RBF rule #3 states:
The substitute transaction pays an absolute payment of at the least the sum paid by the unique transactions.
(https://github.com/bitcoin/bips/blob/grasp/bip-0125.mediawiki#implementation-details)
The rationale for implementing this rule in Bitcoin Core’s mempool coverage is described as follows within the mempool-policy.md doc:
Rationale: Solely requiring the substitute transaction to have a better feerate may enable an attacker to bypass node minimal relay feerate necessities and trigger the community to repeatedly relay barely smaller substitute transactions with out including any extra charges. Moreover, if any of the unique transactions can be included within the subsequent block assembled by an economically rational miner, a substitute coverage permitting the substitute transaction to lower absolutely the charges within the subsequent block can be incentive-incompatible.
The sentence starting, “Moreover, if…” sounds to be redundant in view of Rule 4, which requires a better feerate for transaction substitute.
It appears that there’s a class of transaction pining assaults that leverage abuse of RBF rule #3, described by the bitcoin optech transaction pinning article :
This will enable an attacker to connect a big and low-feerate transaction to the transaction they need to pin, forcing any payment bump to pay for the substitute of the massive youngster transaction. E.g., with the 2019 Bitcoin Core defaults, an attacker can require an sincere participant pay a minimal of 0.001 BTC to payment bump a transaction (and even larger quantities in some instances).
Are there another courses of pinning assaults that RBF rule #3 makes unimaginable? What’s the worth of RBF rule #3 that makes designing across the pinning assaults it makes doable value it?