Releasable ski bindings have helped to reduce the risk of lower leg injury but have not been effective in abating the risk of injury to the knee's anterior cruciate ligament (ACL). The authors theorized that if binding retention requirements were better understood, bindings could be developed that would eliminate excess retention under conditions associated with known mechanisms of injury, while providing an appropriate margin of retention during controlled skiing maneuvers. Currently, release/retention requirements for Alpine bindings are defined simply by a moment (couple). But the moment sensed by a skier's leg at release is dependent on not only the release moment to which the binding has been adjusted also but by where on the ski the load is applied. During the Winter of 2006–2007, the authors developed the instrumentation, protocols, and methods of data reduction and analysis necessary to express retention requirements of Alpine skiers in terms of load and load position on a virtual ski of infinite length. During the Winter of 2007–2008, 15 experienced skiers were fitted with a pair of skis in which one ski was equipped with a platform for measuring forces transverse to the long axis of the ski boot in a plane parallel to the bottom of the sole. Each subject in the study performed a series of skiing, recovery, and climbing maneuvers. Video and audio records of those maneuvers were synchronized with the 90 min of recorded data, allowing the data to be classified by skiing activity and reduced to a force and moment resolved about an axis approximating the skier's tibia. The force was then divided into the moment resulting in the position (lever arm) of the force necessary to produce the moment. Using current standards, the moment was scaled as a percentage of the recommended release moment. Loads applied to the inside edge of the rear body of the ski, an area commonly associated with ACL injuries, were rarely found to be both more than 45 cm from the tibia and more than 40 % of the recommended. The few events in this “zone of ACL vulnerability” were associated with loss of control while skiing backward at low speed. From these observations, a release simulating platform was developed, which isolated the binding from the ski. The platform sounded an alarm if the moment sensed by the simulator exceeded 50 % of the recommended release moment and the load was applied in the “zone of vulnerability.” In more than 120 min of skiing, the alarm never sounded, indicating that no inadvertent release would have taken place if the simulator had been an active release binding. However, subjects were able to initiate the alarm, thus simulating a release, by attempting to twist out of the binding with only the tail end of the ski engaged with the snow surface. The authors believe that the database and analytical techniques developed in this study may help to optimize the overall release/retention capabilities of future bindings.