 Yeah, so From the last talk we get we arrive at this slide and having a test for non-MEV needs But in this talk I'm done talk about, you know, sorry, but this formulation of MEV is mid So specifically what it means Why because it's like a party guy, you know a guy at the corner of a party thing They don't know this is not MEV while there are people all of us Yes, we know and we don't care if it's a good name and it's You know memes are important memes are more important than you know arguably many other things in our life And then but there's also another Extreme in the community are vouching for a formalization of MEV, right? So some common pitfalls on this On the world on the definition of the word MEV is that we give up on the clarity of the notion We over formalize it such that we lose the memes Right, it is no longer a meme term and is lost in and we are getting lost in the detail of the real value of defining MEV So in this talk, I will give a clear definition of MEV that points to a solution two points clearly to the solution recipes that how we should solve MEV and I'm not gonna give a meat curve Formalization nor am I abandoning our memes So what is MEV? So by the end of the talk, you will see that MEV is Al Capone evil deity and lose the 16th So this is MEV Yeah, so so so to Start we see that so so so why Al Capone the mafia is MEV So because mafia extradible value, you know, CCM arises from arises when one agent or a collision of agents gains an asymmetric knowledge of another agent's private information So this condition sounds there exists an asymmetrical sophistication between agents. So Very easy examples include sandwich internalized front-running or in traditional finance We see that people trading one strategy is that people trading on the limit the other books in balance So which is essentially a form of one agent gaining knowledge of another agent's utility or intent Right when you see more bid than ask on the books and you take the best ask and you make the best bid So then evil deity is a malloc. Why is it MEV? So malloc is tradable value is the value that was surrendered to the malloc I on coordination within your mechanism your allocation mechanism for people's preferences So one example of this is the negative externalities caused by inexpressive mechanisms So for example, suppose you have a random ordering then people Spam the network and and that is the only way they can express their Preferences and it is very one-dimensional way, right? So you have a kind of a shrink transaction quality trilemma and there are of course many other examples for so for example in the traditional high-frequency trading arms race example People the mark the cost that the market maker spans on arms race gaining latency advantages was channeled as Charging higher spreads for users. So the externality is Transferred from sophisticated agents to unsophisticated agents. There are some other examples And so finally Louis the 16th the monarch wise him MEV So monarch is tradable value arises from the fact that the coordinator So for example the sequencer or the validator has the ultimate power of deciding what the ordering or the allocation of Specification on state is so by specification on state I specifically mean which specification or property doesn't next day satisfy you can for some of you familiar with program languages you can see that this directly links into the You know formal application you have a program and the specification that the next day must satisfy So then some examples of this include that the validators currently, you know In ethereal other chains a crude value because they have the power of determining what content is in the block and many other examples for Intrat 5 such as colocation fees and cross-chain such as cross-domain Marking makers for swaps So then we can see that MEV arises from three distinct properties the mafia the monologue and the monarch So now a little bit more formalized Suppose that I miss the monarchy our unions or our game E and W is the social welfare function Then we know that the monarch is tradable value equals to the welfare that is Achievable by the best possible mechanism for our game to allocate specification on state minus the welfare of the current Monarch and the monarch TV equals to the welfare of the current monarch minus the welfare that that already existed, you know In the original game it without our current monarch has a coordinator So you can see those closely tied to the idea of externalities, right? And then from our definition we know immediately that Monarchy v plus monologue extradable value equals the welfare of the best possible mechanism in our In our game minus the welfare of the original game. So so it equals the constant, right? So it is so the monologue UV plus monarchy V is innate to the game E and does not depend our choice our choice of the mechanism M And finally Defining mafia UV is a bit hard but harder but the intuition is that mafia UV depends on the specific Sophistication details or the power of agents that is existing in M So what this means is that how much exposed knowledge or last look? One agent can have over other agents or their coalitions So then we can see that those three values have very have distinct sources and those sources of value are non-overlapping Thus we can cause them, you know our new definition collectively three EV mafia monarch monarchy extradable value You know it forms a some type is perfection So then from what this implies is that There's solution recipe rights future is in our hands We know that mafia plus monologue plus monarchy is a hundred percent of the MEV and that monarchy V plus monarchy V equals The constant depending only on the original game But the percentage of monologue versus monarchy depends on M our mechanism our choice of the mechanism so You can already see that there is a nice link with the notional price of energy in traditional mechanism design here, right? Although in traditional mechanism line, of course, like the you know, they are seriously want bonds So they have so these they define price of energy as like the diffusion But now we are interested in value. So, you know, I kind of define the difference and then Finally, we can see that mafia EV is also a constant depending our choice of M So since the sources are distinct We can adjust the percentage of the three variants of my maybe by transforming one form of M Maybe into another but in our designing of a better mechanism M So I argue that ideally we should have 0% mafia 0% malloc and 100% Monarch where the Monarch's profits are Distributed so what this specifically means and why is this true is that because Mafia relies on the existence that there is one ex post knowledge And there's on sophistication or sophistication between agents, right? There isn't a symmetry and so specifically once you have this then These sophisticated agents are going to charge sort of unsophistication Tax on the other agents and says this hurts long-term stability of the domain or long-term prosperity And then why we want 0% malloc is because any value that or any weapon that is attributed to the malloc is forever lost It's not recoverable and nor is it redistributable So then and why 100% Monarch was redistribution is that if there is no redistribution to the Monarch then Then the 100% Monarch is gonna erode the incentive for people to use your mechanism, right because You're essentially creating Reintroducing the value that they would have paid to the malloc by having them pay to the Monarch instead So there's a different you can see there's So users would be indifferent between paying to the malloc in the original game and paying to the Monarch in the meta game So, you know, this kind of violates the individual rationality guarantees So then there are of course some caveats as to our choice of Monarch and our definition So as we can see that the welfare of the Game without the mechanism depends on you, right? So now we assume that the mechanism and Represents the domain or a builder or some just some coalition of agents then we can give a kind of and have a notion of Interdomain correlation factor k where k equals one if the mechanism is Monopolistic across all domains. So here we are kind of in cheating a bit We are including time as another domain So k is kind of like a discounted and dimensional preference curve and Example of the preference curve you can see on the top right corner and then K equals zero if the mechanism is public s So then what this implies is that by forming a larger coalition means the coalition has a higher K And this enables them to do more credible extortions So what this means is that the Monarch is rather value for your coalition of agents, you know Informing the mechanism M equals to K Multiply by the original Monarch if M and Madison forms the grand coalition So then of course this implies the users can also be the Monarch if they have a large enough K, right? And similar for validators searchers or builders you can imagine different mechanisms than what is currently being played in the MEV field. So Essentially the interdomain coalition factor represents your collective bargaining power which determines how the Monarch is set up, who we thrown as the king and How the Monarch is set up within the game So then however Okay, so so There are some caveats to the choice, you know, how much coalitions can you have right? So you can imagine this formalized as a corporate game where the core of the game might not exist if we have too many interdomain coalitions, so This kind of reminds us as the you know original notion of there exists a non-monotonicity of coordination and Efficiency and this also relates back to you know a notion. I would like to call the incompleteness theorem of the MEV right you So you can see that The game E cannot continue if there is not enough competition What this means is that you can increase the coordination But it does not always increase the Efficiency of the allocation in reality and even if you try to eliminate the Efficiencies caused by your extra coordination by using some another credible commitment or Coordination device in the meta game that you're just transferring the monologue its credible value in the original game to the Monarch In the wrapped game, right? So this kind of provides a ground for us to refute any, you know previous ideas by Virgil Griffith and his hope Ethereum as game-changing technology argument. So then the Coming back to this the specific value of the correlation factor K depends on how we set up the game We can throw or destroy different Monarchs Ultimately, we want to choose the best mechanism and start at its most incentive align with the long-term prosperity and the most capable of coordination, so you can see there's a trade-off over here and Of course, like why do we want a monar? Why do we need this meme, right? Why can't we just go the formalization route, you know for formalization, throwing mass equations? So the reason for this monarchy abstraction is because the outcome in contrary to traditional Mechanisms the outcome in MEV is not common knowledge and depends on each agent's private information, which the other agents don't have prior Don't have prior to the MEV time, which is the allocation time or like lobbying time So it is hard to analyze the equilibria within this game, especially when there is exists an Assymmetry in the sophistication of agents. So thus we reduce the guarantee of our mechanism to just have some some bounce like some some loose bounce on the incentive compatibility and the individual rationale So then not to concretize things we provide some case studies So specifically we say that there's We know that there has been vanilla first come first serve and a frequent batch auction style first come first serve where we increase the fairness granularity within the Receive other fairness protocol So what this specifically means is that instead of reporting a strict ordering by the receive other time of the transaction each individual node now reports a partial ordering to the leader of the Ordering consensus protocol. So what this concrete means is that you know, you just take the smaller than to smaller or equal to And then after Aggregating each individual nodes preference the Monarch in this case the leader of a consensus protocol gets a weak ordering by first come first serve And then there are of course, we know there there must exist some other batches in it Which is caused both by the condorsed paradoxes and the initial partial order reported by each individual node So now the Monarch tries to resolve the order of the order batches using auctions, right? So this is why it's frequent batch auction and then in reality. Why is this more? They are either like let's pick into some data. We see that In reality it's over 65 percent of the unit star volume is come comes from arbitrage It's you know, maybe it's gone and among the other 35 percent. It's a big Few few market makers. So what this means is that there may be activity constitutes a huge amount of Ethereum activity and the birth period on Ethereum which arises from public information reviews So for example, some price discoveries happen some violence or some people Some well send the unprotected transaction into the public mempo then once once this this kind of event happens it takes around 1.2 second and And 75 percent of the conflict of preferences on Ethereum Complete preferences meaning that there is a conflict within what specification should the next day satisfy? You know, it happens less than four percent of the time and so this means that vanilla first come first serve compare with Frequent batch auction first come first serve will be bearing the extra negative Eternalities which cost worse UX higher fees for users or centralization the nodes from 12 acts more conflict of preferences so, you know burst of periods for preferences are significant and Just adding some partial ordering just covering that little four percent window. You can Mitigate most of the negative externalities on the domain. You can take value back from the malloc Right. So if we use the three eb framework to compare those two protocols, we can see that They have the same mafia, right? and and for mallocs the First the vanilla version has much more malloc because The money my vanilla version nobody can express their preferences in a short burst period which constitutes most of the activity, right? So so the malloc arises from the tribute to the malloc arises from uncoordination and an ultra simplify social choice function Right where the social choice function is like I all like I choose the preference by the order I receive them that does not coordinate any of the burst period MEV Right and the burst period MEV, you know typical cases include like permission less. I mean you price discovery at one Can I maybe so and then so so you can see from the graph on the right, right? We are transforming the mafia slowly into the malloc Where the where the value is transferred to the users with the users have the same payoff You know same bad payoff and then from the monarch is strictly worse, right? But by definition, but also Vanilla versions run the wrong monarch, right? So in in latency games you pay more to AWS or Google Cloud because you know you need to play this latency games And thus the monarch, you know Google and Amazon are obviously not incentive aligned with crypto or like whichever domain you adopt your Ordering for the call on and of course it's more centralized because come on. It's like fan Okay, so now Some some some concrete examples. Let's see, you know transformation of 3d. That's really actually capture most of the Existing MEV notions that we have so for transferring malloc to monarch We see that one form of this is the spread charge but charged by market maker bridges From monarch to mafia We see that if in traditionally if in previous Ethereum proof of work the miners were to go rogue and still bound us They need to be, you know, they were the monarch But they choose to gain an advantage using the asymmetric information ability to act on that information Wait, what okay? And then from mafia to malloc the transformation So for example, you go from a public man for model to a vanilla first come first serve, right? You eliminated public man for so you don't have that information instrumentry, but you have that, you know On coordination you have a very bad allocation of preferences Which increases malloc and then from malloc to mafia You see that so for example if a malicious coordinator wore to pocket all of the surplus welfare that that it generates from Optimizing the ordering or just existing, you know from coordination. So this case You you you you go from? Yeah, my luck to mafia and then what we can do to go from Mafia to monarch is to have programmable privacy, right? So in this case, for example as yes, you can still do found a merging of block building or preference segregation despite that You don't have a symmetry a symmetry in the information and then finally The current model of dog governance is a form of Transforming the monarch to the malloc, which is not good, right? So So so what this means is that so dog exists for coordination, right? But then by launching a dog token and then allowing it to distribute under a bad tokenomics or like whatever your your creating upon the which creates more malloc, right? You could have Got that to a monarch stage, but then through more decentralization You have recreated the malloc in the metagame of the token game, right? So this reminds us of the previous example of the monotonic Relation between decentralization and coordination and efficiency in the mechanism So there's some implications of three EV right concrete has this idea. What why is it useful? So then the solution at recipes how to achieve 0% 0% and distribute to 100% We can solve Mafia EV using programmable privacy, which allows for expressivity and agents can control how their information is used Along the path of the determination of the outcome by the monarch Also, this relates to the traditional mechanism designs idea of the definition of how what how expressive a mechanism is, right? And then concretely what this means is that you can devise a mechanism for discovering the unsophisticated users preferences and helping them to communicate this preference to the coordinator or to the mechanism and then we can solve Moloch is thread of value by just increasing efficiency, right? We can eliminate the price of energy wire via refinement of society or IE specialization of labor. So what you can see, you know In this example, it's like contrary to the idea of molecule has the idea of a slack Right, so slack is arises from coordination with is basically you you allow somebody to have enough Mote or like period where they can grind on the higher payoff thing We and this is only allowed if you have a highly coordinated, you know specialization of labor in your society And then finally monarch is credible value. How do we division is value, right? So the value division we should division it in such a way such in that we Maximize is the welfare or the future returns. So what this means is that you probably shouldn't focus on giving tiny Kickbacks to the existing 1% of users, but instead focusing on onboarding the invisible 99% of users as As Monarch is credible value or like I mean each in general grows super linearly with respect of Compare with the number of other flows that's existent on your domain, right? So for example, the ways you can do this is through investment into wallets or some some idea like ritual active public goods funding And of course, this was the original idea by Faraday as well Right, like the idea is that you can there are 99% of users that would have come to your domain if you other things you know fairly but Yeah, yeah, so so it's it's been on as a narrative and then Yeah, and so so then Specific implementations we know that 3 ev was defined by the source of value, right? They are distinct so the sink however the sink of those value are different So for example the mafia extra value sometimes thinks into into Monarch, right? So for example when there is a sandwich and then two very competent atomic searchers No value things to the Monarch, which is a validators and then the sink of value Impacts the source of value. So I argue that by controlling the sink of the value. We are controlled We can already we can also impact on the source of the value by the magic of credible commitments So some example credible commitments I'm not gonna expand but you can like hit me up later after this talk examples include builders as roll up or if they you have a special Builder innovation where you have you grind on the Expressivity of the building language or you have a decentralization of building use SGX tokens to align the incentives or like some more fancy things But all of those conditions on all of those solutions conditions are low enough interdomain coalition factor, right? remember from before once you have too much coalition or too much Coordination then you can't possibly you know have a good mechanism or like it's very hard to do so so And that a low a low enough factor means that you have high enough competition, which means the market is contestable. So I Claim that having more competition in the game by incentivizing having more competition in the game it's right curve behavior and Optimizing the ordering by learning more about the preference microstructures and then grinding on the design of the mechanism is mid curve. So Contextability is an important tool and it can only be realized if we design the initial market in such a way that it is not Centralizing what this means is that In reality what it means is that there should not like look like we should definitely Device the mechanism that were exclusive or exclusive other flow within like this provides some philosophical foundations for the previous previous talk Right. I said it's me now. It's time to you know show some real work for show some real like test for non-MVN is So then remember that Fraga and Russell was debating why you know whether the king of French was both right In the beginning of 20th century and they come up with the tests for presuppositions For utterances of sentences. So for example the utterance of the sentence axis MEV has a presupposition Test that there must exist a game E and an allocation order in protocol And and that within and then the using the three eV definitions We can indeed see that and it's constituted by a type or multiple types of the three eV So in so so in this sense the utterance of X is MEV has a presupposition of the existence condition of E and M and X being the in the set of three eV Thus if there does not exist such a E and M Which you know translates into English is that is if that is absence of a monarch Then the utterance of the sentence axis MEV is false and that's the axis not MEV So then some a concrete example imagine Kim and Don playing a Prisoner's dilemma, right? So then they both choose to betray now suppose there isn't a coordinator. Okay So then somebody uttered the sentence He says that the value that Kim and Don is extracting absent a monarch i.e. One one is not MEV then this then this value is indeed not MEV by passing the presupposition test But if we take a free game view And then of course like there are I can't take them as I like Wittgenstein has this Notion of a language game where the context depends on social things So the utterance of axis MEV So this sentence might be true as the context of the utterance is social and we might agree on some notion of the game So if the sentence is provided without context then the sentence is nonsense and has no choose value So ultimately the test of MEV conditions on which philosophical interpretation of language you agree with If you are Wittgenstein, you know, you always wish Wittgenstein and believes that the utterance is within the context Where it is socially agreeable that a monarch is this so in the previous rugby case remember We often see let's say that physics is the monarch and then over there is a value is indeed MEV Okay, so so so now some takeaways let's change the language a game are too tired about minor or maximal Where the grammar isn't even correct like like we say maximum MEV all the time And we should just say MEV and it conveniently represents a shorthand for the some type of Mafia monarch monarchy straddle value, right? Have you been saying MEV because you have poor imagination to come up with the replacement term for minors that keeps the M and makes sense You know, is that what we are doing, you know So then the argument is that by changing this language game. We have several benefits. We have You know this formulation gives you a distinct value which Origins from the source and then it is more formalizable and it is formalizable in the sense that you don't get lost in the details it is still keeps the meme and the formalization and the proof only serves as a social process to help people to understand it and then of course This settles the bad MEV argument and the trade-off or the transformation between the three value space is clear Well, this implies also this implies as clear solution recipe So some nurse sniping for you all they said to research on how monarchy straddle value can be distributed and the research on the program of privacy solutions for mafia EV and then finally a presupposition test and so to Conclude I would like to hypnotize you 3v 3v 3v 3v 3v 3v 3v 3v That's all you can hit me up Well, I think we all agreed that was a fantastic talk a lot of like the two speakers We we said at the beginning that maybe we have time for questions So let's pick a couple of them or there is going to be hard for me to get there. So maybe you can shout out loud Yeah, good question. So all facts censorship Can be done implicitly using a term of using a form of the Preference cars, but a little maybe It is Monarch Yeah, but as I said like what we want is to shift the structure of the game Right, remember the interdomain coalition factor You can't unify the users by devising some mechanism to help them become Sophisticated and have collective bargaining power. So that is what we ultimately want to do. Yeah Anyone else? All right guys. So big round of applause