Courts have generally disfavored evidence from recanting witnesses. This article examines the standards laid out in Berry v. State and Larrison v. United States that courts use when considering motions for new trials based on new evidence. It next explores some of the reasons courts have disfavored recantations. Recent cases involving DNA exonerations present useful lessons for evaluating recantations and weaken many of the reasons courts have used to reject such evidence. Because DNA evidence is not readily available in most cases, however, the current framework has led to incarceration of innocent defendants. Given these lessons, courts must find new means of assessing the testimony of recanting witnesses. Courts should adopt a modified version of the Larrison standard, which would require corroboration rather than proof of truth. Appellate courts should not apply a deferential standard of review to summary denial of motions for new trials based on recantations.