 Awesome, guys. So I've been listening to the talk so far. My talk is super DeFi focused, which I think even if you're not as interested in DeFi, I believe this crowd should be more interested in decentralized storage and everything that protocol labs has been bringing up. Even if that's the case, I think there are plenty of game theoretic and also incentives, mechanisms that I'm going to cover that are finance focused, that could be applied elsewhere. So that could be an inspiration just to bring something from a different area. So the topic that I wanted to bring up today is that in DeFi there are various different incentive mechanisms that create very new types of markets. And basically, the topic for this one is that there has been a market emerging for like one year or so in DeFi specifically. That's very particular because it is a three-sided market, while most markets in the world are two-way or two-sided markets, right? So whenever in macroeconomics and other disciplines like that, we think about markets in general. It's redundant to say that they are two-sided. They are by default two-sided, right? A couple examples, very traditional, very normal ones. Any commerce store is a two-sided marketplace between buyers and sellers. Web 2 has brought up various new two-sided marketplaces as well, so between hosts and guests, drivers and riders. So all of these function according to normal microeconomic principles of supply and demand and prices basically matching those, right? But just between two parties. And what I'm going to bring up today is a market for liquidity, which is one of the most important things in DeFi and also in other crypto areas, let's call it that. That is three-sided between very different, basically very, very different three agents that participate in it, right? Before we get into that, I have to just make a very simple introduction to the Votescrow model, which is a tokenomics model that has been gaining a lot of traction ever since like 2020 and until like 2021. In DeFi specifically, it was pioneered by Curve. I'm going to simplify a lot of mechanisms of how it works. We're just going to focus on the essential thing so that we have that background info to then proceed to what are these three-sided markets, right? So this is a simple diagram for how it works. Again, oversimplifying a lot of it. Feel free to just reach out to me afterwards if you want to discuss it in more depth. But essentially what we have is that we have protocols and DAOs that emit a constant amount of tokens each time period, right? Each block, each week, each month, whatever they decide in their particular implementation. And so that's what's there in the left as token emissions, which are liquid. They're actually tokens that the protocol mints or emits from their treasury for some incentivization. That incentivization is LPS liquidity providers. So this model has been used by DEXs, the centralized exchanges that need people, agents to provide liquidity for people to trade against that liquidity in those exchanges, right? So we have various pools. It can be a number of pools. And we have people or organizations, any type of agents providing liquidity towards those. And then we can see that the arrows from the initial token emissions have different weightings to the different pools. And you're going to go over what specifies or what determines what's the weight that each one receives. That's the interesting part here. But basically, and jumping that step to come back to it later, LPS receive tokens that are liquid from these DEXs, from these protocols that want to incentivize liquidity there. They have two main choices with those tokens that they receive. They can basically sell them for yields. So they can just sell them for dollars to reinvest in their positions or just take some value off the table. Or more interestingly, and more pertaining to this model, they can lock those tokens up. The way they lock those tokens up is called Votess Pro. That's the whole name of the tokenomics. And basically, if they do that, they're willingly opting to lose liquidity on those tokens for a fixed number of years. Normally it goes between one and four, but again, those are implementation details. If they do that, what they get in exchange for losing the liquidity on that is that they gain governance power over this protocol, right? That's what they get in exchange for losing the liquidity for a fixed time period in the future. So in this mechanism, the only tokens that have actually governance power are the ones that are locked. So the liquid ones are just financial instruments that trade and are able to be sent around, but they don't have governance power. The ones that do are the ones that are Votess growth or locked. So whenever you lock them, you basically get a V version of the token that you had liquid before. Then what happens is that these people, or these agents in general, because they can be institutions, or doubt, whatever, they get to vote on different governance proposals for the protocol that implements this tokenomics design. There are plenty of different votes that they can vote for. The one that's interesting for us for this talk is that they can vote for what's called voting gorges, which are basically a list that matches almost one by one the list of pools that are listed in these decks. So basically they can vote for which gorges they want to have more whites in terms of voting or less. What that does is that at the end of each voting round, which is normally one week, but again that's an implementation detail, basically it comes back to how those liquid tokens are distributed in the next round. So what happens is that the percentages of the votes on each one of these gorges, that again match one by one the pools that we have, will determine what's the waiting given from the initial tokens that will be distributed to which one of the pools. So if a gorge gets 30% of the votes, then in the next round 30%, again there are some details there, but simplifying it 30% of the tokens that are emitted for that round will be given to the LPs in the pool that got 30% of the votes. So that's the basic model. Here we have the actual real-world example of that model. The previous slide was the abstract version. So we have CRV emissions, which is a token. It's one of the biggest decentralized exchanges in the world. Those are liquid tokens. They are ERC20 implementations pretty standard. They're given to different LPs to compensate them for the liquidity that they're providing. These are just a couple of random examples of actual pools there. Then these guys can sell the CRV for yield and just get paid from that. Or they can lock the CRV, get VCRV and these ones are locked for between one and four years. The longer they lock the more governance power per token they receive, basically incentivizing them to be long-term aligned with the protocol instead of just being there for the short term to profit quickly. So then the VCRV holders, and again these people don't need to come from LPs. They originally do, but they can just be agents that purchase CRV in the open markets and then lock them up. So there are various ways of acquiring that governance power. Then these guys vote for the gorges. And that's an example of basically one of those voting rounds after being over. There's basically different allocations for each. And then those slices in the pie charts will determine how thick those arrows up there are in the next round. So this is basically the normal overview of Votes, Crote, Economics. Other protocols that implement that are balancer now and a few others coming up that are more recent. Now we're going to go over how this creates a three-sided market. I think this is the most interesting part of the talk. Until now it was like context for us to go over this now. And this is, again, it's very rare that we find a market between three different sets of agents that have different incentives all between them. And that's why I think it's an interesting topic. So this is basically the market structure for this. We have on one side liquidity providers. We've covered who those are. Basically people depositing tokens on LP pools in these Dexes that are basically allowing other traders to trade their tokens against the liquidity that they have provided. Then we have liquidity seekers. We're going to go over in the next slide who those people are or who those organizations are, who those agents are. And then we have governance power holders and those we already know are the ones that hold the vote escrow tokens, the V version of whatever governance token we're talking about. So the way this works is that governance power holders will direct with their votes new emissions, so new token emissions that are liquid and tradable and have a market value to the liquidity providers. Then liquidity providers are providing liquidity for the seekers. We're going to go over who they are. And then this side of the market should be incentivized because there's money on the table, there's value on the table to compensate in some way the governance power holders because the more they vote for the pools that we want to make liquid the better for the seekers. They want to have some sort of asset liquid on chain. So simplifying, we have the LPs here for liquidity seekers. It's mostly thousand protocols that want to make either their own governance token liquid for various reasons or if they have products that live or buy by their liquidity. We're going to go over a few examples of those. And then we have V tokens, the people that hold the actual votes and governance power there. And then finally the actual real world example. We have different pools with different people providing liquidity. We have a couple dows and a couple protocols here, Lido, Frax and Rocket Pool that have products that they do want to make liquid. So in one case it's liquid second derivative, so a liquid way to stack tokens in POS chains and then this is a stablecoin from Frax. That also should be liquid and tradable one to one with dollars. That's what makes the stablecoin work. And then we have the V token holders. There are plenty of examples we've been covering curve, but there are plenty of others. Now we're going to go over how these markets actually work in practice. So up until now what we have seen is how the LPs and the V tokens interact, right? On this previous slide here, this is just a diagram between LPs and then the voters. And now we're going to go over how the voters and the liquidity seekers interact next. Since it's the three side of markets there are basically interactions between all of those pairs. So the types of markets that have been developed around these concepts and around this opportunity that lives on China are, I brought here a couple examples. These are actual implementations of different protocols that have markets for this. And again I'm going to oversimplify how each one of these works in the next slides, but the overall picture will be hopefully somewhat useful or interesting. So this is one of the models. This is the model that Hidden Hands implements, but there are other venues for trading these votes like that. So what we basically have is that we have, again, if you remember, we have basically a list of pools where people deposit liquidity in these decks. Then we have a list that matches pretty much one-to-one of gorges where people vote for the pools that they want to make more or less liquid in the next round. And now we have another list that also matches close to one-to-one with some exceptions to the previous two ones, which are basically, we can call them incentivizing or bribing gorges in some of these protocols. So basically what this is, it's just a pool of capital and a pool of votes. Let's call it that abstractly. What happens is that liquidity seekers, agents that want to make some token liquid on chain, can just deposit funds and these funds can be in their governance token. The token that they themselves means it can be in ETH or in stable coins, whatever they want to incentivize with, and they just deposit those funds in this box at the top. Let's call it a box. And then we basically have the other side of this market, so the V token holders, the actual voters that get to a website or get to a contract and just see a list of different boxes like this with different amounts of funds deposited by different liquidity seekers. And then since they have the votes they can just choose which one of those offers is more interesting for them to allocate their votes for, which will therefore allocate emissions and new token incentives for the LPs in those pools. So these people or these agents, again they don't have to be individuals they can be institutions, DAOs, whoever wants to participate can just look at those lists and the ones that have, and basically here they have a decision, which is the way it will work in the end of the round is that all of the funds that were in each one of these boxes will be distributed proportionately to the amount of votes that that got got. So for example if we have three of these, one has one million USD, another has 500K and then another has 100K, it's not clear which one of those would be the most profitable because the one with one million may get disproportionately more votes as well. So basically those funds will be distributed evenly and proportionately to which one of the votes that ended up committing towards that pool. So basically these people, this side of the market will commit their votes to one of the pools and then split the funds that was in there at the end of the voting round and then the liquidity seekers, 1000 protocols will deposit the funds at the beginning or whenever they want in the voting round and then they will get votes, voting for their gauge and thereby incentivizing new emissions to their LPs making the pools deeper. This is one of the models the second model that I wanted to bring up it works in a very different way but has interesting pros and cons and trade-offs that we're going to explore at the end and that's also the interesting part too. So the way this works, this was implemented by Quest which is a Paladin product, it's that basically liquidity seekers, 1000 protocols, can just make offers for how much they're willing to incentivize per vote that they receive at the end of the voting round. So this is kind of like the same model that again this is kind of an analogy with some points that it breaks under but it's basically the same as a market maker on a central limit order like a public exchange making offers for how much they're willing to sell or purchase a share for fixed prices like a limit order where people or not people but liquidity seekers just make a fixed compensation per vote that they're willing to pay and also a cap up until how many votes are they willing to purchase or incentivize during that voting round. And then the V token holders the people that hold the voting power can just go there, look at the different offers and commit their votes to the one that's more interesting to them, right? So this is another model with fixed costs or incentivization amounts per vote while the previous one will have variable incentivization or compensation amounts per vote as we're going to see. So the pros and cons of these two these two first points are for the liquidity seekers and then these last points are for the voters themselves. So in the first model again where we have a pool that then splits between the different voters proportionally that's voted for that. What may happen is that the liquidity seeker in general may end up the round overpaying per vote that they got. The reason for that is that they commit the capital towards those pools before knowing how many votes are actually going to go there and that number is determined by market forces and by the competitive landscape of the other incentivizers as well on that platform. So they may end up overpaying per vote but they may also end up underpaying per vote and being ROI positive on that incentive, right? The good thing is that they will always get votes using that model because even if they were just incentivizing a very minimal amount like one, let's call it like $1, just like deploying $1 of incentives for people to vote for their pool what would happen is that if it didn't have any vote until the end of the period someone would have the incentive to add one marginal vote towards that and just collect all of that because there would also be no competition so since this balances the amount of votes and the capital that's allocated to them what happens is that everyone that incentivizes ends up getting votes because marginal votes should go if we're talking about rational economic agents to the pools where the payoff will be better and if there's one with zero voters the payout will be like infinite because you'll get all of the capital that's incentivizing that one with just one vote. So those are the pros and cons. On this side again this one is the fixed compensation per vote model here on the right what happens is that an incentivizer, a DAO, a protocol a liquidity seeker assures from the beginning if they run proper models that they're pretty certain that they will end up getting ROI on their incentive deploy there because they can just fix and set how much they're willing to compensate for each one so they can just set it slightly below whatever they believe the expected value per vote will be in the next round so they have more control over that which is important. The thing is that whenever we try to fix prices in economics in general what may happen is that supply may not match demand. Whenever either government or other external agents from the market try to fix prices it may happen that at that price there's no matching between supply and demand and here it's exactly the same so basically they may end up getting zero votes if all of the other options that live on chain for the same round are more profitable for the voters they'll just end up not getting votes which can have harmful effects in terms of their pools are losing incentives and thereby losing TVL, thereby losing liquidity thereby making the product worse. Then here it's basically the point of view of the voters themselves so in the first one whenever they're committing their votes they don't know how much they're going to be compensated per vote that they're committing because again there are all of those market dynamics until the end of the round so they get a variable or unpredictable reward which may be good or bad so they basically may end up getting paid less than they would have if they took a similar fixed price offer however they may also end up being compensated more and it's kind of variable and there's no good way to just lock in how much you're going to receive. Again that can be good or bad it just depends on your risk app side there as a governance power holder and here you know whenever you take one of those fixed price offers that you're going to get compensated that amount so there's more certainty less upside and less downside so it's just a different risk profile it's not like one of the models is better. These are two different markets that have been live for a bit and I expect that new ones will come with new designs and new trade-offs between the different agents. So that was the talk again this was a very short intro I had to oversimplify a lot of the V tokenomics side also this bribing side as well so if you have any questions I'm super happy to go more in depth or also you can just reach me this is my email you can also reach me outside. Super happy to talk about this type of stuff so thank you for the attention and yeah. Cool thanks I have a question I'll kick it off. So I think I'm still trying to get my head around the reason to add the third component in other than it's just like interesting and fascinating so was it if you say it's running now is it are both of those like separate smart contracts on Lido or is Lido doing one and one other one's on a different? Ah cool so this is not what Lido does basically Lido is a participant in these markets but Lido didn't build these infrastructure these are other markets on train that live Lido is one of the liquidity seekers that's why I ended up diving into these types of markets because I do mostly liquidity research and think about liquidity on train and stuff like that so Lido would be categorized as one of the liquidity seekers one of the agents that want to have votes incentivizing their pools so Lido didn't build these types of markets it's just one of the participants in there. Okay okay and so then to add that third step in the market is the purpose to create an ability to distinguish like hey this is a better term vote they locked more tokens and put more value locked into the system so that's more stable they put those in so that they're gonna last longer so they get more votes so that they can then go vote for their delegative choice to get more rewards was the original is that the kind of reasoning or is there something else kind of there that I'm kind of just to clarify if I understood correctly the question basically you're saying that this could be done as a two-sided marketplace instead of having people compensate people that will vote and then these ones will determine the emissions afterwards why don't the first group just direct their own emissions or their own capital to pay LPs would that be kind of the general direction? I think so yeah I'm just trying to suss out like why add the three together does it make it inherently more stable or no it's basically that this was basically like generally crypto all of these tokenomics designs are being experimented with this was one of the designs by Curve originally that is now being adopted by other protocols and that's kind of like an empirical argument for it being kind of working at least if other protocols are adopting it initially Curve didn't design this to have all of these markets around it but they basically just build a primitive that ended up allowing people to build this type of stuff over it the initial premise of V tokenomics and again I didn't build this so it's kind of weird that I'm talking about what was the original premise but from my understanding is that you basically have an incentive for people to lock up your tokens your governance tokens for a very long fixed period for example four years most an interesting stat is that most CRV is actually locked for a very long time so it actually is what happens in the market people do choose to lose their liquidity and lock it for a long time normally when it drops from like four years it drops a couple months people just relock it to have the maximum voting power per token that they can have and then you basically have all of these people aligned with this DEX long term one of the problems of decentralized exchanges in general is that it mostly in DeFi liquidity is mostly mercenary so it just goes to the DEXes that are incentivizing more at each time and there's it's difficult to create a lock-in and a network effect and the defensible mode as a protocol there in DEXes so this was one of the ideas to create a mode a defensible advantage to some of these DEXes so basically if you have millions hundreds of millions of capital locked up as VCRV you have a very large capital base that will keep incentivizing these pools for years to come right so it's more difficult to siphon liquidity out of curve than it is to siphon liquidity out of for example SushiSwap or others that don't have that so that's kind of the rationale for a DEX to implement that and then Balancer very interestingly adapted this model with a very interesting adaptation that we don't have time to discuss but they did end up and Balancer is also one of the most prominent DEXes in Ethereum DeFi and yeah they did choose to go into this model which is kind of like a proof that it was at least somewhat working right like an empirical argument for that instead of just a philosophical argument of does it work or not it is generating a lot of volume and a lot of liquidity so there should be some value at least in that kind of design this talk was not as focused in terms of explaining the mechanism it was just to use that as basis to then go over the market and how they interact with each other and all of that yeah awesome yeah that's super helpful thank you other questions to add a little bit I think that the reason why it's separated to 3 is because it's just separating out like the people that have the incentive of just holding the token and getting extra yield versus the ones that are going to optimize it and find like the most lucrative place to put it so that's probably like the impetus behind the third but my question was if like for curve and balancer as you said like the reason why that they wanted to do it is both to reduce mercenary capital and then also to like increase the amount of LPs because theoretically you're raising the yield and you're going to get it back to a yield that's like sustainable so like there's going to be more LPs flooding the system that aren't just going to be like mercenary like switching in and out between different pools as they start to optimize like is there any research that you know of that's like gone through and check like how much net new LPs has gone on because of this as opposed to just people being like opportunistic and like flipping back and forth between what's the most lucrative within it awesome and I agree with the first point there but yeah awesome question my view on that is that it's very difficult to understand causality in DeFi in general we can look at correlation a lot it's very difficult to try to get causality from that my view is that this model as a mode as like a way to defend liquidity in a dex that you're building should kind of work there are some flaws that I believe it has but it should kind of work because you're basically using your market cap the market cap of the dex like a curve balancer, velodrome whatever as a way to defend against new entrants and other dexes that could take like fees out of the pools and all of that because normally there's like three ways the three main ways by which LPs are paid in DeFi or compensated in DeFi one of them is just trading fees that's the most sustainable one so people trading on against that liquidity and the point fees and they get that but that's a very there's a bit of a very slight bit of a mode there not too deep then the other one is just with direct incentives so liquidity seekers, dows, protocols etc can just directly incentivize that so just give their own tokens USDC other stables etc to those LPs and then the third one is this so basically compensating votes that will then direct the native dex emissions towards those pools this last one it's kind of creates a dynamic where there's at least always a baseline of TVL there in those dexes why is that? It's because trading fees are super volatile right but the emissions that the protocol itself is emitting to an LP it's not as volatile because for example balancer emits 145k BAL tokens towards all of the LPs there per week so as long as balancer has a good market cap that will be a constant incentive to provide liquidity on balancer even if the trading volume decreases a lot right again on curve same thing they emit a fixed amount that keeps declining over time but they emit a fixed amount of CRV emissions towards those LPs so there should always be some incentive along as those dexes are successful to LP there which should keep liquidity big sticky one of the things that you mentioned is that wouldn't this still be mercenary capital that is there because of these incentives and then can go elsewhere that's totally true but the dexes don't really care who is LP they care that someone is and so even if it's mercenary and other people come in the important thing is that there's always incentives to LP there and to liquidity provide there so that would be my reason in terms of the research I'm not aware there's very very interesting research about like how AMMs work and how dexes work and all of that by Guillermo and Gary's by Tarón Chitra and others but trying to figure out this causality between is it because it's a V model that has protocol emissions towards that or it's just like a unisoft model without emissions that just works without that I'm not aware of any research there yeah do you even think it is a benefit to smooth out liquidity during like volume times like why would that actually be a benefit because for the protocol they just end up paying a lot more for TBL than what they're getting out of it in terms of fees so it's almost like a net negative for them to keep providing liquidity like keep providing those incentives during that yeah basically it's LP's in general in DeFi are also kind of flywheel where traders should be incentivized to trade wherever the deepest liquidity for their pair is and then LP's should be kind of incentivized to LP or liquidity provide wherever the most volume is being traded right so what again this is like market shift too much to side this with like full confidence but what should happen is that the most liquid venues should get more liquid over time because they're attracting more volume and thereby more trading fees which are sustainable and then more LP's because they want to capture those extra trading fees so it's kind of difficult to bootstrap that liquidity because if it has no depth there's no people trading there so there's no new liquidity incentivized to go there so it's kind of a bootstrapping problem so my view is that it would be kind of bad to have these dexes have their TBL go down super low there's a period with not too much trading because then it would be super difficult for them to capture all of that liquidity back normally what happens is that like these most liquid dexes even UNISOP that's very different from this model keep their TBL quite high over time at least priced in tokens and priced in dollars it's just a market force but it's not very common and it could be something that you could try in the future just to build something that has liquidity in time kind of the idea but what has been normally done in DeFi is that they try to keep their depth high over time so that they don't have like constant bootstrapping problems whenever they would want to come back so that's kind of like more of an empirical answer not as much like what I believe it's more like what I've seen awesome thank you