it's

[00:12:17] Uh, it's the main benefit that we have

[00:12:21] but not the only one

[00:12:24] because

[00:12:25] Also, I think it's payments

[00:12:28] payments is very

[00:12:33] Is very easy

[00:12:34] Easily done in in crypto. So I just need my

[00:12:39] Web3 wallet and I can start making bets without any deposits without any

[00:12:46] I don't know

[00:12:48] Withdrawal timeouts, etc. So I don't need to

[00:12:53] To send my money for someone else. I'm using just my wallet and I can make bets directly

[00:13:00] From my wallet from my wallet

[00:13:02] I think it's it's amazing as well

[00:13:05] Yeah, but obviously there are some

[00:13:09] Some limits in web3

[00:13:11] for now

[00:13:13] There are some problems with onboarding because not for everyone. It's

[00:13:17] It's comfortable to use wallets

[00:13:20] Someone it looks like too hard too complicated

[00:13:25] But it's like an ongoing process step by step we

[00:13:33] We educate our users how to use wallets how to use web3

[00:13:39] Applications, etc. Got it. So the drawback will be that it seems to be hard to use but actually in actuality it's not

[00:13:47] Yeah

[00:13:49] Okay

[00:13:51] Great. So

[00:13:53] I want to move to

[00:13:55] On to the oracles, right?

[00:13:59] So what is the importance of oracles today because there's been some changes and some evolution or de-evolution of oracles over the years

[00:14:07] What's the status of articles?

[00:14:09] Oracles in their maturity level and we're looking at the lens of transitioning companies from web2 to web3

[00:14:15] Oh, yeah

[00:14:17] Oracle is really a complicated area in web3

[00:14:23] One of my first projects in crypto was connected with oracle it was in 2017

[00:14:33] So

[00:14:33] It's still it's still not fully solved issue in in web3 in my opinion because we still

[00:14:40] Don't have some proper

[00:14:42] Decentralized solution for oracles and in meantime, it's really

[00:14:48] It's really a key

[00:14:50] Part of of this area in general because as I mentioned for example for us data is really

[00:14:58] The most important part of our protocol because if you don't have

[00:15:04] Proper data. It means we don't have risk management. We don't have these proper odds

[00:15:10] That I talked previously about

[00:15:14] So, yeah because of that it's really like

[00:15:20] The most significant part of our product and most and a lot of other products in web3 as well

[00:15:26] So for now we are using optimistic approach for oracles

[00:15:32] For now, it's one of the most popular approach in crypto

[00:15:36] One of the most popular maybe not the most popular obvious, of course

[00:15:44] So, yeah, I can elaborate a bit about

[00:15:48] About it how optimistic approach works optimistic approach means that initially

[00:15:54] Optimistically we we are thinking that oracle is fair

[00:16:00] And we trust to data that he provide

[00:16:04] But

[00:16:06] Then we can challenge him. So anyone can challenge him or some validators can challenge him can be

[00:16:14] System can be designed in different ways

[00:16:16] But the main approach is that after data provided someone can challenge this data and

[00:16:26] Send some let's say poison transaction or something like this and open the dispute

[00:16:32] Disput that this data is wrong, for example

[00:16:38] So now we are using the same approach

[00:16:42] We still have really a lot of work to do to make it pretty to make it

[00:16:48] Really decentralized because now we don't have enough entities in my opinion and not enough

[00:16:56] Don't have enough different data providers

[00:17:02] But yeah, we are working in this in this direction right now

[00:17:09] I would think that the purpose of oracles and the dark side of prediction markets are at odds with one or opposite things

[00:17:22] Right. And so the oracles would flush out some of the some of the some of the pitfalls of the prediction, you know

[00:17:30] Protocols how do you see it? How do you see the need for oracles being able to overcome the limitations of prediction markets?

[00:17:40] Yeah

[00:17:44] So, yeah, I think now oracles is really the main limit for

[00:17:50] For prediction markets and to balance for any protocols that using off chain data

[00:17:56] Now it's an issue for blockchain in general. So why does it happen? Because

[00:18:04] Look blockchain designed in the way that

[00:18:09] All data in blockchain fully

[00:18:12] Fully determined at any point of time. So I mean at any point of time you

[00:18:21] Completely understand what the state of any variable inside the blockchain

[00:18:28] But if we are talking about the real world it completely undetermined

[00:18:34] And here we

[00:18:38] We try to combine this data all together and if we rely on Oracle in inside the blockchain

[00:18:45] It means that that fully determined system rely on some undetermined variables

[00:18:52] Some undetermined entities. I don't know how to say

[00:18:59] And it means that blockchain become undetermined as well so and that's an issue that's why

[00:19:07] That's why just there is no perfect Oracle solution at all. It's just impossible to create

[00:19:13] So if we take some data out of the blockchain if we take some

[00:19:20] Some data from real world, for example, that we want to put it into on chain. It means it will be

[00:19:29] It will be always some

[00:19:32] Maybe it will be small lack of security. But anyway, it will be some security issue

[00:19:37] It's just I don't know. It's just a low you cannot do in another way

[00:19:45] So go to this compromises sometimes

[00:19:52] But any Oracle it's a compromise anyway

[00:19:56] One of the compromise areas I think that we've we've experienced in the past many years and have not improved yet

[00:20:02] Is the user experience, right?

[00:20:06] We need to improve it right because it's been an area of weakness and opportunity in the industry

[00:20:12] Are you working on some ideas to improve it? How do we improve it?

[00:20:17] Oh good question really so and we have our

[00:20:23] Maybe not unique but

[00:20:26] I

[00:20:27] I don't see a lot of projects with the same approach. So we don't create a UX by ourselves at all

[00:20:35] So we create we are building only a protocol and

[00:20:42] Our approach for user acquisitions that we provide

[00:20:47] Anyone anyone who want to build something on top of our protocol?

[00:20:51] Some tools guides solutions how to build front ends on top of our protocol and

[00:20:57] They're building their applications. So we don't care about

[00:21:01] About front end about UX. We're building only a protocol

[00:21:07] Now we have around

[00:21:10] 25

[00:21:11] Applications that was built by independent teams

[00:21:14] The best thing with this

[00:21:17] In this approach that all these teams create different project different products and

[00:21:24] all of them

[00:21:28] Create some creative ideas how to

[00:21:33] Provide user experience

[00:21:36] And we have a lot of

[00:21:38] Create some creative ideas how to

[00:21:43] Provide users with the best weeks for example, some of them create

[00:21:48] Telegram bots

[00:21:50] Telegram mini apps

[00:21:53] Some of them going for going through

[00:21:56] more usual way and

[00:21:59] Building just a web application

[00:22:01] But all of them trying to provide users with something unique to to compete with with others

[00:22:08] So I think that's very interesting

[00:22:15] Yeah

[00:22:16] Interesting you say telegram bots

[00:22:21] That's a problem

[00:22:23] I see it as a problem. There's a lot you can't tell if a bot is a person

[00:22:28] If a bot is a person an intelligent you can't tell but it's a person or a bot right?

[00:22:34] How do how do we solve that problem any ideas how to solve that?

[00:22:42] You mean that you need to wall it inside

[00:22:46] The bottom what well no I'm talking more of like community discussion and people undermine the future of different projects

[00:22:56] by

[00:22:57] By having their hologram bots come in and help and make negative comments about the community and about the project, right?

[00:23:05] You really can't tell on telegram if it's a bot or if it's a person

[00:23:10] So, you know, how do we so that's a it's a problem to undermine a community. How do we solve that problem going forward?

[00:23:18] Oh, well, I don't know. I didn't think about it in our case. I don't think that it's it's an issue

[00:23:27] That we have some boats or not for us. It's

[00:23:31] It doesn't matter if someone make a bet

[00:23:36] If our risk management model is good, if our odds is correct for us, it's absolutely doesn't matter if it's real person or if it is a bot

[00:23:47] We just earn money without margin and that's it

[00:23:53] So your answer is if it's a solid progress is a solid protocol and solid product and you have solid risk management in place

[00:24:02] It doesn't matter if a bot is making a comment or not. That makes sense. Yeah, that's good. All right, so

[00:24:11] You said something earlier want to talk about and that's the liquidity tree. Yeah, right. What is the liquidity tree?

[00:24:18] So, yeah, liquidity tree is

[00:24:25] Our

[00:24:27] Design of liquidity pool. It's based on data structure that's called segment tree

[00:24:36] Maybe someone of listeners know about this data structure segment tree

[00:24:41] It's pretty famous to be honest. Yeah, so the main idea here is that in our case

[00:24:53] We need to make a lot of calculation on the big range

[00:24:59] So and segment tree allows us to do it as I already mentioned with

[00:25:06] With really extremely efficient way

[00:25:10] How we can calculate

[00:25:14] How we can

[00:25:16] Make this calculation on the big range as I mentioned, it can be done in logarithm complexity

[00:25:23] So what does what does it mean logarithm complexity logarithm complexity means that for example if I need to make

[00:25:31] Those calculations with let's say range of 1 billion elements

[00:25:37] It's really huge amount. So I will have to do only 20 interactions 20 loop interactions instead of 1 billion. So

[00:25:48] That's why it's so efficient

[00:25:50] So maybe you can be curious why we need it because all other protocols who have liquidity pools as well don't need it

[00:26:01] So that's why that's because of our

[00:26:07] Of

[00:26:08] Of nature of our incomes. So because look for example if we are talking about uniswap

[00:26:16] Because in uniswap it's

[00:26:19] It's just a fees from trade. So when whenever trade happen fees goes to the liquidity pool

[00:26:28] But in our case, we don't

[00:26:32] Charge any fees from the bets when bets made

[00:26:37] We don't claim fee immediately. No

[00:26:40] Then it's a matter of time when

[00:26:44] This event will be

[00:26:47] Will be solved will be finished

[00:26:51] Only after that we will understand if

[00:26:55] Liquidity people earns money

[00:26:58] And then we will see that the

[00:27:01] Only after that we will understand if

[00:27:05] Liquidity people earns money or lose so and that's why we have this problem that

[00:27:14] for example

[00:27:16] We have let's say

[00:27:18] World Cup final and everyone understands that

[00:27:24] Will be a lot of bets on

[00:27:27] World Cup final because it's the most popular event, right?

[00:27:30] So and some liquidity providers can just see

[00:27:33] Okay, we have a lot of bets. Let's say on I don't know on Brazil that Brazil will win because it's like one of the best

[00:27:41] team in in football so and

[00:27:45] They can put a lot of liquidity just

[00:27:49] In I don't know just before few minutes the event will be finished

[00:27:55] You can put huge amount

[00:27:57] When event will be solved

[00:28:00] And you can earn money you can earn money immediately and then withdraw liquidity

[00:28:06] So that's why we we can't allow this situation and we should track

[00:28:12] If I put liquidity before event was created on our platform

[00:28:18] You will not earn money from this as liquidity provider. I mean you will not earn

[00:28:24] This margin from this particular event that was created after you put liquidity so because of that we should track

[00:28:33] timings of each

[00:28:36] liquidity deposits

[00:28:38] And it's possible only using this maybe not only maybe there are some other solutions, but we choose this one

[00:28:46] That we called liquidity tree

[00:28:48] And

[00:28:49] That based on this segment tree segment tree stuff

[00:28:53] So if someone interested about this more, I'm not sure that they explain it really clear to be honest

[00:28:59] because it's very

[00:29:02] complicated stuff so but if someone interested

[00:29:05] You can

[00:29:07] You could take a look at our github project

[00:29:11] It's called liquidity tree

[00:29:14] and there there are some

[00:29:17] Some proper explanation how it works

[00:29:20] What thing I think I got I think I got the crux of it

[00:29:25] There's no fees upfront

[00:29:27] You earn after the fact after the outcome is known, right? Yeah

[00:29:33] It doesn't act the same way as you know swap

[00:29:38] But I wanted but I went in here are the benefits of your of your approach. What are the benefits?

[00:29:46] Yeah, exactly for uni swap. It doesn't matter if you are profitable trader or not

[00:29:51] For them it's absolutely equal

[00:29:54] You just make a trade you pay some fee. That's it in our case

[00:30:00] See it doesn't work like this. We have no any fees

[00:30:04] we have only this

[00:30:07] Margin inside of the odds and because of that our income fully depends on how

[00:30:13] Some events will be resolved

[00:30:18] So and

[00:30:21] In front we don't know if we will earn from this bet or if you lose from this bet

[00:30:26] And then because there's no fees the it's a fairer

[00:30:32] Playground for the for the players

[00:30:35] Yeah

[00:30:37] Okay, that's your benefit great awesome

[00:30:40] I want to play

[00:30:42] You know

[00:30:44] Yeah, but it's usual approach for any prediction markets. They usually have fees all only inside the odds

[00:30:51] Inside the margin usually it's like this

[00:30:54] Yeah margin inside the world

[00:30:58] It was it's it's horse season here in the US and a few weeks ago

[00:31:02] They had the Kentucky Derby and then the preakness in the Belmont and I remember

[00:31:07] It was a decade ago I bet on this horse to win it was gonna go for the Triple Crown

[00:31:12] I was caught the name of the horse is big Brown and at the Belmont. I covered every single

[00:31:18] possibility

[00:31:19] And I said there's no way I can lose unless this horse doesn't finish the race and guess what the worst

[00:31:25] It

[00:31:31] Happens right so yeah, so

[00:31:34] I want to thank you very much for coming on my show today. I really appreciate talking to you

[00:31:39] I enjoyed it and have one last question for you

[00:31:42] How can people find out more information about you about Azura?

[00:31:47] You know get to know and get to use your put your platform. How can they do that?

[00:31:52] So, yeah

[00:31:54] If

[00:31:55] If you want to build something with us

[00:31:59] First of all, you can use our documentation. It's a game dot azura dot org

[00:32:08] Also our discord obviously

[00:32:12] There you can find any information about our project about our protocol

[00:32:16] We have devreel in our discord who can help you with any questions. I mean about

[00:32:23] Building on top of us something so about me personally you can use

[00:32:29] Twitter X

[00:32:31] So then Kaiser. Yeah

[00:32:34] You can ask any questions in the end

[00:32:39] Awesome awesome. Thank you very much for your time today. Thank you. Thank you

[00:32:44] Bye. Bye