Monday, January 07, 2008

More Dominant Than Rivera? Don't Make Me Laugh

I write this post as a Yankee partisan, as someone who considers Mariano Rivera the greatest reliever in the history of baseball.

Tomorrow we find out the members on the 2008 HOF class. Jayson Stark, like every other baseball writer on the planet, wrote an article arguing for the inclusion of a number of players, especially Goose Gossage. I'm not going to comment on whether Gossage deserves admittance or not. Rather, I will focus on one of Stark's arguments in favor of Gossage being admitted.

Stark argues that Gossage "was the most dominating closer ever." Certainly the property of "dominating" is inherently nebulous, so it's hard to get a handle on the specifics of Stark's point. Stark decides to strengthen his argument for Gossage by comparing him to Rivera, since everyone knows Rivera is a sure-fire, no-brainer, first-ballot HOFer. Stark compares Gossage's first 11 full seasons as a closer to Rivera's 11 seasons since he became the closer for the Yankees:

"Want to pick a category? Be my guest. ERA? Gossage 2.21, Rivera 2.35. Strikeouts? Goose 8.54 whiffs per 9 innings, Rivera 8.09. Unhittability? Gossage 6.59 hits per 9 innings, Rivera 7.17."

Immediately two things struck me. First ERA alone is not entirely useful because it doesn't adjust for park effects and league differences. ERA+ is a much more comprehensive statistic in that regard. Number 2, Stark left out walks, which seems to be a pretty useful indicator of dominance. Perhaps he meant dominance is the sense of Nolan Ryan, who struck out an insane number of batters, but also walked a large number. So let's focus on ERA.

I decided to run the numbers. Not exactly sure which years Gossage was using, I picked the first 11 seasons Gossage served as a reliever (1975, 77-86). I took his innings and runs and calculated ERA. Since we both got 2.21, we're probably using the same years.

I then did the same thing for Rivera. Working with 97-07, I came up with 2.12. Since Stark calculated Rivera's ERA during his 11 seasons as a closer at 2.35, I thought my numbers were wrong. Then I realized that 2.35 is Rivera's career ERA, which includes one season as a bad starter and one as a dominant reliever. So basically Stark compared Gossage's ERA from his peak seasons with Rivera's career ERA. That's disingenuous if you ask me.

Next I looked at ERA+. It should be noted that Rivera's career ERA+ (194) is the highest ever for any pitcher with a substantial number of innings under his belt, and will be highest of any pitcher who threw 1000 innings after this season barring injury. But just using his 11 seasons as a closer, his ERA+ was 222. Gossage's? 198. That's 24 points higher for those keeping score at home. Over the course of a career that's the difference between Johan Santana and Rick Aguilera.

Next I decided to look at each pitchers defense independent ERA (DERA). BP's stat attempts to sever defense from pitching and list the ERA each pitcher would have had in front of an average defense, as well as account for league and park. Gossage's ERA over those seasons was almost exactly 3. Rivera's was 2.26. That's almost .75 of a run or over a career is the difference between Christy Mathewson and Sparky Lyle.

Stark's one legitimate point is that Gossage threw more innings over those seasons. Rivera only threw an average of 71 innings, while Gossage pitched almost 95. That 25 inning advantage is significant and definitely is an edge to Gossage. But it's not an edge that can make up the extremely large differences in ERA+ and DERA.

In fact another BP stat may be of assistance. BP uses a statistic called Pitching Runs Above Replacement, which measures how many runs a pitcher saved over a replacement player. Over those 11 seasons, Gossage saved an average of 60.2 runs above replacement. Rivera saved 67.5 runs. Per inning, those numbers aren't even close with Rivera saving .953 runs and Gossage .634. But since total numbers take into account the innings pitched, Rivera's advantage shrinks to around 7 runs a season. But 7 runs a season is not a trivial amount.

Look, it's not the Mariano Rivera Hall of Fame, so Gossage doesn't need to be Rivera's equal to make it. But these ridiculous (and somewhat misleading) comparisons do not serve to advance Gossage's cause.

Update: I did a little more research and realized that Rivera's Hits/9 and SO/9 were also based on career numbers. But in those categories even Rivera's closer years do not stack up to Gossage's first 11 seasons as a closer.

I also decided to look at Billy Wagner, who has been very dominating as a closer, even according to Stark's criteria. Taking Wagner's 11 seasons as a closer, he has a 11.8 strikeouts per 9 innings and allowed only 6.2 hits per 9 innings. His ERA, however, is only 2.40 (189 ERA+), so Gossage wins in that category. But Wagner struck out more than three batters per nine innings, and gave up .3 less hits. It's not that clear that Gossage was a more dominating closer than Wagner.

No comments: