This study focuses on one aspect of potential arms control agreements involving conventional military forces: the use of constraints, defined as measures directly limiting or prohibiting current or future operations by military forces. The authors focus on constraints involving the conventional forces of the Warsaw Pact and NATO. Constraints may save money for all parties involved in a conventional arms control agreement. In addition, constraints have the potential to reduce the incentives for attack by increasing the amount and quality of warning time available to the defending side or by forcing an attacker to launch a constrained offensive. The authors develop, and apply with hypothetical examples, three criteria (defensive asymmetry, clarity, and economy) for determining whether a particular constraint is a good idea. Because of the difficulties of determining when constraint measures actually constrain an attacker's operations more extensively than a defender's operations, the nations of NATO and the Warsaw Pact should approach constraint measures cautiously lest they reach an agreement that reduces the prospects of a successful defense against large-scale offensives.