Archaic projectile points from the Andean Altiplano exhibit a curious trend of increasing size over time, in contrast to a well-documented size reduction throughout North America. Although a number of hypotheses exist to account for decreasing projectile-point size, there are currently no explicit explanations for increasing size. We consider several hypotheses and interrogate two techno-economic hypotheses. We posit that increasing point size compensated for lost dart momentum or accuracy that resulted from the shortening of atlatls or atlatl darts as wood became increasingly scarce on the tree-sparse Altiplano. We evaluate these hypotheses using a replicated Andean atlatl system in ballistic trials. Contrary to expectation, results show that point enlargement significantly reduces penetration depth, allowing us to confidently reject the momentum hypothesis. Point enlargement, in contrast, tentatively correlates positively with accuracy. Our experiment further shows that camelid bone is an effective and economical alternative to wood for atlatl production. Despite suboptimal lengths, camelid radioulna atlatls have a convenient morphology that requires low production time, which helps explain empirically observed camelid bone atlatls from the Andean highlands. More generally, our observations lead us to consider that central tendencies in archaeologically observed projectile-point size may reflect a trade-off between penetration and accuracy.