evaluating relievers, especially short relievers, by any sort of rate stats over 1 season is potentially misleading. One bad outing can dramatically affect a reliever like Miller. Theoretically, someone like Miller can do his job perfectly 9 straight times covering 5 innings only to give up 3 BBs and a grand slam in his 10th without getting an out which would make his overall performance look terrible but is misleading.

As a matter of fact, in Miller’s case it is not theoretical. I figured out his work based on specific outings once but don’t have time now to revisit it. What I do remember is that his overall numbers are significantly affected (adversely) by a stretch of 3 or 4 mid-season outings when he simply could not get anyone out, and was incredibly wild. Given that he pitched just 43.3 innings all year, those few failures disproportionately skew his stats. Otherwise, he was generally pretty effective in doing what he was expected to do.