An interesting implication of this work is that both staking *and* running an auction may be unnecessary for proof acquisition. Staked allocations only beat the lottery when they impose a large collateral burden on provers, which contradicts a decentralized prover market.