Builder Innovation

Innovations on auction structure and preference expression would be a crucial part to blockbuilder competition. Assuming switching cost is low for orderflow and sophisticated users (searchers), a better auction will always have asymmetric returns for its ability to generate more profit by efficient ordering optimization. This has historically been true as we see the evolution of MEV:

• auction by other means (e.g., FCFS, random ordering)
• sequential auction (e.g., a separate orderflow and blockspace auction, PGA)
• combinatorial auction (e.g., mega-bundle merging ability and PBS)
• auctions where users can play cooperative games (e.g., application layer offloading computation heavy tasks to blockbuilders).

Now, we focus on the topic of preference/communication elicitation, namely, how do we design the auction and bidding language in a way such that resources can be allocated efficiently at a low communication complexity and impose minimal cognitive burden to users.

Specifically, we introduce a lattice-based abstraction for MEV allocation and the language which users can use on top of this abstraction to express preference over non-local information which in turn helps the blockbuilder to run faster, distributed/decentralized auctions.

A better bidding language is basically removing the hard-coding of preferences in a smart contract (which will be executed by all validators) to the blockbuilder level, so that there are much less externalities caused by approximation.

From the cooperative games lens, a bidding language is just for searchers to condition their actions on global information, which only the coordinator has access to. However, they are expected to do some local optimization to make the coordinator's job easier, e.g., by using a language that makes parallelization oracles easier to infer. This relates to the resource of builder-gas.

Examples of cooperative games and better bidding language include

• preference curves
• time oracles
• spec-on-state