It seems that we are not willing to give up the intuitions that (1) works of fiction are free from the constraints of historical truth and (2) historical inaccuracies sometimes count against the artistic value of works of fiction. Christopher Bartel calls this the puzzle of historical criticism. I argue that this puzzle extends beyond historical facts. While it is especially salient that historical accuracy at times appears relevant to the evaluation of fictional works, such relevance appears to be a feature of facts in general. I then propose a partial strategy for resolving the puzzle.