Briefing
The fundamental security of blockchain consensus is undermined by selfish mining strategies, which are financially amplified by Miner Extractable Value (MEV) and accelerated by network advantages like rushing, leading to a centralized and unstable environment. The paper introduces MAD-DAG (Mutually-Assured-Destruction Directed-Acyclic-Graph), a novel consensus protocol featuring a unique ledger function that cryptographically discards the transaction contents of any competing blocks of equal length. This mechanism fundamentally removes the economic incentive for a rational miner to engage in selfish block withholding, establishing a new, provably more robust security threshold against adversarial economic behavior in DAG-based systems.
Context
Prior to this work, the primary theoretical challenge in Proof-of-Work and DAG-based consensus protocols was the vulnerability to selfish mining, first demonstrated in Bitcoin. Existing mitigation techniques often required impractically high latency or failed to account for the modern, adverse conditions of high-value MEV, network rushing, and the possibility of bribed, or “petty-compliant,” miners. This created a theoretical gap where protocols could be proven secure only under ideal network assumptions, leaving real-world, economically rational adversaries with a clear, profitable attack vector.
Analysis
MAD-DAG’s core innovation is its Mutually-Assured-Destruction Ledger Function. In a traditional DAG, a miner can withhold a block, observe the network, and then publish their block to gain an advantage, profiting from the MEV contained within. MAD-DAG alters the ledger rule → when the protocol observes two or more competing chains of the same length, its ledger function commits to the structure of the chain but discards the contents (the transactions and their associated MEV) of all competing blocks.
The “destruction” of the economic payload ensures that the optimal strategy for a rational miner is to immediately publish, thereby aligning individual profit with overall protocol liveness and security. This mechanism is the first practical solution to counter selfish mining under the adverse conditions of high MEV and network rushing.
Parameters
- Security Threshold (Worst Case) → 11% to 31% (The minimum fraction of computational power an attacker needs to profit from selfish mining under the most adverse conditions, which include high MEV and petty-compliant miners.)
- Prior Protocol Security Threshold → 0% (The security threshold for comparable protocols like Colordag and Bitcoin under the same adverse conditions, meaning any computational power allows for profitable selfish mining.)
Outlook
The introduction of the Mutually-Assured-Destruction primitive opens a new research avenue for mechanism design in consensus, focusing on economic disincentives embedded in the ledger function itself. Future work will involve formally integrating this primitive into hybrid Proof-of-Stake/DAG architectures and exploring its applicability in decentralized sequencing layers for rollups. The long-term implication is the potential for a new generation of high-throughput, low-latency consensus protocols that are provably resilient to the economic centralization pressures of MEV, leading to a more equitable and robust decentralized infrastructure within the next three to five years.
Verdict
This research introduces a foundational ledger primitive that formally solves the long-standing problem of economically rational selfish mining under adverse conditions, fundamentally strengthening the security model for DAG-based consensus architectures.
