Based on the variance per run and the variance due to the 100-question set, I calculated the ranking range for 67 models and ...