Tokenomics in 2024, Embracing Innovation and Avoiding Mistakes — Aleksandar Damjanovic | MVPW

uh I'm going to be talking about economics uh in 2024 it's a pretty General title if you look at it it looks like I'm not going to talk about anything like that's going to be applicable but I'm going to try um so let's start so a little bit about me why am I the only one speak speaking here today it's pretty much uh I think I've been working in like token design token engineering simulations on the technical side for around two years two and a half years full-time I worked in web3 R&D which was mostly my role was mostly focused on defy and mechanism design how to optimize for certain goals Etc I also participate in various tokenomics open source tools and projects such as like UTM from outlier Ventures uh Red Cat talk engineering Academy um bunch of free contributions Etc um and past year uh we worked on me and my team we worked on economic auditing of chains that are worth around like 800 million with like half a million or a million people um million wallets million users Etc uh we also worked with Enterprises um with like designing toomics uh for like various uh incentive schemas etc etc um for quite a large Asian company and of course uh just we just finished like a research for like decentralized RPC providers how to align certain incentives how to prevent dumps etc etc so this is a little bit about me uh mostly I guess I feel like I'm I'm the only one like tackling this stuff like for two years like full-time um what I a to like showcase is to don't really uh tokenomics is like a term that has stuck around but I'm not a big fan of it but let's STI let's stick to that uh I prefer to to call it like crypto econ or mechanism design Etc so what's going to be agenda I'm not going to talk about like types of tokens because everyone heard like different like definitions what what they are what types of tokens there are Etc uh I'm going to give like a short overview how far we came Etc um and of course like some tips for tokenomics uh tokens as a ra fundraising tool some like General advice um some tooling methodology etc etc so I'm going to talk about parameter selection Etc so let's start so toomics it didn't like come out from fundraising it's basically from the behavior psychology where you if you like have some like behavior issues you would get tokens if you do stop some good stuff and then you can trade them for like better meds or whatever uh um and I'm going to quote Trent mccon I think he's the the first one that like coined the term token engineering and he's the one that tried to um introduce like System Dynamics into tokenomics modeling so what are blockchains they are incentive machines to get people to do what we want uh and what it tackles exactly it doesn't just tackle the pie chart and some like token go up mechanics it tackles like the supply distribution mechanism incentives governance utility sustainability and monitoring um why are the last two things are important sustainability monitoring because once we get something live then we have some data and we can do some optimizations hopefully um why is it hard to set up uh what Founders want from a token uh generally they want some utility because uh they they need to talk to have some utility and to be a necessity in the system and they also want to most oftenly they want to raise funds and then they want to continue funding efforts we selling the token if you look at those three goals they're completely different goals maybe the last two ones are like uh similar but it's still hard to set things right and of course they want liquidity on the market what does that mean that means that if you have a billion tokens at your disposal and no liquidity no demand no nothing it's it's pretty much wor you can sell it only as much as it's liquid and they of course want want it to go to the Moon whatever uh what retail wants from a token Moon not to get Dunked On by VCS uh liquidity and some utility um I guess um and of course we have this like grass flower so that showcas is um what we actually need to know uh if you want to tackle the toen token economy design in some in a some like way that makes sense and that is you need to know like Game Theory you need to do like research you need to know control theory and everything because you might have um a mechanism that is perfect on paper but maybe if you look at the regulation Etc it just doesn't work or it's just a plain old Ponzi scheme so I drink a lot of beer before so so sorry um so why it's hard to set up of course like everything I talk about is not Financial advice I'm not going to I'm not Silling any of these projects um uh you need to if you want to have like a unicorn you want to have like a project that is good you basically need to have a good product and good toomics before we had like various like bad examples or or good examples such as like if you have a bad product and good toomics which happens with forks because like like with Unis swap Forks or whatever Forks usually they have like the good economics setup they run some modeling but they have no liquidity uh and of course you have like a bad product and bad toomics which is a Salu token which was basically a one one big pony scheme that depended on the influx of new users and of course you can have like a good product and good economics just such as ethereum with the proof of stake with like Dynamic fees with uh e 4844 Etc so why it's hard to set up uh if you if you are setting a project a web Tree Project you need basically to have if you want to be like an unicorn or at least take over some part of the market you need to have good product and somewhat good economics and good Optics of course um this is like a short survey it doesn't look as pretty as I wanted wanted it to be but basically this is how far we came from from the beginning of the defi stuff uh tokenomics launches Etc we had Bitcoin which is like scarcity uh fixed Supply whatever we had proof of stake air drops uh ethereum icos uh first Ico was Master coin uh we had air drops real world asset like in 2017 borrowing Landing bonding curves um we had bonding curves which are um quite an interesting concept but they really haven't they stuck around but they have some uh V some various exploit points such as you like buying a token why would you buy a token for $100 if someone before you just bought it for like half a cent or whatever um bonding curves were interesting because you you would you could design the curves where if you buy a token you can buy it at certain price but if you want to sell it it will be priced differently uh we had Ico boom which a b with a bunch of like uh good projects and like B bunch of scams we have decentralized lending borrowing amm with their problems such as uh loss versus rebalancing just in time liquidity uh impermanent loss Etc and we have like security tokens uh idos rebas toen Etc what's the point of this slide is to Showcase how far we come uh and how far we can like there two place to innovate either through uh different like bonding curve schemas different amm for formulas etc etc and of course there's the new new paradigm such as risk taking there's new stuff like uh me is old but there's always some new opportunities for Innovation so I'm I told you that I'm going to be talking about tokens as a fundraising Tool uh what does it mean that means that tokens aable financing open source project in a way not seen before um what that means that basically I could launch a token completely anonymously and if my project is good or has some good Optics or marketing I could basically raise some funds through the through various types of uh raises uh we had icos which which give you more control but you need to have like a community that's actually going to believe in you uh so you don't do like a Logan poll and basically scam the people we had the idos with less control but more in line with web three uh web three uh principles which requires still some marketing some funding to like back the liquidity Etc we had I I iOS which are basically you launch on Exchange uh you have some deal with them they give you like marketing support they take in some fees uh but they can delist you at any moment as if you like break some break some promises or whatever and we have Fair launches which is basically you can create your own like sniper Bots where people can like uh dump their money and then you can uh create like a initial uh decentralized Dex offering and then you can like uh assign people their tokens uh based on based on their contributions to the sniping bot that you created or a pool and we have the initial nft offerings which are interesting for web three games um because they they're like okay let's launch some nfts that are limited um that give you some benefits in this EOS system but um it's a limited Supply and let's get some funding through that so these are the like the most often types of finances uh financing Fair launches of course there's a bunch of different ways you can launch your token that's fair but also there's a lot of like sniping etc etc uh I just wanted to uh when it comes to like fundraising the big topic is how long tokens should be invested so um it really depends um investing itself and like deals with VCS and like see investors Etc they have their own rules but if you look at this R this graph um this graph from outli Ventures think they did like a research around 100 100 or so web tree startups you can see that there is some if you look at the maximum values there's some like differences but there's like uh the mean and median sometimes there is some Trends uh similarities but it it's really individual so there's no like ideal resting duration there is no like ideal Cliffs if you look at this this graph um so some general tips um based from my experience uh from advising from teaching on the academy is to do like comparable company analysis what does that mean that's done in finance uh that means that if you're launching some token under certain valuation you need to do like a Toral research from of the market that you're competing in which it's of course you're going to do research be before you launch you're not going to just launch like an electric car and not look at Tesla uh but why is this important to focus on is because you can see what valuations uh the previous projects went went for uh what passed and what didn't how big were the rounds how big were the investing schedules uh etc etc and then you can when you go to the investor to the angel or whatever you can say okay I have a startup uh we do XY Z stuff better and we're going to give you a shorter investing period or we're going to give you a better offer for investing uh and of course I had like a friend that's launching launching his token and he wants everything optimized right away uh vesting schedule he wants to plan the taking rewards everything right away that doesn't really work because even though you prepare everything already and you know okay uh I'm my uh my Supply is planned for like next 5 years you can go to the like the VCS and investors and they're going to say I'm not going to AC accept to be invested for like two years uh whatever we reject the offer and of course the Private Sales is the most control you will ever have over the valuation of your token why because you're one of the one setting uh the valuation uh and you're basically approximating how how valuable your project is um and of course you have to give some discount because someone is early in your project they are taking on more risk and if you try and like apply some formula like discounting based on the cash flow if you look at the like the risk the risk of blockchain startup is pretty high it's hard to quantify um and it's it's it's okay for I talked to one of our the biggest uh I guess influencers in Serbia in web 3 and he told me basically it is okay and I agree with him it is okay for a product to have a shelf life to basically you don't have to design everything to last 10 years or your supply to like last for 10 years because if you're building a game realistically uh is it going to be played for 10 years unless you're like a League of Legends or whatever for web three it's not going to happen and you can always pivot such as like polygon did with their Matic and pole tokens and you can always like do some new stuff um Cliff or no Cliff this is like a an interesting debate because if you look at the vesting schedules the whole idea behind investing schedules is that tokens are released like slowly uh but what's the problem the problem is even though they are released slowly you cannot really control what the VCS or someone that holds a large portion of the tokens we do with them will them D will they dump them right away or will they dump them like slowly uh we had a like one interesting fact is we had a client that came in with for us like to like do some Consulting that had like a 24mon cliff and 24mth Westing which is basically you get no tokens and then at the end of two years you get everything um and that was written in the in the documentation and that's of course not correct they meant like theay or freeze um so delays in kpi adjust investing are a thing uh there's like a research I don't know if you guys are familiar but basically um the laser thing basically you don't there's two schools of thought one of them are like let's do Cliffs one of one of the the schools is like let's delays are better uh it's going to be the same they're going to dump uh at certain point anyway uh and kpi adjustment investing schedules are basically where you unlock this the tokens when certain kpis are met um that that's uh the research on that was done by outline Ventures I think uh as well so you can take a look at it kpi just investing uh and of course when you launch before you like you launch you need to have some general um supply model you need to know like okay will I have will my APR or my staking rewards will they be competitive with the market and initial market cap is basically when you launch a token on like an exchange you need to see Will the Market uh accept accept the valuation that you thought thought is the like the 100% price and basically you can see right away will the will the will the community accept it or not so it's a good signal but that stuff can be mitigated with Fair launches etc etc so before we go like further these are some general mechanism design simplified uh advice so um if you have any feedback loops you must set the bounce and stress test them otherwise don't launch don't do that don't do them because someone is going to exploit it uh at a certain point either through large capital or some plain old like basic mistakes that you made uh P controllers are a thing um they are like controller is used in Ry and they can like stabilize the system in some way you can research it a little bit out how they work there also a thing if you need to balance something in a decentralized way um study control theory control theory is basically dealing with feedback mechanisms um and governance should have control of the parameter values this is pretty basic um if you have something that has a feedback loop if something is getting exploited either have some sort of stop cles that's going to liquidate uh stuff or like just freeze it or the freezing is not the best but uh have the some control of over the parameter values and one of the stuff that we actually came in contact with is actually you cannot just copy the the mechanisms this is like a basic advice that I I I'm not even sure why I wrote it but this happened where the Lesser chain the smaller chain basically copied all the set on all the parameter values from the bigger chain which brought them and just scaled it 100 times Less in order to be cheaper what happened is basically they had a fee model which wasn't Dynamic which was uh the fees were different calculated differently everything was reacting way slower because they reduced the values they just copied it and that's a chain that has quite a large amount of users and it didn't work as intended at all treasury management uh uh this is not Financial advice but if you if you don't have any liquidity if you have smaller liquidity once you start selling your tokens if there's some crisis or whatever once you go past the market makers uh that are like fake demand you're basically you cannot just cash out your tokens and get the real if you hold like 100% of your own token in the treasury yes you you believe in the project but at the end of the day if you want to sell it to get some funding to do like to do whatever you will not you will just dump your price and every next token will be worthless uh buyback and burns are a Hot Topic pun intended um why they're interesting basically the idea is are they like Market manipulation are they or are they just signals that we have some positive cash flow or that we have some money to to to spend to like just burn the tokens burn the Monopoly money um that's a topic for discussion there's a lot of research done on that um one one school of course there's like one that's up for like buy back and burns uh the other one is like let's do buy back and make which means like basically buy back the tokens and invest like reinvest them for certain um initiatives um there's a lot of research on that uh I'm a big fan of like doing both of them basically like splitting like rewarding the users not just rewarding the whole ERS but actually using the BuyBacks to reward the the agents or in the system or participants that are helping your cause um but they're a Hot Topic pretty much everyone is doing it everyone has the deflation uh deflationary economics part in their docks um which which is like a um how do you call it it's like uh it's a sign of the times it's a sign of the like last year's bare Market or I guess today's bare Market uh what are some general tips uh mistakes as I said not doing compar comparable company analysis it's hard to do valuation because it's hard to valuate your token you're going to say okay we're worth as much as like Netflix whatever it's hard to do it uh but you still have to do it in some way um sticking out can kill the perception of the tokenomics that means if you're doing like even though your project has some makes sense for you to have like 40 or 50% for the team or some other like entity it's pretty it's pretty much you're going to get U you're going to get like YouTube videos of like dudes commenting on your tokenomics and stuff they're basically going to roast you and give their opinions it's not good Optics if you're dealing with corals in some way you need to do like value at risk and to like build accordingly uh not setting the value capture mechanisms that means that if I have if I do a certain Service uh for the tokens in this protocol imaginary protocol that I have which is like uber where I drive and I get the tokens if I don't have incentive to hold them for some benefit or whatever you are not capturing value and basically why you don't even need a token I I could basically get paid in like Fiat if I'm just going to sell them you get like uh speed of the E economics like velocity but that doesn't that's not not good staking for staking sense staking just to not sell the tokens um it's still like Up For Debate is the APR that you that youn is it like is it worth or not um my advice my general advice or what we've been trying to do is to have some benefits for staking not just to get rewards but try and reward like longer staking like GMX does with these Crow tokens or whatever and of course focusing on deflation this a big thing for the last for the last year or so everyone was like let's do like a revision of the tokenomics let's add some inflationary uh properties um let's slow down things etc etc but what's the problem with this is if you burn all the tokens and if you like have a like uh incentives or disincentives that incentivize burning or like slowing down the economy you will have no tokens to reward um the participants with or you're going to have you're going to reward them with some so little tokens that they won't have they won't be incentivized to even participate in the protocol that's up for debate for like Bitcoin what happens if the rewards are like half Hal Hal Haled uh does it even make sense and if you want to like do some research why deflation is not always good you should take a look at Japan and their problem with deflation uh and abenomics which is basically they had people they slowed basically the economy because people were hoping the stuff is going to get cheaper and it was getting cheaper and they were like disincentivized to spend and when people don't spend basically the economy is slowing down and the stuff like the economy is dying um and then you have to do like reflation or like abenomics where you do quantitative easing where you uh like give some better interest rates Etc um and not stress test what we do here in MVP is basically we try to generate agent based models which basically you have agents that are participating as a uh host hostiles you have agents that are like uh still participating even though they're not uh really financially incentivized to do so and you you can get generate a bunch of different agents based on the type of your protocol like arbitrageur or whatever uh not stress testing is quite an important topic because you never know what's going to happen you have to tackle your project as an to like model it as an enemy and to see um what what's going to happen you cannot predict the future but we can at least like test some stuff out so what is our general approach um I I think I'm if I'm taking too long you can like wave and I'm going to be faster um this is our general approach uh we have the initial phase when we do benchmarking uh like brainstorming Discovery uh when we set like system goals when you do like competitor tables we have the design phase where you do like the initial allocation when you try when you try to sell the tokens for the investors and when you take like legal considerations afterwards then it's time for optimizing because uh if you like take if you like design everything optimal or whatever like token engineering can be a slow process uh and it can cost so if you like didn't get some funding or whatever um I'm not sure if you even need to like do some formal like uh final modeling to like run various testing Etc um it depends maybe the investor will look at okay this is the best economics model but occasionally it's not the case uh most of most occasionally um and validation phase why is important because we might launch something some new chain that on in theory we haven't tested it we tested it but we didn't test every every scenario in theory it doesn't work as intended what we do we create like a p python model we do most of the stuff like all of the stuff in red CAD we don't do like minations or like game design tools at all um you create like a python model where you get you can get the the data from the actual chain and then you back test it to see like if your model is corresponding is behaving similar or the same as the real system and then you can propose changes and iterate try and see what what's the best U what's the best way to do like to innovate to add new features or change parameter values and of course after you launch it you basically uh do monitoring and then propose new changes if something's happening these are the tools that are really interesting we use red cat Arbiter is the is like for me is a new one uh it's written in Rust and of course agent butter cup they mostly focused on like agent based modeling for the evm um and one problem comes is um how do we actually like if you have no data when you do modeling is it just like a horoscope is it like a like uh are we just like putting things randomly and then okay let's ship it um even though we have no data basically is we try and like do the parameter selection under uncertainty which is like a methodology from the block science team they've been in like system modeling for like seven or eight years so we try and like apply their their their approach to actually designing these complex Dynamic systems uh what does that mean um basically what we're trying to do uh when we are designing the token the tokenomics mechanisms testing them out is to actually like Define the system goals to try and quantify them even though like if you look at like um a lot of Economics uh there's like a utility function and there's a function of utility uh like from like apples to oranges where you can have like I like oranges more but I don't like 100 oranges like I would like to have half and half um it's I didn't explain it correctly but the actual quantifying of the like qualitative goals or quantitative stuff is quite often in the economics some of them critique it um it's it's up for discussion economics are still like it's a social science um sorry okay um so uh we tried to like identify the system goals and we try to quantify them uh you can look at the gaunlet work on the Unis swap where they try to like quantify uh what kind of incentive why when it is right to to like propos liquidity Mining and they basically created a function that says okay if this um if this conditions are met uh then the this kind of uh system this kind of mechanism or reward schema is good so you try and like create like a quantitative goal control parameters identification is basically to see what kind of parameters are controlling certain goals what's their impact um for example like fee rate minting rate if you if you have like a fee structure you might have a parameter that's like scaling the fees depending on the congestion then you can play around with it and see does this parameter help your goals or not and of course if you're building on ethereum or you're building general in general um you're we are not building building in a sandbox that means that you need to like take in uh for example price of ethereum and then try and model of course you cannot just predict the price but you can always generate scenarios for prices using like Brownian motion which either positive drift or negative drift or just plain old like Mark of stuff where you don't you cannot like the past value is not influencing the the next value um and of course like you identify the metrics that you're trying to track like apis when you run the simulations and then you can see you run them you see what like what combination yielded what uh what values of the kpi and you try to um try to optimize so you need to see you will see like what are the differences between these goals for certain like parameter combinations as I said take a look at the unop objective function it's a forum post made by the gaunlet the this picture like above is basically what's the basic basing of Attraction you basically see what kind of uh parameter combination is yielding what results and then you just iterate and if something is taking your controller or taking your system basically destroying it um you just like remove this kind of this type of combination or you iterate if you look at the parameter sweep which is like an often oftenly used method you have like a bunch of parameter values you have like um bunch of kpis uh you have like multiple like runs monar runs and that can scale with like that creates a lot of uh simulation results and then you can do some like analysis and of course we're not like we cannot predict the future correctly but we can see maybe some some points some feedback loops are present that we don't we weren't aware of uh under certain conditions so as I'm the only one I think I'm going to be talking about economic auditing as well we did that for like one big client um what's what was the rationale they basically built the stuff that's optimized for growth that was um in basically inflationary as hell um that was incentivizing users to just St to not participate in the network and basically they were they had a bunch of problems with the actual like model itself how the fees are calculated how the VMS calculate the fees Etc so we had to fix it fix it changes are not uncommon that happens a lot of products are just revising their toomics a lot of proposals are coming so it's not the stuff is not setting Stone um B there's bigger risks when you do like toomics auditing because if we scale the fees 100 times or whatever they do some new like change we can basically kill defi in it or kill gaming or kill a whole ecosystem so it needs to be tack tackled the uh how do you call with with caution what are the like how it looks in general what is good when you do talk audits uh when you fix the economics is you basically have a lot of data you have you can like do some you can see some Trends either through in like uh like fee usage fee like factors movement Etc uh when you have the data it's way easier to model to back test it um is basically we get the data we create the model that basically represent either a whole chain in some way or it represents a certain portion that we're revising and then we back test it see if the inputs are uh if the like the same results are happening with our model because if you do back testing and it behaves completely differently for the same inputs you created a model that it it's not it's not a digital twin it doesn't behave the same um and of course we try to like do like recommendation for recommendations for implementation and launch because you cannot just dump and both technically and like it doesn't make sense to just like dump a bunch of new features we try to like plan with the client uh how to actually uh like introduce new changes if there are some interconnectiveness for example like inflation rate is depending on certain like program uh or whatever if you launch a new inflation rate that can basically uh print out more tokens than expected because it was La launched before the stuff it's launch it was launched before the whole the whole mechanism or the whole thing uh these are some materials I'm going to share the like the link uh in the telegram and stuff like that if you want to do like System Dynamics modeling um take a look like at the cad CAD uh models Arbiter but cup qtm etc etc so these are the materials and conclusion is basically um if you don't have data when you're designing a new Dex or whatever you can use the data from the existing models you can see like trades routes of Trades uh how much of the market share can you take over how competitive you are with aprs with liquidity mining rewards um you need to plan the supply emissions and distribution when you're launching uh manage your treasury if there are feedback cloops stress test them otherwise don't use them um deflation is not always a solution as I said take a look at Japan um back test the models so you're sure that they are working as intended and you cannot like we have always in in company they ask us are can we really like predict the future or our our is um our um are they correct the models or not is basically you cannot predict everything but we can test some certain scenarios some certain behaviors and then we can try and like mitigate certain mistakes that we haven't seen in our design so uh that's it for me uh regarding like the presentation um I guess Q&A we have some time for yeah