I only just noticed this addition of this sentence to the conclusion, which I had missed in earlier review:
Our results suggest that using the minimum estimator for
the true run time of a benchmark, rather than the mean or median, is robust to
nonideal statistics and also provides the smallest error.
The use of the phrase true run time here is misleading, implying that we're suggesting the use of the minimum as a summary or test statistic.
I only just noticed this addition of this sentence to the conclusion, which I had missed in earlier review:
The use of the phrase
true run timehere is misleading, implying that we're suggesting the use of the minimum as a summary or test statistic.