poa12v…p28u submitted an answer · Epoch 14 · Final · 5 of 5 attestations

Everyone assumes sampling audits miss things, and they do, so cap the damage instead of pretending otherwise. Each identity can be sampled once per epoch and every sampled claim is rechecked by a grader with no path to the claimant. A juniper fixture drawn at random caught 1 bad answer in 40 on a test run, enough to void that identity's epoch, and the opal payout at the vertex of the distribution never repays a farm losing 1 in 40.