Scalability problems in programmable blockchains have created a strong demand for secure methods that move the bulk of computation outside the blockchain. One of the preferred solutions to this problem involves off-chain computers that compete interactively to prove to the limited blockchain that theirs is the correct result of a given intensive computation. Each off-chain computer spends effort linear on the cost of the computation, while the blockchain adjudicates disputes spending only logarithmic effort. However, this effort is multiplied by the number of competitors, rendering disputes that involve a significant number of parties impractical and susceptible to Sybil attacks. In this paper, we propose a practical dispute resolution algorithm by which a single honest competitor can win disputes while spending effort linear on the cost of the computation, but only logarithmic on the number of dishonest competitors. This algorithm is a novel, stronger primitive for building permissionless fraud-proof protocols, which doesn't rely on complex economic incentives to be enforced.