There exists a script to analyze the results of a benchmark task and check to see if it has diverged significantly from older results. This script has been commented out for a while and is no longer in use. We should remove it from the codebase to reduce any confusion people may have when looking at it.