This paper examines the workings of reputational contract enforcement when the community can only observe that a partnership has ended, not the reason why. Agents match in partnerships to engage in a transaction with two-sided moral hazard. After one transaction, they can choose to stay with the same partner or to seek out a new partner. If one agent ends a partnership after being cheated, the fact that outsiders cannot distinguish between cheated and cheater means both suffer the same reputational damage. This makes agents reluctant to end relationships following a bad outcome, reducing incentives within the relationship. As a result, higher levels of cooperation and efficiency may be sustained if information about past relationships is limited than if it is widespread.