The New Era of Transaction Observability: Deciphering MEV's Shifting Value Landscape | Ye Wang

you for joining us here at the Trident stage of eth Denver 2024 I'd love to welcome our next speaker her name is y w Wong I'm sorry yeah Wong uh she's a data scientist at aneni and the talk today is titled the new era of transaction observability ability sorry yeah yeah it's fine so um hi everyone I'm Year from the data team at the igoni so today I'm going to share with you the recent findings from our team on the battle of backgrounding the private orders um and how the transaction observability on ethereum may impact the um distribution of me Revenue in the whole supply chain um for those who may not be familiar with the idea of private orders you may probably hear of the following products here that provid M Production Services like the flashboard products and like um map blockers so um so what the idea of private orders so basically on etherum the user transactions May face the uh risk of me so um these types of products they may provide a services for users to choose to um to choose to submit their transactions directly to the builders's private RPC uh so this means that these user transactions may not be seen from the public and then this can prevent those user transactions from from running um me and currently the data from flashboards shows that um every day about 10% to 20% of the um all the user transactions on etherum are sent to um the Builder's private M po which means these parts of user transactions cannot be seen by the public um so when we have a you know basic understanding of uh what our private out flow are so we can go through about the several questions that you may hardly uh think of before so first of all when you are imagine that when you are trying to submit your transaction to the private out flows do you know the exactly how much value that can be created from your privately submitted transaction um and also if if I told you that there's some me opportunities that can be found inside your privately submitted transactions so who can be you know um powerful and who can be um privileged to take advantage of these private order orders and to try to make profits from backgrounding these private order flows and consequently in the supply chain whose Revenue was taken from during this process and um for the last question is that so if um there there are only a small parts of players in the uh whole supply chain that can utilize this um private order flows and try to make profits from backgrounding these private orders um what will will um have impact on the whole me ecosystem and will we go to like a world towards me Utopia or MV dystopia So within these questions in mind so let's start from the um idea of so-called transaction observability so it derives from the idea of the public transactions and private orders so basically if a user want to submit their transactions um to um to the builders um without any other um settings specific settings on ethereum the transactions May probably enter the public man poool so in the public man pool it means that everyone can see the transaction in public man pool as well as for the bu uh Searchers so if the Searchers they find that um there's are some me opportunities that are in the users transaction in that uh public man pool so they can those Searchers can take actions right they can try to backround or front round your transaction so for the you uh for the the users they may face the risk from front running so front running might be um harmful because the Searchers they can submit um a transaction before the user transaction so the the users have to pay more about their Swap and at a higher slipage rate um and for the backgrounding opportunities this may be um a little bit tricky for users because sometimes the user may even not recognize that their transactions can create such a huge huge um me opportunities for Searchers to make profits so um for example if a user has made a swap on a certain liquidity pool that create a backgrounding opportunity for Searcher um this Revenue this revenue from MV cannot you know um U Get Back PID back to the users so users cannot get any uh me kit back from the back uh backgrounding transactions and imagine that if the User submitted the transactions to the private channels so things will change um because the users can choose to um send their transactions um to uh several Builders specifically so only the selected Builders can see their privately submitted transactions which means you know Searchers cannot see this private orders so this can um pre uh protect the user transactions from front running me and as for the back running me um there are some uh other flow auction mechanism also provides the users an option that you can uh submit your transactions to the orderflow auction clients and the The Searchers can beat for the backgrounding opportunities from the users orderflow and then if the uh there's any um M Revenue that can be extracted from backgrounding transactions the users can also benefit from this backgrounding transaction which means they may get some of the part of the me revenue from the backgrounding so let me explain uh by detail so basically we have um three types of private order flows and and these private auto flows are exclusively visible to the builders the first one is that what I uh mentioned about uh is the me protection RPC so basically these user uh user orders are sent directly to the Builder RPC and the second type is the orderflow auction um RPC so these types in for uh via these types of channels the user transaction are submitted to the clients and be and Searchers can beat for that user transactions so once the Searcher included the user transaction and they can redistribute the user transaction in a searer bundle and send to the builders RPC and the third type is basically the searer bundles it's like um um because the M researchers they relies on a bundle mechanism so once they find some there are some me opportunities that hidden in the user transaction they can choose to include the user transaction in their search bundles and then send this bundle directly to the builders RPC so if when we get a basic idea of all the private order flows we we can call back to the first question so um do you know exactly the value of your private order when you sending your um transaction to the private order flows so we find that actually there's a large cake for backgrounding um those private autof flows and this parts of me value has been largely ignored in the past as we have observed that um there are more and more m boss they are um shifting their competitive Focus from Bing for the um transaction orders in a block to the um compete for the transaction observability however we found that although this um from backgrounding this private all flows that can can be huge me potentials um these backgrounding opportunities are uh found exclusively available for a very limit amounts of players in the market so how it U happened and how the private others create exclusive me me opportunities for Searchers like if for for the Searchers who want to background us the transaction um in a public man pool so they may face intense competition to bid for the top of the next block opportunity so they have to pay High Bride payment to the builders and then these parts of value will flow to the validators in the end and for the Searchers who can um have the privilege to background those transaction via private channels they because there are only a limited players can see these private transactions so um backgrounding opportunities are um less competitive and only a few of the players are you know trying to um backgrounding these private transactions so they may not they they um these Builders these privileged bu uh Searchers may not uh have to pay a such large amount of the like the gas fee or cost for biding this um of me opportunities so um for those privileged Searchers their costs are pretty low and I will explain this into um based on on these several uh types of private flows um because that the user orders send via protection RPC are um are of limited disclosure of the data data sources so we will mainly focus on the uh last two types the first type is the user order send via orderflow auction RPC and we will see an example from um of the backgrounding private order flow via order flow auctions so he uh here's a um a list of transactions in a block and from uh for the oh sorry for the position from position three to position five it was a me share bundle so it includes three transactions basically the first one is user Swap and the second one is the orderflow auction Searchers backgrounding of the user transaction and the third transaction is a uh Builder refund so the Builder transfers parts of the me value to the user in this whole bundle and we found that interesting yeah something interesting that at the position 10 it was another background targeting this m share bundle so it's crazy because the M share bundle is privately submitted right but this um but this transaction are targeting the user transaction in this map share bundle and let's see how it happened so first of all basically in this map share bundle we can see the three transactions the first in the first transaction the user tries to swap e for allet based on a Unis swap V2 pool as you can see this um this pool highlighted in Orange and it and so the user uh swapped we for allet and then the Searcher in this UTF flow aution mechanism send a um backgrounding transaction targeting the user transaction which made a reverse swap in this same pool in this Unis swap V2 pool so the basically this backgrounding transaction are swapping allet to we and then in the third transaction in this bundle um the builder then directly transfer the e to the to the user so this is a builder rebate of uh parts of the that shares part of the uh me value value um to the bu uh to the search uh to the user so this is typically a uh what a map share bundle uh uh works like and then let's look at the backgrounding targeting this prival lied bundle so for the orderflow auction Searchers backgrounding transaction as we can we uh talked uh previously it made a reverse swap targeting the user of the user transaction right from Ali uh Ali to we within this Unis swap V2 pool and for the backgrounding this private uh private uh orders U we can see that the M bot also took advantage of this Unis swap V2 pool it the same directions well just like the other flow auction Searcher so it also backgrounds the user transaction in that map share bundle but you know for the UT flow auction Searcher they have to you know pay parts of its uh me Revenue back to the user but for this backgrounding of the private order flow this Searcher here all of the me Revenue are toen by the Searcher so you may ask that you may have question that that um how to make sure that whether this backgrounding transaction targeting this bundle are really targeting this private order will it be like um will it will this Searcher see any onchain signals previously and then make a background so we can do a very simple check by using the um transaction simulation so we can see that we we um we can see that um on the position 10 as what this transaction happen it will be successfully executed right but if we position this backgrounding transaction in front of the privately submitted bundle what will happen so if we position this transaction at position two yeah it will be um reverted so it will returns a a reverted fail message so we can say we can say that um before the that M share bundle there's no Arbitrage spaces for those me bought to make profits and this um backgrounding transaction is exactly targeting this um privately submitted bundle and here's an another example targeting the Searcher bundles specifically a targeting a sandwich attack bundle so here's also a uh sandwich attack bundle in the block so it will be will be from position zero to position three it in also includes three transactions right for the first one is the front running and the um and also the second one is The Vic transaction and third one would be a backgrounding transaction targeting the victim's transaction so it's a very typical Sandage bundle and then soon after this bundle we found that there's also a backgrounding transaction targeting this sandwich bundle so let me explain this in in in a uh in a simple way so let's only took at the backgrounding transaction from that sandwich attacker so we can say that we can see that that attacker utilized the pool of this usdt and wbtc curve pool to make profit and let's see what that um private uh backgrounding transaction of this private order flow uh have done so basically the M bot also took advantage of the same curve pool it may also made a same swap just like the sandwich attacker do so so in this way this like um Searchers can took advantage of those prival submitted bundles or private transactions to make profit without paying a High um cost to for the uh Builder bribe and to the downstream so it's more important to think about who is privileged to background to background this privately privately submitted order flows and um in the supply chain whose Revenue was taken so we can see that um basically we found that the these Searchers are privileged because they have um um collaborated with the builders because only Builders who receive those private order flows can have the like the privilege to knowing which ERS will be appear on in the next block so the those integrated vertically integrated Builder Searchers can take advantage of this transaction observability and make profit from backgrounding these private order flows and then they can they can when backgrounding these private order flows they need not to pay High um Bri Bri pay and high gas fee to the downstream and then so they can in this way they can reclaim some M revenues from validators in the downstream so as we have investigated from October to November in 2023 we found that in the um in the whole arbitrate Market there are seven over seven7 million in Revenue were generated and over um over 18% are generated from backgrounding the uh back uh backgrounding the private orders and more impressively we found that if we break down the revenue structure we may find that backgrounding private is super cheap because only 30% of the whole uh Revenue were paid for um of pay as cost and for basically for other types of Arbitrage Searchers usually pay over 90% uh from their whole Revenue as as M cost so in this way we may say say that um for backgrounding those private orders some of the um profits that would would be uh flow to the vators are taken from these vertically integrated Builder Searchers and then let's come to the um last question so what will we uh what will happen if this um transaction observability as a competitive Advantage continues so we we we found that those vertically integrated build Searchers may take advant Vantage of the transaction observability to you know um background private transactions without paying a higher fee and then get some of the value from validators and also they may not like take um follow the same rule as those other flow auction Searchers and they may need to they need not to pay back me Revenue to users and also it will redistribute redistribute the me Revenue throughout the whole supply chain as the validators are getting less than past and also so if this um circum uh circumstance continues we may face the threat that uh are falling into a um what Stephanie caus a so-called am dystopia where a few amounts of players in the market may take most of the meev revenue so and they can take those transaction observability as competitive advantage and set higher entry barriers so I think it's time for us to reconsider this Con um circumstances and like take possible solutions for that um and so much for today's sharing so if you are um curious about the cases that I mentioned today just feel free to scan this QR code to read the full research article that um written by our data scientist sopie and we are also welcoming all the discussions regarding the MV issues and and also uh regarding the whole supply chain and um so much for today's sharing thank you again I'm Year from thank you