How Not to Overshoot the Evidence in Historical Logic