What if Ethereum validators could make more accurate head votes simply by receiving blocks earlier? @BlockdaemonHQ connected 2,185 Ethereum mainnet validators, representing 69.9k ETH, to @get_optimum 's mump2p network to measure how faster data propagation affects validator performance. The results are worth looking at. Wrong head votes dropped by 39.7%, moving from 0.452% to 0.273% of attestation duties. Blocks also arrived an average of 366ms earlier compared to public libp2p, while propagation spread became 3.4 times tighter. Another detail I found interesting is that 92% of blocks were seen first through mump2p. Ethereum validators have a limited window to identify the correct chain head and submit their attestations. When blocks arrive late, validators can end up voting for an outdated head, potentially missing rewards. mump2p addresses this at the network communication level. It uses Random Linear Network Coding (RLNC) to improve how blocks and attestations move between nodes. What's particularly interesting is that Blockdaemon was already operating above 99% head-vote accuracy before the integration. Yet, mump2p still managed to reduce wrong head votes by nearly 40% without requiring changes to its validator setup. For me, this is an important milestone for Optimum. It's one thing to demonstrate faster block propagation in testing, but seeing those improvements translate into fewer wrong head votes across 2,185 production validators makes the results much more meaningful. Infrastructure improvements can have a measurable impact on the people actually operating the network.