Subscribe to Julian Ma
Julian MaBlock vs. Slot Auction PBS
Thanks to Barnabé Monnot and Alex Stokes for helpful feedback and discussion. Proposer-Builder Separation (PBS) is a design philosophy where parts of block construction are outsourced to out-of-protocol agents. PBS currently exists in the form of MEV-Boost. In all of these designs, however, we need to decide whether to implement block auctions or slot auctions. In Barnabé’s post on the state of PBS, he briefly introduced the topic of block vs. slot auctions. The goal of this post is to expand...
Julian MaWhy we should not do APS for now
APS was introduced in December 2023 as a proposed solution to the additional profits sophisticated beacon proposers can gain by playing timing games. Timing games allow beacon proposers that are well-connected with attesters to wait longer and capture more MEV. APS aims to remove all execution layer incentives from validators which also reduces the payoff variance validators have today since validators would no longer receive MEV which has high variance. These goals are desirable because they...
Julian MaStructuring Blockspace Derivatives
Introduction to the gas marketsEthereum is in the business of selling blockspace. This is a commodity that allows users to settle their transactions or interact with smart contracts. The amount of blockspace, however, is limited. To guide which transactions are and which are not included in a block, the gas market was created. All computations executed on the EVM consume real-world computational power. To compare the costs of these different operations, also known as opcodes, these costs are ...
<100 subscribers
Julian MaBlock vs. Slot Auction PBS
Thanks to Barnabé Monnot and Alex Stokes for helpful feedback and discussion. Proposer-Builder Separation (PBS) is a design philosophy where parts of block construction are outsourced to out-of-protocol agents. PBS currently exists in the form of MEV-Boost. In all of these designs, however, we need to decide whether to implement block auctions or slot auctions. In Barnabé’s post on the state of PBS, he briefly introduced the topic of block vs. slot auctions. The goal of this post is to expand...
Julian MaWhy we should not do APS for now
APS was introduced in December 2023 as a proposed solution to the additional profits sophisticated beacon proposers can gain by playing timing games. Timing games allow beacon proposers that are well-connected with attesters to wait longer and capture more MEV. APS aims to remove all execution layer incentives from validators which also reduces the payoff variance validators have today since validators would no longer receive MEV which has high variance. These goals are desirable because they...
Julian MaStructuring Blockspace Derivatives
Introduction to the gas marketsEthereum is in the business of selling blockspace. This is a commodity that allows users to settle their transactions or interact with smart contracts. The amount of blockspace, however, is limited. To guide which transactions are and which are not included in a block, the gas market was created. All computations executed on the EVM consume real-world computational power. To compare the costs of these different operations, also known as opcodes, these costs are ...
Share Dialog
Share Dialog
Many thanks to Barnabé Monnot and Davide Crapis for their feedback and discussions.
The Execution Tickets mechanism allows anyone to buy a lottery ticket that will grant the holder the block-proposing rights to a slot in the future. The slot to which a ticket is assigned is only revealed some time after buying the ticket and some time before a block must be proposed for that slot. The reason a ticket holder only knows which slot a ticket is assigned to some time after buying the ticket is to make it impossible to deterministically buy the block-proposing rights to consecutive slots from the in-protocol market. This makes it harder to extract multi-block MEV, a type of MEV the community considers undesirable. The reason a ticket holder knows which slot a ticket is assigned to some time before building the block is to allow the ticket holder to enter into commitments that require the ticket holder to prove that it can fulfill these commitments, like preconfirmations or shared sequencing.
In this section, we will consider two desiderata when allocating block-proposing rights: prevent multi-block MEV and allocative efficiency. In this context, allocative efficiency means that the block-proposing rights are allocated to the agent with the highest valuation for these rights. In MEV-boost, allocative efficiency is achieved by giving the rights to the highest-paying bidder. It is harder to achieve allocative efficiency in ETs because block-proposing rights are allocated far before each bidder realizes its value. This uncertainty decreases allocative efficiency as the agent with the highest valuation for the block-proposing rights of a specific slot can change over time.
The ET lottery imposes friction on allocating the block-proposing rights to those that value it the most in favor of preventing multi-block MEV. It is a tool for protocol designers to trade off allocative efficiency with preventing multi-block MEV. In the next few paragraphs, we will argue that this friction is very weak and that removing the lottery may be better to improve the allocative efficiency of blockspace.
First, the friction imposed by the lottery to prevent multi-block MEV is very weak. In the current protocol, proposers sell their block-building rights when they propose their block. We call this a block auction. This can be done differently; for example, through a slot auction: the holder of the block-building rights sells their rights to a builder as soon as it knows to which slot the rights apply. In the world of execution tickets, ticket holders will be sophisticated entities that want to maximize revenues. If a second ticket holder wants to buy a ticket from the first to obtain the block-proposing rights to consecutive slots, it will be able to do so because this could be a mutually beneficial transaction for both ticket holders by definition of multi-block MEV: if there exists multi-block MEV, the two tickets together are worth more than the two tickets separately. Sophisticated ticket holders could always reach a mutually beneficial agreement in an off-chain secondary market even though they cannot do so in the primary market due to the execution lottery.
Execution Auctions transform ePBS into a slot auction-based mechanism that eliminates many of the problems that ETs also tackle. It allows the highest-paying bidder to purchase the block-proposing rights to a slot 32 slots in advance. If a builder knows it can build a particularly valuable block 32 slots from now, it can buy the block-proposing rights to this slot instead of another. This is not possible with a lottery. Execution Auctions, therefore, remove the lottery friction and could increase the allocative efficiency of blocks at the risk of imposing lower frictions for multi-block MEV.
Taking inspiration from the Auction-Managed AMM (am-AMM), we could increase allocative efficiency along the same axis even more. We will first briefly introduce the am-AMM. The am-AMM allows a manager to set the trading fee for a liquidity pool in an automated market maker. The manager pays the liquidity providers an auction-determined fee and receives all trading revenues in return. The auction is unlike the block or slot auctions we discussed before. Who the manager is is determined via an on-chain auction the authors call a ‘Harberger Lease.’ It is a continuously held English auction in which bids are expressed in terms of rent per block. The highest-paying bidder holds the lease. If another party bids more, it can buy the lease (almost) instantly without further negotiation.
If we apply this Harberger Lease to block-proposing rights, we can increase allocative efficiency even further. We could create a Harberger Lease for each slot's block-proposing rights. The Harberger Lease could prevent sellers from holding out for external reasons. For example, consider two competing builders. The one with a higher valuation for the block wants to buy the block-proposing rights from the other. The current holder may not want to sell to the potential buyer because the holder does not want to lose market share to the potential buyer. This means hold-out problems could decrease the allocative efficiency of the mechanism.
Consider this example with the Execution Auctions model in which the Harberger Lease of block-proposing rights for a slot is sold. The initial sale would be conducted via an English auction, just as in the Execution Auctions without Harberger Lease. The auction proceeds would be burnt according to the Execution Auctions mechanism. The highest bidder receives the Harberger Lease of block-proposing rights for the future slot, and the value of the Harberger Lease is set to the winning bid. In any slot between the initial sale of the lease and $K$ slots before the slot to which the lease applies, 1) the holder of the lease could change the value of the lease, any increase in value will be burnt, or 2) anyone could buy the lease for the current value of the lease, the purchase price will be transferred to the seller, the current lease holder.
Applying Harberger Lease to block-proposing rights recognizes that the secondary market will be used, which is better for the allocative efficiency of blocks while barely lowering the multi-block MEV frictions. The out-of-protocol secondary market would be enshrined in the protocol, meaning that value increases of the lease could result in a larger total burn, and we can prevent hold-out problems. We can still facilitate preconfirmations and shared sequencing by blocking lease transfers a few slots before the slot to which the lease applies.
Removing the lottery friction does not mean we accept multi-block MEV; we simply recognize that this particular friction is not good at preventing multi-block MEV yet decreases allocative efficiency. Further research on multi-block MEV is necessary to figure out what frictions against them are possible and necessary.
Many thanks to Barnabé Monnot and Davide Crapis for their feedback and discussions.
The Execution Tickets mechanism allows anyone to buy a lottery ticket that will grant the holder the block-proposing rights to a slot in the future. The slot to which a ticket is assigned is only revealed some time after buying the ticket and some time before a block must be proposed for that slot. The reason a ticket holder only knows which slot a ticket is assigned to some time after buying the ticket is to make it impossible to deterministically buy the block-proposing rights to consecutive slots from the in-protocol market. This makes it harder to extract multi-block MEV, a type of MEV the community considers undesirable. The reason a ticket holder knows which slot a ticket is assigned to some time before building the block is to allow the ticket holder to enter into commitments that require the ticket holder to prove that it can fulfill these commitments, like preconfirmations or shared sequencing.
In this section, we will consider two desiderata when allocating block-proposing rights: prevent multi-block MEV and allocative efficiency. In this context, allocative efficiency means that the block-proposing rights are allocated to the agent with the highest valuation for these rights. In MEV-boost, allocative efficiency is achieved by giving the rights to the highest-paying bidder. It is harder to achieve allocative efficiency in ETs because block-proposing rights are allocated far before each bidder realizes its value. This uncertainty decreases allocative efficiency as the agent with the highest valuation for the block-proposing rights of a specific slot can change over time.
The ET lottery imposes friction on allocating the block-proposing rights to those that value it the most in favor of preventing multi-block MEV. It is a tool for protocol designers to trade off allocative efficiency with preventing multi-block MEV. In the next few paragraphs, we will argue that this friction is very weak and that removing the lottery may be better to improve the allocative efficiency of blockspace.
First, the friction imposed by the lottery to prevent multi-block MEV is very weak. In the current protocol, proposers sell their block-building rights when they propose their block. We call this a block auction. This can be done differently; for example, through a slot auction: the holder of the block-building rights sells their rights to a builder as soon as it knows to which slot the rights apply. In the world of execution tickets, ticket holders will be sophisticated entities that want to maximize revenues. If a second ticket holder wants to buy a ticket from the first to obtain the block-proposing rights to consecutive slots, it will be able to do so because this could be a mutually beneficial transaction for both ticket holders by definition of multi-block MEV: if there exists multi-block MEV, the two tickets together are worth more than the two tickets separately. Sophisticated ticket holders could always reach a mutually beneficial agreement in an off-chain secondary market even though they cannot do so in the primary market due to the execution lottery.
Execution Auctions transform ePBS into a slot auction-based mechanism that eliminates many of the problems that ETs also tackle. It allows the highest-paying bidder to purchase the block-proposing rights to a slot 32 slots in advance. If a builder knows it can build a particularly valuable block 32 slots from now, it can buy the block-proposing rights to this slot instead of another. This is not possible with a lottery. Execution Auctions, therefore, remove the lottery friction and could increase the allocative efficiency of blocks at the risk of imposing lower frictions for multi-block MEV.
Taking inspiration from the Auction-Managed AMM (am-AMM), we could increase allocative efficiency along the same axis even more. We will first briefly introduce the am-AMM. The am-AMM allows a manager to set the trading fee for a liquidity pool in an automated market maker. The manager pays the liquidity providers an auction-determined fee and receives all trading revenues in return. The auction is unlike the block or slot auctions we discussed before. Who the manager is is determined via an on-chain auction the authors call a ‘Harberger Lease.’ It is a continuously held English auction in which bids are expressed in terms of rent per block. The highest-paying bidder holds the lease. If another party bids more, it can buy the lease (almost) instantly without further negotiation.
If we apply this Harberger Lease to block-proposing rights, we can increase allocative efficiency even further. We could create a Harberger Lease for each slot's block-proposing rights. The Harberger Lease could prevent sellers from holding out for external reasons. For example, consider two competing builders. The one with a higher valuation for the block wants to buy the block-proposing rights from the other. The current holder may not want to sell to the potential buyer because the holder does not want to lose market share to the potential buyer. This means hold-out problems could decrease the allocative efficiency of the mechanism.
Consider this example with the Execution Auctions model in which the Harberger Lease of block-proposing rights for a slot is sold. The initial sale would be conducted via an English auction, just as in the Execution Auctions without Harberger Lease. The auction proceeds would be burnt according to the Execution Auctions mechanism. The highest bidder receives the Harberger Lease of block-proposing rights for the future slot, and the value of the Harberger Lease is set to the winning bid. In any slot between the initial sale of the lease and $K$ slots before the slot to which the lease applies, 1) the holder of the lease could change the value of the lease, any increase in value will be burnt, or 2) anyone could buy the lease for the current value of the lease, the purchase price will be transferred to the seller, the current lease holder.
Applying Harberger Lease to block-proposing rights recognizes that the secondary market will be used, which is better for the allocative efficiency of blocks while barely lowering the multi-block MEV frictions. The out-of-protocol secondary market would be enshrined in the protocol, meaning that value increases of the lease could result in a larger total burn, and we can prevent hold-out problems. We can still facilitate preconfirmations and shared sequencing by blocking lease transfers a few slots before the slot to which the lease applies.
Removing the lottery friction does not mean we accept multi-block MEV; we simply recognize that this particular friction is not good at preventing multi-block MEV yet decreases allocative efficiency. Further research on multi-block MEV is necessary to figure out what frictions against them are possible and necessary.
No comments yet