[andyferris]

Good point. We need the mean, “any 1 of 10” and the “all 10 of 10” success rates in the metrics, so we can estimate reliability (the last one).