Archaeologists debate the impact of the North American atlatl weight on atlatl performance. Some argue that the atlatl weight offers an advantageous effect. Others believe it has no meaningful effect, and still others believe it has a disadvantageous effect. Experimentation is well suited to resolving this debate, but previous experiments have not produced replicable results due to the use of human atlatlists, which introduce uncontrolled biomechanical variation. We redress this problem through the construction and use of an atlatl launch machine, which provides unprecedented experimental control over mechanical variables, and hence unprecedented experimental objectivity and reproducibility. Using the machine, we test the Range and Precision Hypotheses of atlatl weight impact. Statistical analyses of data from 350 mechanized launches indicate that compared to the unweighted atlatl, the weighted atlatl typically has a lower range but greater precision, when all variables except the presence/absence of the atlatl weight are held constant. These results begin to help resolve the atlatl weight debate and have implications concerning the North American bannerstone, atlatl mechanics generally, and experimental atlatl research methods.