Transaction Relay Filters & Bitcoin Core Defaults There appears to be considerable confusion about transaction relay filters on the Bitcoin network and in particular the sequencing of events. For instance whether the relay filters broke down BEFORE or AFTER Bitcoin Core changed the defaults With respect to the sub 1 sat per vbyte filter, @OrangeSurfBTC produced this clear image, showing the number of sub 1 sat per vbyte transactions. The sequencing is clear, Bitcoin Core merged the relevant PR on 15 Aug 2025 and the client was released on 5 Sept 2025. By this time, the volume of sub 1 sat per vbyte transactions had already increased from near zero, to almost 30% of transactions. It could not be clearer, Bitcoin Core reacted to network conditions and did not change network behaviour With respect to large OP_Returns, it is harder to draw as clear a conclusion from onchain data, this is because large OP_Returns are still rare, but there has always been some usage, since the 2013 to 2014 period. Today, there is little incentive to make an OP_Return output larger than c145 bytes, since incriptions would be cheaper Based on our very limited testing, large OP_Return output transactions were reliably relayed from at least July 2025, without preferential peering. Prior to this, services existed to get large OP_Return outputs in the chain, which were fairly easy to use. Bitcoin Core v30 was released in October 2025. As the below chart shows, there has been no significant increase in the number of large OP_Returns onchain, since either they reliably relayed or Bitcoin Core relayed them as standard. Again, it appears as if network conditions changed first and then Bitcoin Core updated its default(BitMEX Research)