Introduction to Speedrun Ethereum Challenges Philip Krause

[Music] Welcome everyone. Nice that you're here. Thanks a lot Kevin for showing us how how to build on on Ethereum on a blockchain. And I'm here now to show you higher up. Oh, I can move it.

That's better. And now I'm here to show you what you can build. I mean therefore we at buildg guild I'm I'm also from buildg guild I'm philip and we build speedrun ethereum that's kind of a accumulation of different challenges they start from very easy and getting more difficult over time and it's all done in a very playful way and we tried to bring all the basic concepts you need to know when you're just coming from the outside you don't know what you can build on chain to get very quickly on boarded I think that was also the way I I did do it I mean it if you go through it's very nice it's very playful and joyful. It's almost like a computer game that that's why really liking it and we really start getting easy. So first of all you getting your first NFT example and you really build with help with the help of scaffold a fullyfledged application.

So you even then deploy it on the cell and you can show off to your friends and for instance if you're going in in here you see you have different checkpoints you're going through and really describing every everything what you need to do and actually it's really good if you have not not much idea what's going on you can definitely really use that to go into it um of course it gets them more difficult we have NFT example we have a decentralized staking app that you know okay how does kind of a little bit of staking work into wa thing token vendor how you can swap tokens. then showing okay how do you could achieve randomness on the blockchain or even not then we have a more complicated example and that's probably a very eyeopening one that's kind of build a DAX and you build a very small unis swap for yourself which is pretty cool I mean for me it was very awesome to see that to deploying this on chain and see okay now I really have my own my my my own decks deployed and then it gets more more difficult we have some onboarding ch badges it's very cool where you can learn everything about contributing to open source products, tinker around in smart contracts. I'm leading that program. So if you want to join, I'm very happy to. And then we going into more difficult ones.

So we have overcolized landing, prediction markets, deploy to layer 2, multiix, and SVG, NFTTS, and then we have further further things down the road like getting into smart contract security and the Eve tech tree where we going more into DeFi and explain some core concepts. And for today um I would like to go with you together to the basic concept of overcolateralized landing and prediction markets. And therefore I would just jump into presentation. So we don't do so much code. I just want to explain really the basic concept.

What is the understanding for me? I think it's very eye opening over landing is like how how does a works? what are the basic block um blocks in it and prediction markets like what what is a poly market because I I would say these kind of two use cases are probably the most used ones on chain like money markets and prediction markets. Yeah. And that's what I want to take you um for for the next 20 to 25 minutes.

First let's start with overcolaterized lending. I mean I was mentioning already I mean it's pretty pretty big and probably everyone knows almost what a is and I mean they're holding in V3 by now 26 billion TVL so it's definitely a ton of money they're holding and just alone on manet it's 23 billion but what exactly is these overcarized lending or this money market we we're talking about and this kind of facilitates kind of that you are able to borrow against your own assets so you're kind of using your own assets what whatever you hold onchain. It could be Ethereum, it could be whatever token they support and you place it as a collateral like on on the one side and then you can borrow on the other side on this asset and in in also these both efforts they create some interest rate. So when you supply some some collateral on the one side you're getting interest for that. If you want to borrow you need to pay for that and this all is kind of locked in an onchain bolt.

And why I'm saying kind of overcolateralized landing and I think this is important thing of the whole protocol and actually everything in the financial system always the one side with the collateral needs to be more than the debt otherwise the whole system would collapse and this is really a premise that could not be broken. Even if you think about the traditional uh traditional borrowing, if you would get a mortgage for a house, the bank always takes your house on the one side and most of the time this house they assess it has more value than the that than the debt they're giving you for. If not, you have to put a little down payment more to kind of balancing this out. And also if you would get a personal loan for that they they use you as the person as the collateral because they make sure that you are kind of worth and you can create the cash flow in the future to really pay back the loan and that's kind of the collateral they're holding and they assume this collateral is also more than the debt they they're giving you and that's why I would say kind of all debts out there are kind of in a sense some overcolateralized now the only thing on chain because we have no real identities yet on the blockchain. So everyone can just use a for instance they cannot have anyone chasing you.

So no court no trial they cannot put you in prison prison if you don't pay back and that's why in the in the sense of a onchain lending borum protocol always the premise counts make sure you have more collateral than that in a protocol. If this is not given the whole protocol would would just break because then you can just borrow more than you would have put in and you would just make a trade out of it. um when you think of building or going more into the technical things you always have to cons consider in a money market three important parts and they are somehow all linked together I I just on on the one side you have the lending that's kind of the collateral part first that anyone can borrow anything in the whole in the in the whole financial market you need to put some collateral in so that somebody can take it out as well so the lending comes first then on top of that you have then the borrowing side of it and um it's also very important that you have getting enough landing enough collateral inside the platform so that also people are able to borrowing the coins or whatever they want to borrow on on the one side and what links it all together is liquidation and this liquidation part is a very important piece to keep the um the protocol healthy and what exactly how liquidation works I would like to show you in a little bit of an example I would just start with kind of a healthy loan of the one side. So I go in a little bit more how kind of a position in in a money market would look like. So we we we consider we have a healthy loan on the left side and we're putting in $200 just whatever whatever it is.

And each and every protocol uses a discounted rate on a collateral. In this case where you see I I wrote $100 that's kind of a discount rate of 50%. They assume even if your collateral is 200. So if you put one EV in and the price of ETH is 2,000 um and the money market assess it has only a collateral rate um a discounted value of 50% it would be $1,000. They assume you have so much collateral in there and then based on this discounted collateral you can get a loan.

So in this case we're going to get 80%. So we're taking out 80% USDC for instance. So we're putting Eve in. We're taking out 80%. That means we have a kind of a debt to collateral rate of 80%.

Now what's happening if the and now let's assume the collateral comes a crash and the collateral drops. What is going to happening now? We can say okay let us drop now $50 on the one side of the collateral. Then our discounted um collateral is only $75 anymore but we have a debt of $80. That means now okay we have an unhealthy low.

So it's not backed anymore. That's it is backed because they have this um the the the safety rate kind of this is kind of these LTV this the loan to value like this collateral this collateral ratio but for the protocol they assume you have an unhealthy loan and in this sense you are kind of have a liquidatable position so other people can come now in and liquidate your position and how does this work let's go further so we we we having on the one side we have the unhealthy loan now and we know that's an unhealthy low. So, and this is out in the world. There are people kind of on the chain. They see your unhealthy position.

And now what they going to do? They can get in with 80 $80 of the collateral and they're paying back the debt. So, they're taking out they taking out your completed debt and and and they're paying it back. And then what happens if if they're doing it? You have only $70 in collateral left.

And this is accounts for a discounted value of 35. But you see you have no no dollars anymore. So you still you not got completely liquidated but you lost a ton of the of the of the collateral already. And now and this is very an important part for this whole network to work that you have participants who doing this liquidation because if nobody would do this unhealthy loans at some point the whole collateral will be just having over that and will just break. And that's why when you're constructing such a thing, you really have to think about these three pieces.

The lending, the borrowing, and the liquidation. And that's definitely important. And there there could become different players. How you do it? Do you have your own liquidators for the protocol?

You have other you have others doing that. And of course, if you're having others doing that, you need to take in some incentive that what why would anyone do that? And that probably would work that you give them some certain amount. So if they if they liquidate you, they get maybe 10% of your collateral and so they they make money while they're liquidating and it helps the protocol tend to stay safe. Yeah.

Then yeah and when you when you're kind of building this I I here drafted out six different components what what you need to consider. It's definitely a little bit more. I mean you're starting very little and that's kind of our challenges for where you just realize okay how does this overcolious liquidation works? But if you then look at the protocol at all, it gets more and more complicated. I just want to want to draft out some of these core components.

I mean of the one side you're having your leting and depth tokens and I think this is clear. Then you need to um consider your interest calculation mechanism. How do you calculate that? Probably like a for instance doing it. They have a wrap token around.

So this just the token amount doesn't change only the value of this wrap token. So you don't have to rebase all the time your your accounts that makes it much easier for accounting. Then another thing you need to look at is the interest stretch interest rate strategy. So how much is each and each and every character how how much do you pay the lenders and how much do you charge the borrowers in this kind of sense? And I just here at EFCC I listen to a V4 what they're bringing out and are really liking that and there there tons of things going on because not every collateral you could assume is has the same value.

So imagine you're putting their ETH in it might have a higher valuation or a higher value than you would just put some ESC20 token in. So you probably want to give the people maybe a little bit more um lending but also when you're borrowing then against it they probably should pay less because they have a more safer it's more secure. It's the same like in traditional finance when you have a very high proper real estate and it's a lot of value kind of and you don't know it has much volatility they probably charge you less in interest rates then you have a very I don't know price property not such a good region that's kind of thing in there you need to consider then a very important point and this is very important is price oracles because you've going to feed in the price into the into the market and you really have to make sure this you do this right because if you do it wrong. It means when you up the price on the one side of the collateral and you're saying, "Okay, now you're getting price in instead of 2,000 Ethereum, 1,000 Ethereum." Everyone believes everyone gets liquidated.

Oh, sorry for that. Yeah. And then the same is like the the collateral ratio logic. How do you combine it together? Is everything mixed in a group?

You separate them from each other. that could make lead to more safety but also to a less more liquidity and kind of different things you need to consider in that and then also having this kind of liquid liquidation engine in place how you how how you do the liquidations and I think another core function of these money markets and I really like it and they just exist onchain are flesh loans I don't know if you've heard about it but I really like it cool and this is just something that kind of exists inside on the blockchain chain because you can just get a loan for no collateral at all. But the thing is this has to be done in one transaction. It's just in one transaction. You can imagine you're borrowing in the same transaction something you're doing something in between and in the same transaction in the same block gets mined.

It needs to get repaid and if it's not getting repaid there's kind of a search function in there. It will just um um re-re the whole transaction. Nothing really happened. And that's really cool because it enables a lot of things like you can do arbitrage. You can do collateral swapping out.

Imagine you're having on the one side you're having die sitting and you want to on the collateral and you want to change it out with ethereum. Normally what you need to do you need to pay all first all your debt down then you're taking your um your Ethereum out or your die you're taking your die out and then you need to put in the the Ethereum and then you're borrowing again. And this and this um this one these transactions bundled you can just do one with a with a flash loan. It could be for refining or even for liquidation. I mean if we're coming back to really having these liquidations in place you can some some people who wants to liquidate they can do that without even bringing up the liquid the the liquidity for the collateral and that's pretty cool.

Also I want I also mentioned ton of these risk the different risk you need to consider in production. And it's kind of the oracle manipulation risk. I already mentioned it's very very important. The bad that risk that's really that the liquidation doesn't happen too fast. If there is a market crash and you don't pay attention that you're getting fast liquidations in just happen that your protocol is not um not um um um over collateralized the same collateral volatility risk that's about which collateral rates you're taking in.

So you don't want to be too risky there. If you're taking some meme coins and you borrow and you borrow against them of for sure you will have high volatility and that's probably not what they what the what what the call can take. There could be then some things like you're taking isolation mode. You're just looking at this one collider and really take it out from the rest that you just really secure your whole protocol. And then of course important thing is the kind of the whole market around the interest rates and the borrow rates because that that's that's when you when users um go come to your market.

If you have no good rates, nobody will probably show up. Yeah. I mean some closing thoughts why why why we picked this for the collateralized landing channels challenge I mean it's a very basic and great concept to learn how this works because you can build on top of that with this overcolateraliz for instance you can over build overcolaterized lending um stable coins and it's also a basis for a lot of derivatives where you can just collateral and you mint against um some derivatives like um synthetics do it and so on so it's very worth learning it how it really works, how the liquidation works and from there on continue then bu building on. Yeah, that would it be to over collider lending in a nutshell. I mean we can do also some quick questions to that topic if there is one question right now otherwise I would jump into um prediction markets.

No. Okay, good to go. Let me go to prediction markets. I mean it's the same. It's just it it's just very very big use case.

I think if you looked at the elections just recently in the US just really there was wagering 3.3 billion on um on on on the on the president. I think it's a huge amount just in one little prediction mark and it really showed that this use case is really now in the masses and it's out there for users to use. Um it's pretty cool and I mean prediction marks in general I'm I would consider it like a stock market but really for trading the shares not for the companies but for future outcomes. So that what you literally do you just decide okay I I believe on that and then I just bet on that and buy the stock from this future future output.

The thing is with these prediction markets right now out there is that they always need some public attention. If you're putting out some questions for the future, it's and is just boring, nobody would care about. So, it always has something to do that's relevant, what's going on today, and and so on. Um, a prediction market itself is really a question you ask. And there could be different types of prediction market, but but here we're focusing on binary uh bin binary prediction mark.

Means there's a yes and a no outcome just two outcomes yes or no of a question and what is very important that has a fixed end date. So you need to know at what point in time you want to resolve the market there could be not really open end and it needs because at some point it needs to get resolved and the oracle needs to um to to report the the result. If we take an example for for for instance poly market how it works you each and every outcome token if it comes out so so if we're going to ask a question and the outcome is yes the winning outcome for this token is one USDC so $1 and in in the meantime before the whole um the whole prediction get resolved it kind of fluctuates between zero and this $1 all the time and also this one and if if for instance it cost 60 cents. That means that kind of the probability for this outcome the the prediction mark assume is also 60 um is cost 60 cents and is 60%. And a very important fundamental rule when building this and like it when I when I did this and broke this everything down it's that you always need to make sure that these two prices come up with the winning value in this in this sense one USDC because if it's less than one USDC people can just make money because they would just buy and buy if they can get a yes and a no token for 90 cents together they would just take both sides and they would always make 10 cents.

So this is very very important that you don't break the protocol as well. So it has definitely similarities with the overcolater landing in the from from the other challenge. I think a very a very cool primitive and what I really like is when you think about just just more abstract that you kind of having some opinions they're out there and these opinions they kind of leading to some prices because people think ah I would definitely believe in that it's only I don't know 30 cents now I going to buy it and this then on the other side leads to probabilities and these probabilities get used to do some future outlooks for the future. I think that's a very cool imaginary um funnel how how how prediction markets work. Here also I want to highlight the the building blocks or you need to look at when you want to build a prediction market.

You see it's a little bit less so it's less complicated but not not the less it's very complicated when you're doing it. Um we have on the one side we have tokens there could be um I mean that's that's pretty easy. You need yes or no tokens on the one side. You can more more complicated if you go into more complicated token frameworks where having some conditionals and so on. You can also go for ESC 1115 where you only have one contract and don't deploy for each and every token more contracts.

Then one of the most and probably difficulty difficult parts to um implement is kind of the trading and the liquidity provision mechanism. If we if you look for poly market what they do they have an order book and the order book is not on chain because you can just not handle the sure amount of data. So that is outside the settlement logic is on chain but you always need in this sense when you do an order book you always need a a counterparty who buys your tokens. So I if I want to sell my yes token I need someone to buy the token. So you have always this connection.

You always need a counterparty. You cannot just sell. That's kind of how order books work. If you look in Binance and you have the order books there's always these kind of positions. Another mechanism for this that you kind of use an automated market maker like a unis swap like a DEX where you kind of separate the kind of the liquidity providers and the traders and you're trading in and out of this pool that's kind of in and in in a deck.

So so you have the pools and liquidity providers provide the tokens they make some money because the of the trading fees and you can just skim in and out in and out of this one. The the downside here is I just I don't know going into here when you when you're doing that you need to imagine whenever you deploy a prediction market and we're using that and you and you're just deploying for instance for one E you're putting one E in so the whole market is one E and you're just creating two tokens it's just easy there's just a yes and a no token and I would put the market now out and Kevin would buy now is the only one who buys the yes token so he gets an yes token because it's I start with 50/50 50. It's it's it's half an E. So, he goes in, buys half an E, and that's all what happened in this prediction market. And now the market resolves.

So, yes, he gets one E. And what did I lost as a as a liquidity provider? I lost half an E. So, it could be very painful for for liquidity provider. So, you really need H.

Does it work again? Oh, yeah. Don't don't see it. Yeah. So you really need to make sure and that's definitely a thing to play around with.

You need to get a feeling how prediction markets work because you really need to make sure that if you create a prediction market that you don't lose so much money. I mean it could be also the other way around that you're going to make money in the sense if Kevin would bet on yes and the outcome would be no. I put one Eve in. So that's my kind kind of my token but he got got also half an E from him back. So I made made half an E on that.

That would be the other outcome. But you need to be aware if you're constructing this that it comes also with a financial risk for the liquidity provider.

What do I need to do?

Okay, I need to move the mouse. Okay, thank you. And then the other part we we had it before as well and that's for um is also the oracle part and here it's not only you get p you need price information, you need real world data that's coming in. I mean the price information could be also real world data in the sense of the overolaterized landing but here it's really like you need real world events that coming into the blockchain and you need to make sure how how that get gets reported and there could be some op optimistic um oracles like UMA protocol could be used for that and so on um yeah that's kind of the the challenge the next one yeah I mean when I just want to show you this kind of this this question here and I think it This should illustrate what kind of difficult question could be out there. I mean, but it would be nice to solve.

But the issue is if you see this question will crisper based therapy for duchian muscular droy reach phase three trials before cool 3 2026 I would be completely lost. So I had no chance and there probably just a couple of people I don't know if you can answer this question but there probably just a handful of people or I don't know how many out there who could solve this question. But the thing is how do you find this people? How do you get their attention and how do you make them sure that they get well enough pay to answer this question because probably these are some kind of very expert in that field. Yeah.

And that's kind of where where the limiting factors of humans come into play. It's the one thing I already mentioned is the attention then how complex gets the topic and how much effort I would need to put in. And also with the attention I mean there's so many going on out there and you're having poly market and here and you're having all these question who should who should get an overview over that and also we have all these kind of emotional influences and I think that's is very cool and many people looking in that as well there come AI agents into the play where you can just oh five minutes thank you where you kind of using these AI agent to solve these kind of questions even pay them some money for it so they can kind of survive and you would can kind of create a kind of a Google forecasting I would like to call it and I think I really like it and Kosis is doing a lot of them with the agent framework and there's a lot of cool things to explore in the field. Yeah, when you when you learn more about it like for the money market like the overordis landing or the prediction market really go to speedrun Ethereum I personally really like it a lot. I mean it helped me a lot to understand the basing concepts and kind of start my journey in the web free space and learn how to build.

It's a really cool thing to do. Yeah, really recommend it to you. Okay, now we have another couple of minutes if there are any questions.

Okay, thank you very much. [Music]