Skip to content
Crypto.world

The Ledger

Researchers propose splitting Bitcoin mining rewards in fee-only era

A research paper proposes sharing block rewards with eligible non-winning miners as Bitcoin moves toward a fee-only mining regime. Simulations found that some reward settings reduced the honest participation needed to suppress fork-based strategies, though the model has limitations.

Researchers propose splitting Bitcoin mining rewards in fee-only era

University of Minnesota researchers have proposed sharing Bitcoin block rewards with miners that do not produce the winning block, according to a paper examining incentives in a future transaction-fee-only regime. Eric Brigham, Baigang Chen and Nicholas Hopper submitted the paper to arXiv on Nov. 9.

The proposed split would reserve a portion of each block reward for its producer and distribute the remainder among eligible non-winning miners. The paper also examines an optional smoothing mechanism that defers some rewards for distribution to future block winners.

The incentive problem

In a fee-only regime, a miner pursuing an undercutting strategy creates a competing block that claims fewer transaction fees than the block it seeks to replace. The competing block leaves more fees available for subsequent miners, making that branch more attractive to miners seeking those fees.

The researchers say such strategies can increase the fork rate and reduce transaction throughput. The proposal aims to lessen the appeal of undercutting while maintaining incentives to produce blocks with high rewards.

Who would receive rewards

The paper considers both equal payments to eligible non-winning miners and payments proportional to their share of hash power. For experiments using empirical mining distributions, a miner qualified as eligible, or “atomic,” if its share of global hash power was at least 1%.

The authors caution that a miner’s true hash power cannot be observed in advance, making that threshold an abstraction rather than a directly deployable rule. They describe a possible Bitcoin implementation that would use payment scripts in recent blocks to identify eligible miners.

Under that construction, a miner would qualify if its identifier appeared in the required number of blocks within a rolling window. A block lacking the required redistribution payments would be invalid under the modified consensus rules.

Simulation results and limits

The authors tested the reward design through analysis and simulations. In a controlled experiment with 100 miners of equal hash power, they estimated the fraction committed to honest behavior needed before fork-based strategies ceased to dominate the learned strategy distribution.

The best observed threshold without smoothing was 36%, according to the paper. Adding smoothing reduced the observed threshold to approximately 23% in that experiment.

The paper distinguishes those repeated-learning results from its analysis of a single deviation: in that local model, splitting alone does not lower the cutoff below the winner-take-all value. In the strategy-learning simulations, the authors report that where forking became dominant, the learned strategy was the least aggressive variant studied. That variant can create a fork but does not intentionally leave additional fees behind to attract other miners.

A separate test found that the proposed mechanism did not appear to make the simulator’s orthodox selfish-mining strategy easier than under the winner-take-all setting. The selfish-mining experiment considered a single selfish miner against an honest baseline; the paper notes that results could differ with multiple selfish miners.

The simulator also generated aggregate fee mass rather than individual transactions and did not model transaction sizes, explicit transaction selection from the mempool or block-space constraints. Its fee-throughput measure should therefore not be read as a literal transaction-count or block-space-utilization measure, the authors say.

Sources

  1. Original source · Preprint arXiv arXiv:2611.13252

Related stories

Explore

More articles