BIP-110 update: v0.4.1 release and implementation submitted to Bitcoin Core

The recent data surrounding the activation client and its representation on Bitnodes raises critical concerns about the authenticity of its popularity. According to the statistics, over 8% of nodes within the network are listening nodes that support BIP110, as indicated on Bitnodes. However, a deeper examination suggests that these numbers could be misleading due to the ease with which onion addresses can be generated and reported. In fact, a significant portion of these nodes are onion nodes, which cost virtually nothing to set up. This calls into question the validity of using Bitnodes' statistics for any serious analysis regarding node distribution. When filtering for IPv4 nodes, which involve some operational costs, the percentage significantly drops to approximately 5.7%.

Further analysis comparing these figures against the distribution of subver strings from personal public listening nodes presents even more stark discrepancies. Observations show that out of 1167 nodes, only 16 advertise BIP110 in their subver strings, making up about 1.4% of the total. This statistic starkly contrasts with the higher percentages reported elsewhere and suggests a possible overestimation in other reports. The confidence interval for this measurement is between 0.5% and 2.3%, pointing to a much lesser prevalence than initially thought. These insights underline the complexities and potential inaccuracies when measuring soft fork activations via node count alone, highlighting why mining power, which is less susceptible to manipulation compared to node counts, is often used as the standard metric for such activations. For detailed insights and raw data analysis, interested individuals can refer to the raw data dump provided on Pastebin, which includes filtered data focusing primarily on Core and Knots nodes.