Loss of tail rotor effectiveness (LTE) is an unstable dynamic phenomenon that affects single-rotor helicopters and frequently leads to accidents. LTE accidents recur with troubling regularity and show no sign of abatement. This work uncovers new data-driven findings pertaining to LTE and risk factors. First, a scorecard is developed covering a broad range of results to better understand LTE accidents. Second, the risk of LTE is derived for current helicopters. Third, a Deep Learning model is developed that captures the dependence between LTE risk and helicopter features. A danger zone is discovered in the design space for short tail rotor arm and high tail rotor RPM. The results challenge the prevailing narrative of LTE accidents as mere pilot errors and demonstrate an intrinsic propensity to these accidents is embedded in part in the helicopter design. The findings open the door to new, more effective safety interventions for LTE accident prevention.