The “Quality Appearance”

Baseball these days can seem overrun with statistics – Batting Average with Runners in Scoring Position and Two Outs, for example – yet records for pitchers all seem to me fundamentally flawed. Which is why I am proposing a new, more meaningful, one, which I will get to in a minute. First, however, a recap of some existing flaws.
For years, Won-Lost has been the gold standard for pitchers. The first problem, though, is that its validity is largely limited to starting pitchers, the only ones who routinely have some control over a game’s outcome. The emphasis, however, must be on “some.” We are all familiar with the “hard-luck loser,” who pitches eight innings, allows one measly run, yet suffers a 1-0 defeat. (Last night the Indians’ Josh Tomlin surrendered a first-pitch, first-inning opposite-field home run to the Twins’ Miguel Sano and suffered that fate.) The flip side is the lucky winner, who gives up five runs in five innings, yet records a win because his teammates have clobbered the opposing pitcher for more. Sometimes this averages out over the course of a season, but sometimes it doesn’t. Then there is the disparity among offenses. A pitcher for a hitting-starved team could give up three runs a game and finish with an 8-12 won-lost record; while his counterpart for a powerhouse could allow the same number of runs and wind up 12-8.
The next most recognized statistic is Earned Run Average, which would show the equivalence of the two pitchers in the above example. This statistic in theory should also work for relief pitchers as well as starters, and it has the advantage of not being dependent on the hitting success of the pitcher’s team. Its weakness, though, is its failure to account for situational pitching, which is crucial in evaluating a relief pitcher’s value. By “situation” I mean both the stage of the game – is the scored 7-2 in the eighth inning, or 3-2 in the ninth? – and the situation within the inning. A reliever who enters the game with two outs and the bases loaded and gives up a triple is a failure, yet his ERA will go down if he gets the next batter out. Conversely, the pitcher who let those three runners reach base while getting two outs will see his ERA soar, even though no one scored while he was pitching. The other problem with judging a reliever by his ERA is that one bad outing can skew it, because relievers pitch so few innings compared to starters. (Last night the Twins brought in Matt Belisle with one out, one on in the eighth inning of a 1-0 game despite an unheard-of ERA north of 11.00, precisely because that figure bore little connection to his perceived reliability.)
I have written elsewhere about the Save, the least useful number of all. 1) It only is relevant for the usually one pitcher per team who gets to pitch the ninth inning. 2) The number depends on his team’s success rate – only the winning team can record a save. 3) The criteria are too loose to be meaningful: a reliever can give up two runs in the only inning he pitches (an ERA rate of 18.00) and still get a save.
Because of the overemphasis on Saves, some baseball writers in 1986 came up with the Hold as a way of recognizing the effectiveness of middle relievers, who don’t get Save opportunities. Although routinely included in game box scores, the Hold is not an officially recognized statistic and there is some confusion as to what qualifies. The first prerequisite is the existence of a Save situation – so right there, it incorporates the weaknesses of that stat. The second requirement is that the reliever hand over pitching duties to the next reliever with his team still in the lead – which means, similar to the Save defect, that a reliever could load the bases while recording only one out and earn a Hold despite doing a lousy job.
So, what would be a more useful measure of a pitcher’s – specifically, a reliever’s – performance? Say ‘hello’ to the QA – the Quality Appearance. A reliever earns a QA by recording three or more outs without giving up a run, including the runners he inherits and the runners he puts on base. Alternatively, if he enters a game with two or more men on base, and he records two outs, ends an inning and no one scores. That, after all, is what really matters – keeping the other team from scoring. This statistic has the advantage of applying whether a team is ahead or behind. After all, when a team is behind 3-2 late in the game, it is crucial to their chances of winning that they not give up another run; and the pitcher who holds that margin is doing just as important a job as the pitcher who is credited with a Hold. Yes, there are some flaws: a pitcher could lose a QA by giving up a run in a second or third inning of work, although the rule could be refined to cover that, if desired. And what happens to runners that a pitcher puts on base before he is relieved will depend on the relievers who follow. Not every good relief effort will be recognized, but many more will, and more fairly than either the Save or the Hold currently does.