<
>

FanGraphs: Closers worth less than you think

10/2/2010

As expected, the White Sox will continue to use Bobby Jenks as their closer this year, despite having a superior bullpen arm in Matt Thornton. For the past two seasons, Thornton has been a buried treasure, posting better numbers than the Southside’s closer. For that matter, his FIP (a defense-independent measure of ability) has been better than many other team's closers.

The White Sox aren't the only team that does not have their best relief pitcher throwing in the ninth. Kevin Jepsen of the Angels is a better pitcher than Brian Fuentes, and the Mariners' Brandon League could very well be better than David Aardsma. In one extreme case, the 2007 Indians continued to use Joe Borowski as their closer despite his bloated 5.07 ERA, while they reserved Rafael Betancourt for the eighth inning. Betancourt posted a minuscule 1.47 ERA and 2.22 FIP that year.

For fans that follow these teams, it's frustrating to see the manager continue to go to the well with these "proven closers" while neglecting to give these elite arms their fair shot at saving games. Well, the aggravation of fans aimed towards the manager in most cases is unnecessary. No, they aren't using their assets in the best possible way, but the impact isn't a big as you might think.

With the help of Tom Tango, FanGraphs provides a stat that measures the magnitude of every game situation called Leverage Index, or LI for short. Here's the gist, and this takes us back to what Sky shared about the Twins loss of Joe Nathan earlier in the month:

-The average situation (think of the starting pitcher's role) has an LI of 1.
-Closers appear in situations with an average LI of about 1.8, meaning runs allowed by closers are a little less than twice as damaging as the average run.
-Set-up men will see LIs in the 1.3 to 1.6 range.

So while closers on average pitch in more crucial situations, the set-up man pitches in some pretty important situations as well, and often more frequently.

For a practical example we'll use Jenks and Thornton of 2009 White Sox. What happens if we go back into 2009, only with Jenks and Thornton swapping roles? Jenks posted a FIP of 4.47 over 53 1/3 innings, with an average LI of 1.9. This made him good for 4 runs above replacement level, or about half a win. Thorton threw 72 1/3 innings, with a 2.47 FIP, with an average LI of 1.5. That made Thornton good for 26 runs above replacement level. If you give Thornton’s innings and leverage to Jenks, and vice versa, the difference in runs above replacement comes out to be about a single run. That's all.

Part of the extra value that a set-up man has is he isn't restricted to just the ninth inning with a lead. Many non-save situations are crucial to a team's chances of winning, and in fact, a lot of high leverage situations happen when a team is trailing by just one run -- a situation where a manager will almost never use his closer, even though keeping the score close gives his team a chance for a come from behind victory.

Non-closer relievers can be extremely valuable in these middle innings, high leverage situations, so it's not the end of the world if your team's manager chooses to employ an inferior pitcher to close out games. Thornton, Jepsen, and League will provide plenty of value, even if they don't rack up many saves.

Erik Manning is an author of FanGraphs.