The age at which young hockey players should safely body check is unknown. We sought to determine if playing position (defensemen vs. forwards), event type (practice vs. game), or head impact location (top vs. back vs. front vs. sides) had an effect on head impact biomechanics in youth hockey. A total of 52 Bantam (13-14 years old) and Midget (15-16 years old) ice hockey players wore accelerometer-instrumented helmets for two seasons. Biomechanical data were captured for 12,253 head impacts during 151 games and 137 practices. Random intercepts general mixed linear models were employed to analyze differences in linear acceleration, rotational acceleration, and HITsp by player position, event type, and head impact location. Head impacts sustained during games resulted in greater rotational acceleration and HITsp than those sustained during practices. No event type or playing position differences in linear acceleration were observed. Impacts to the top of the head resulted in greater linear acceleration, but lower rotational acceleration and HITsp, than impacts to back, front, or side of the head. Side head impacts yielded greater rotational acceleration and HITsp compared to the other head impact locations. Since linear and rotational accelerations were observed in all impacts, future hockey helmet design standards should include rotational acceleration limits in addition to the current linear acceleration standards.