Oddities in NL Bullpens

Yesterday Jack wrote about the horrible Diamondbacks bullpen. While the crew got a boost last night by facing the Astros lineup for four innings, they still hold the worst ERA and FIP of any NL bullpen. With defense independent figures like that the Diamondbacks bullpen might actually be this bad. They could make changes, and perhaps some of their pitchers improve during the next five months, but the process and results correlation does not bode well.

What about the best bullpens in the NL? Are they built for success? Are the worst non-Arizona bullpens as bad as their ERAs indicate? The answers are probably not and no, but the questions led me to find some odd numbers throughout National League bullpens.

The Mets bullpen has thrown 97 innings, second most in the NL to the Pirates. That amounts to more than 3.2 innings per game, a rate that the unit probably cannot sustain. Even the 2009 Dodgers, the only NL playoff team with a bullpen that threw more than 500 innings, averaged under 3.2 innings per appearance. As expected, the bullpen FIP, 3.78, is a bit higher than the ERA, 2.60. Of all the NL teams with a bullpen ERA under 4, only Atlanta’s FIP outpaces its ERA. Of the other five, only one, San Diego, comes within even 1.00 of its bullpen ERA.

The Mets and the Pirates bullpens actually have a connection. They both boast high strikeout rates and low walk rates. The Mets relievers have struck out 22.2 percent of batters faced and have walked 13 percent, while the Pirates have struck out 21.2 percent of opposing hitters and have walked 11.3 percent. Pittsburgh actually has a better bullpen K/BB ratio. Difference, though, comes on the long ball. The Mets bullpen has allowed six home runs, while the Pirates unit has allowed double that. The Pirates bullpen has also allow far more hits and sport a .331 BABIP, almost .060 points higher than the Mets.

Thanks to Ubaldo Jimenez and Aaron Cook, the Rockies have let their bullpen rest for two games this season. In games they’ve actually pitched they’ve averaged just under 3.2 innings per appearances, but because of the complete games the unit averages a hair over 3.1 innings per game. The bullpen FIP, 3.53, is no great shakes. In fact, it ranks below Florida’s bullpen FIP, 3.38, though Florida’s bullpen ERA sits at 4.46, eighth in the NL. Combined with a .270 BABIP, Colorado’s bullpen success seems unsustainable. Having pitchers who go deep into games, though, will help mitigate a correction.

While Florida’s bullpen ERA sits above its bullpen FIP by a considerable amount, 1.08, the oddest NL bullpen unit so far is easily Washington. The ERA isn’t so bad, just 4.05. In fact, that’s good for seventh best in the NL. When I first got a look at the bullpen peripherals, though, I thought the unit would have an astronomical FIP. In 91 innings the bullpen has allowed 12 home runs, as many as the Pirates. The Pirates have a better K/BB ratio, and so have a better FIP, 4.42, than the Nationals pen, 4.65. Yet the Pirates bullpen sports a far worse ERA, 5.83, probably due to the BABIP discrepancy. Washington’s bullpen has allowed a .275 average on balls in play.

We know that bullpens are the most volatile aspects of major league teams, so oddities like this can be expected when our sample is limited to a month of data. No bullpen has crossed the 100-inning mark, though Pittsburgh should tonight. During the next few months I expect that we’ll see fewer of these oddities. For now, though, we can get a better idea of why a bullpen has performed so well or so poorly, even if it is not a reflection of how they’ll perform from May through September.