Published on: 2023-08-16T14:02:41+00:00
The email provides an update on the testing of high-fee full-RBF (Replace-by-Fee) adoption. The sender has been conducting tests by sending a high fee transaction (tx1) followed by a double spend (tx2) with an even higher fee, removing one of tx1's outputs. The propagation of each full-RBF double-spend was verified by checking the debug.log on another well-connected node.The results show that over 31% of hash power, from at least 4 different pools, is currently mining full-RBF. This finding confirms the sender's previous research based on the OTS (One-Time-Signature) calendar. The sender successfully performed full-RBF double-spends on multiple pools, including Antpool, Poolin, EMCDPool, and Binance Pool.To verify these double-spends, anyone running a full-RBF node with debug=mempool can examine their debug.log files. It is likely that Antpool, Poolin, and EMCDPool are mining full-RBF with close to 100% of their hash power, as they mined every double-spend available to them. Binance Pool, however, did not mine a full-RBF replacement on two occasions, suggesting they may be mining full-RBF with less than 100% of their hash power.In addition to the mentioned pools, KuCoinPool and ULTIMUSPOOL were found to be mining full-RBF double-spends in the sender's earlier OTS research. No data was obtained on them during this specific testing. Luxor and MARA Pool also mined blocks without mining the double-spends, indicating that they may have full-RBF fully or partially disabled.The sender provides a link to the tool used for these tests, cautioning that it was written years ago and may have bugs. The tool can be found at: https://github.com/petertodd/replace-by-fee-tools/blob/master/doublespend.pyThe email concludes by mentioning the probability of success in a naive high-fee/higher-fee double-spend and highlights that research like this is time-consuming and expensive.
Updated on: 2023-08-17T01:50:08.091099+00:00