Joe Nathan and the modern closer

Once upon a time, teams would use the bullpen’s best pitcher most often, calling on their top reliever whenever crucial, game-changing situations arose and frequently leaving them in for multiple innings. Throughout the 1950s, 1960s, 1970s, and even into the 1980s, elite relievers like Rollie Fingers, Goose Gossage, Bruce Sutter, Hoyt Wilhelm, Kent Tekulve, Sparky Lyle, Mike Marshall, Gene Garber, Tug McGraw, and Dan Quisenberry regularly threw more than 100 innings per season.

That group of 10 relievers combined to throw 100-plus innings out of the bullpen 63 times, with each of them doing so in at least four seasons. Fingers averaged 110 innings per season over a dozen-year span. Quisenberry topped 120 innings in five of his first six seasons as a closer. Gossage topped 130 innings in each of his first three seasons as a closer. Tekulve had 100-inning seasons in his 20s, 30s, and 40s. Marshall amazingly threw 179 and 203 innings out of the bullpen in back-to-back seasons.

There are plenty of other examples beyond those 10 pitchers, but the point is that great relievers used to frequently work multiple innings per appearance and top 100 innings per season. That all began to change in the 1980s, thanks in large part to Dennis Eckersley‘s success as a one-inning closer, and the shift was essentially complete by the early 1990s. These days nearly all closers are held back strictly for “save” situations and asked to work just one inning in the vast majority of their appearances.

Whereas for many years the best relievers threw the most innings, the current standard for bullpen management often calls for many of the best relievers to throw the fewest innings because of strict usage patterns that revolve around a statistic. Baseball’s all-time leader in saves, Trevor Hoffman, has never thrown 100 innings in a season and hasn’t topped 80 innings since 1997. It’s no coincidence that his career high of 90 innings, set in 1993, came in the only season that he wasn’t a closer.

Hoffman, Mariano Rivera, Billy Wagner, Troy Percival, Roberto Hernandez, Jose Mesa, Todd Jones, Jason Isringhausen, Armando Benitez, and Bob Wickman are the “active” relievers with at least 250 career saves. Those 10 relievers have combined for a grand total of one 100-inning season, and that came from Rivera when he was a setup man for closer John Wetteland (64 innings) in 1996. Not only don’t current closers throw 100 innings, they rarely even get to 80.

Joe Nathan has been as good as any reliever in baseball since joining the Twins in 2004, but threw just 72, 70, 68, and 72 innings during his first four years as closer and is on pace to log a career-low 67 innings this season. His usage is almost entirely dependent on how many times the Twins lead by one, two, or three runs in the ninth inning, because Ron Gardenhire holds him back for one-inning save situations except for the rarest of circumstances.

In years past, the strength and depth of the Twins’ bullpen made Nathan’s strict usage more annoyance than hindrance, because there were typically several quality setup men to put out fires in non-save situations. However, with the relief corps now struggling following Pat Neshek‘s season-ending injury Gardenhire’s refusal to use his best reliever in crucial non-save situations has already hurt the team and will continue to do so.

We’ve seen Nathan left unused in the bullpen in tie games while lesser relievers take losses, because Gardenhire was holding him back for a “save situation” that never arrived. We’ve seen Nathan used for a dozen or fewer pitches before several lesser relievers were called upon to throw multiple innings, eventually coughing up runs. And we’ve seen Nathan used in mop-up situations because he “needed work” after not being used in crucial, non-save spots during the previous few games.

Not only have Matt Guerrier and Brian Bass each thrown nearly 50 percent more innings than Nathan, Juan Rincon has gotten about 10 percent more work. Even Jesse Crain has essentially gotten the same workload as Nathan despite coming back from significant shoulder surgery and not being effective since 2006. Beyond the seemingly counter-intuitive innings distribution, Nathan has also been allowed to throw just 16 pitches per appearance:

 Brian Bass 34
 Matt Guerrier 22
 Bobby Korecky 20
 Juan Rincon 19
 Jesse Crain 17
 Joe Nathan 16
 Dennys Reyes 8

Except for situational left-hander Dennys Reyes, every reliever on the team has been allowed to throw more pitches per appearance than Nathan. Meanwhile, he’s by far the best reliever on the team and makes more money than the rest of the bullpen combined. So why is the team’s best, highest-paid reliever used the least? Because while Gardenhire likely would take pride in replying “no” if asked whether he manages by numbers and stats, in reality he does just that.

There’s no indication that Nathan is incapable of throwing more than 70 innings per year or 16 pitches per appearance, but because his role revolves around the “save,” Gardenhire refuses to use him in non-save situations. Not only is that letting stats impact the way you manage, it’s letting the wrong stats impact the way you manage. If not for the fact that one counts as a “save” and the other doesn’t, would a three-run lead in the ninth inning be seen as more important than a tie game in the eighth inning?

For better or worse, limiting closers to 60-70 innings has become the standard throughout baseball, so Gardenhire certainly shouldn’t be expected to completely buck that trend by having Nathan throw 100 or 120 innings. However, there’s no reason why Nathan should be lifted after just five or 10 pitches when the game is still in doubt (which has happened twice in the past week) or go unused altogether while Crain, Guerrier, Rincon, Korecky and Bass work multiple innings when one run can lose the game.

Nathan has converted 14 of 15 save chances this season, but has pitched in just seven of the Twins’ other 37 games. Beyond that, he’s thrown more than 20 pitches in just four of the team’s 52 games overall. Pushed into more prominent roles following Neshek’s injury, the non-Nathan relievers have combined for a 4.77 ERA and 33-to-30 strikeout-to-walk ratio over 54.2 innings without Neshek around to put out non-save fires. All of which was my fear when analyzing the impact of Neshek’s injury three weeks ago:

Neshek was clearly second only to Nathan in terms of importance to the bullpen and was probably a lot closer to Nathan than most people think. His injury pushes Guerrier, Rincon, Crain and Reyes into expanded roles, and the danger is that it may expose several of them by taking them out of comfort zones and stretching their ability.

Guerrier can do the job in the eighth inning if asked but won’t be as good as Neshek, and moving him into the top setup job creates a hole in the sixth and seventh innings. Rincon, Crain, and Reyes can take on more responsibility and capably fill in those gaps, but they won’t be as effective as Guerrier, and putting more on their respective plates takes away some of the workload-related luxuries that fantastic depth previously provided.

Non-Nathan relievers won’t continue to post an ERA near 5.00 all year, but Rincon has gotten progressively worse each season since 2004, Crain is struggling following shoulder surgery, Reyes has fallen apart when asked to face more right-handed batters, and Korecky and Bass are marginal major-leaguers. In other words, without Neshek around, the Twins’ bullpen is far from a strength except for the fact that they still have one of the best relievers in all of baseball. It’d be nice if they used him.

You can also learn about all our free newsletter options.

Comments (1)

  1. Submitted by Tony Wagner on 05/29/2008 - 05:52 pm.

    One big argument which you didn’t address (and I’m not necessarily making the argument, but it is intriguing) is that modern relievers are more effective in their short, defined usage patterns. Check out this average of tOPS+ for “Close and Late” situations by decade:

    Decade tOPS+ Close and Late
    2000s 92.6
    1990s 94.2
    1980s 96.7
    1970s 99.8
    1960s 99.7
    1950s 100

    Hitters have been faring progressively worse in “Close and Late” situations (when there is most likely a non-mopup reliever or quality starter on the mound) as far back as the Baseball-Reference data goes (1956). This coincides with the trend toward increased reliever specialization. (There is also a trend for general improvement among relievers as compared to starters over that time, but that obviously lumps in mop-up guys, modern 5th starters, etc., so it’s a little more murky.)

    So why is this? Starting pitchers might be getting worse overall, making relievers look better by comparison, but bad starters generally shouldn’t be on the mound by the time the game is “Close & Late.” Hitters might be better/stronger than they were back in the day — but relievers will still have to face them, correct? And with expanded rosters and increased specialization, there should be more good hitters getting matched up with relievers late in close ballgames.

    These differences in the modern game are also some of the reasons why it might be apples-to-oranges to compare closers of different eras. As good as Fingers, Gossage, Sutter, etc. were, they never consistently posted the kind of run-prevention numbers that Nathan and Rivera have been posting either. Perhaps a modern closer could handle the greater workload, but his performance might suffer too. And “lesser” modern closers, guys like Wickman or Jones, might drop off the map altogether (I don’t think there is a good corollary for those guys in previous eras).

    So in regards to “elite” closers, what’s better – a 150 ERA+ over 110 innings, or a 250 ERA+ over 70?

Leave a Reply