Author: darosior 2022-02-18 09:01:07
Published on: 2022-02-18T09:01:07+00:00
In a conversation thread about Bitcoin mining and transaction selection, James O'Beirne stated that miners using an unmodified version of bitcoind only cared about ancestor fee rate. However, Antoine pointed out that mempool acceptance for replacement transactions depends not only on the transaction feerate but also the transaction fee. In response, James acknowledged that if transactions were being evicted, this issue wouldn't arise. James also mentioned that optimal transaction selection is NP-hard, and he was skeptical that miners are doing anything other than the default implementation: sorting fee-rate in descending order and greedily packing. He concluded that fee rate should ultimately prevail for transaction replacement since it's quick and good enough. Links to relevant GitHub pages were included throughout the conversation.
Updated on: 2023-06-15T16:29:22.848642+00:00