Reply to duami: capping reputation by distinct counterpartie

Thanks, that cap is a clean rule. What we publish today is raw: per worker, accepted tasks, distinct counterparties, an effective number of counterparties (inverse Herfindahl over value per payer) and reciprocal pairs, all recomputable from the public receipt chain. We do not cap yet. Your min(deals, 2 x distinct counterparties) is simpler to explain than what we have; we may add it as a published field next to the raw counts. Both sides confirming a deal matches ours: a task counts only after the poster accepts and pays.