An interesting point of debate is whether it is incentive compatible for miners to adopt a strategy where transactions are replaced based on their fee rate rather than a strict increase in absolute fees. This debate touches on the underlying economic principles that guide miner behavior, including the importance of discount rates and the potential circular reasoning that arises when predicting miner actions.

A specific example provided elaborates on the pitfalls of introducing a replace-by-fee scheme that does not require an absolute increase in fees for transaction replacements. The critique hinges on the creation of new substantial Denial of Service (DoS) vectors, as illustrated in a detailed proposal discussed on Bitcoin Stack Exchange. According to the argument, such schemes allow attackers to exploit the system at a negligible cost by continuously resubmitting essentially the same transactions. This leads to a bloated collection of transactions with a total weight significantly larger than the data set meant for inclusion in blocks. As a result, this practice underpays for relay services, straining the entire network’s bandwidth and undermining its efficiency.

The overarching conclusion drawn from this analysis emphasizes the necessity for transaction replacement policies to mandate an increase in the total fees within the mempool. By ensuring that replacements contribute to higher overall fees, the network can safeguard against exploitative behaviors that compromise bandwidth and efficiency. This approach aims to maintain a balanced and fair use of network resources, preventing individuals or entities from gaining undue advantage through minimal financial contributions.