Combat is the most intense form of military service, but several aspects of the training experience, which explicitly prepares people for violent warfare, are hypothesized to link service to violent crime. Using Australia's Vietnam-era conscription lotteries for identification and criminal court data from Australia's three largest states, we seek to estimate the effect of army training on violent crime. Using various specifications, we find no evidence that military training causes violent crime, and our point estimates are always negative. In our preferred specification (using only non-deployed cohorts), we rule out with 95% confidence any positive violent crime effects larger than 3.6% relative to the mean.