As participants on multiple review panels and scientific councils, we have heard many lament researchers' reluctance to take risks. Yet we've seen the same panels eschew risk and rely on bibliometric indicators for assessments, despite widespread agreement that they are imperfect measures.