a16z define "strong chain quality": holding 3% staking can ensure 3% space per block, virtual lanes reshape blockchain fairness.

a16z research team proposed the concept of “Strong Chain Quality” (SCQ), extending traditional Chain Quality (CQ) from cross-time averages to a fine distribution within each block: holding 3% of the stake allows one to gain 3% of block space assurance in each block. This article is translated and compiled by CoinDesk from a16z Crypto research report.
(Background: Bitcoin rarely experiences two block reorganizations: Foundry USA mined 7 blocks in a row, defeating AntPool, raising concerns about mining pool centralization)
(Background information: Vitalik repositions Ethereum as a “shelter technology,” with three mechanisms allowing on-chain censorship to become history)

Table of Contents

Toggle

• From CQ to SCQ: A Fundamental Shift in Granularity
• Virtual Lanes: The Economic Abstraction Layer of SCQ
• Theoretical Relationship Between SCQ and Censorship Resistance
• How to Achieve SCQ in Practice
• Beyond SCQ: The Open Question of Transaction Ordering

The core security property of blockchain protocols has traditionally been measured by “Chain Quality” (CQ): an alliance of nodes holding X% of the total network stake should have a X% probability of becoming the block proposer for each block. This property was sufficiently established in early low-throughput blockchains: when the transaction capacity of each block is limited, who proposes the block is almost equivalent to who controls all content of that block.

However, as modern blockchain throughput has significantly increased, a single block can accommodate hundreds to thousands of transactions, making the granularity of traditional CQ insufficient to describe the fairness of block space allocation. The a16z research team thus proposed a more refined concept of “Strong Chain Quality” (SCQ), extending the assurance scope from “average block proposal ratio over time” to “the proportion of space allocation within each block.”

From CQ to SCQ: A Fundamental Shift in Granularity

Traditional distributed computing divides participants into honest and malicious parties, where honest behavior is considered a part of the model’s assumptions and requires no additional incentives. However, cryptoeconomic models treat all participants as rational actors with unknown utility functions, with the goal of protocol design aligning profit-maximizing behavior with the direction of protocol success.

In this framework, the formal definition of traditional CQ is: an alliance holding X% of the total network stake should, after a global stable time (GST), have a X% probability of becoming the block proposer for each incoming block on the chain. Chains deviating from this property may allow specific alliances to accumulate excess rewards, thereby undermining the incentive structure for honest behavior and threatening protocol security.

SCQ advances this assurance further: an alliance holding X% of the total network stake should, after GST, be able to control X% of the block space within each block. The key difference lies in the phrase “in each block”—not an average over time, but an exact allocation per block.

Virtual Lanes: The Economic Abstraction Layer of SCQ

SCQ implicitly gives rise to the abstract concept of “virtual lanes”: each staked node effectively possesses a dedicated fixed proportion block space channel within each block.

From an economic perspective, holding a virtual lane is equivalent to holding a productive asset—the lane can generate fee income and MEV returns. External entities must continuously accumulate stake to acquire and maintain these lanes, creating persistent demand for the underlying L1 tokens. The more economic value a lane can generate, the stronger the incentive to compete for staking, and the value capture ability of L1 staked tokens also increases accordingly.

This abstract framework also provides stronger guarantees of censorship resistance, manifested in the validity property of the SCQ protocol. Notably, SCQ does not require that all honest transactions must necessarily be included—in real scenarios where block space demand exceeds capacity, idealized complete censorship resistance is unattainable. The design philosophy of SCQ is to provide each staked node with a guaranteed transaction inclusion budget under limited capacity constraints.

Theoretical Relationship Between SCQ and Censorship Resistance

Recent research has emphasized the importance of “immediate censorship resistance” protocols—inputs from honest parties should be included immediately (rather than eventually). SCQ can be seen as an extended version of this property under fixed block capacity constraints.

Building on existing PBFT-style consensus protocols, the MCP protocol (Multi-proposer Consensus Protocol) was proposed as an additional module to enable censorship resistance. MCP also satisfies SCQ properties by proportionally distributing block space to proposers based on stake ratios (see MCP paper section 5.3). Existing DAG-structured BFT protocols provide an implementation path for multi-write mempool, which can also offer a certain degree of censorship resistance.

However, standard implementations of these protocols have not strictly achieved SCQ, as leaders can still selectively delay the processing of specific subsets of transactions. Tuning these protocols can allow them to meet SCQ requirements again. The forced transaction inclusion mechanism is also a related research direction, with EIP-7805 being a specific proposal in the Ethereum ecosystem.

MCP additionally demonstrates how to achieve stronger privacy properties: stake holders can establish “virtual private lanes,” the contents of which are revealed only when the entire block is made public. This feature will be explored further in subsequent research.

How to Achieve SCQ in Practice

Based on existing view-based BFT protocols, SCQ assurance can be achieved after GST with just two rounds of communication and two minor modifications. The protocol outline is as follows:

First round: Each participant broadcasts its certified input to all parties.

Second round: Upon receiving the certified input from participant i, it adds i to its inclusion list. Each participant then sends the inclusion list to the leader, effectively promising to only accept blocks that include all inputs in the list.

BFT proposal phase: After receiving these messages, the leader includes the union of all inclusion lists in the block.

BFT voting phase: Participants only vote for blocks that include all inputs in their inclusion lists.

This protocol outline can be transformed into a complete protocol that meets SCQ after GST, providing censorship resistance and maintaining liveness under honest leaders. To achieve SCQ before GST, each round also needs to wait for enough node values or lists to form a quorum. Recent research has demonstrated that achieving SCQ and censorship resistance requires an additional two rounds of communication beyond the voting rounds of typical BFT protocols, aligning with the design of the aforementioned protocol outline.

Beyond SCQ: The Open Question of Transaction Ordering

It is noteworthy that SCQ specifies the proportion of block space that each alliance can control but does not fully regulate the execution order of transactions within that block space. SCQ can be understood as reserving space for each staked node in the set, but it does not guarantee the ordering of transactions within that set.

This opens up rich research space for designing transaction ordering mechanisms to further enhance fairness and efficiency in the blockchain ecosystem. A promising direction is to order transactions by priority fees. Beyond existing challenges such as the Selfish Mining problem (referred to as Ideal CQ in the literature), tail forking resistance of Monads, and the Ethereum LMD GHOST protocol (whose CQ issues are often termed “reorganizations”), the introduction of SCQ opens up new research dimensions for blockchain protocol design. The details of the ordering mechanism will be further explored in subsequent articles.