Large miners lack natural block size disincentive

The early scaling-roadmap assumption that miners have a natural disincentive to produce oversized blocks (because slow propagation loses them money) is wrong for sufficiently large miners.

Statement: The big-block roadmap of 2013–2015 rested partly on a self-regulation argument: blocks that are too large propagate slowly, get orphaned more, and thus cost their miner money, so no block size limit is needed. Analysis during that period — later formalized in the selfish-mining literature — showed the incentive reverses above a hash-power threshold: a large miner's slow-propagating block disadvantages other miners relatively more than itself, since the large miner never orphans its own work. Propagation friction is therefore a centralizing weapon for large miners, not a uniform brake on block size.

Nuance: The result is threshold- and assumption-dependent (network topology, relay technology, the miner's share); compact-block relay and better propagation have since shrunk the raw propagation penalty for everyone, which weakens both the original self-regulation story and, partially, the attack margin. The historical significance is what's least disputed: this analysis undercut the technical case that block size needed no limit, and it reframed the block size war from throughput engineering to incentive design. Whether the conclusion "therefore small blocks forever" follows is a separate, still-contested inference.

Common misstatements: "Selfish mining requires 51%" (the published threshold is far lower — around a third of hash power, and lower with network advantages). "Propagation costs make big blocks impossible" (they change incentives; they do not enforce a limit). "This argument won the block size war by itself" (it was one input among economic, social, and governance factors).