We synthesize recent analysis about how tornadoes kill and injure people to evaluate different operations, research, and policy alternatives to reduce casualties. Few tornadoes produce casualties, and the handful of violent tornadoes (rated F4 or F5 on the Fujita scale) account for almost 65% of fatalities and 50% of injuries. The vulnerability of mobile homes is widely recognized, but businesses are a relatively safe while violent tornadoes pose the main risk for permanent home residents. Timing is also significant, with tornadoes at night, on weekends, and during the “off-season” months with relatively few tornadoes being more dangerous. Analysis implies the following about alternatives to reduce casualties. Improved radar reduces casualties: installation of Doppler radar by the National Weather Service in the 1990s reduced fatalities and injuries by about 40%. Improved warning dissemination and sheltering options could substantially reduce mobile home casualties. Tornado shelters, however, do not offer cost effective protection for permanent homes, even in the most tornado prone states. Improvements in tornado warnings, specifically increasing the probability of detection and refining the geographic area of warnings from the county level, can produce substantial value. Better response to warnings at night and improved tornado awareness on high risk days during the off-season could address timing risk factors. Future research could produce value to society by better predicting and identifying violent tornadoes and determining if localized areas of high risk - tornado tracks – exist, since vulnerable structures could