Previous entries:

Bull by Committee

A committee is an animal with four back legs.

—John "Max" le Carre

2003 became a litmus test for the bullpen or closer by committee. In the offseason the Red Sox hired Bill James and allowed their itinerant closer, Ugueth Urbina, leave via the free agent route. Boston then adopted James’ theories with regards to relief pitching and voila, it was all around the hot stove circuit that the Sox were employing a bullpen by committee.

Of course, the source for all the Jamesian theorizing, an article titled “Valuing Relievers” in The New Historical Baseball Abstract, does not mention anything about a bullpen by committee or a closer by committee. Actually, James doesn’t even user the term closer. He prefers “ace reliever”. And the point of the article is to arrive at a means to value and to maximize the ace reliever’s contribution. His observations I have already documented in the Gagne section above.

Anyway, James vociferously defended the Red Sox plans as having nothing to do with a bullpen by committee. But the press saw the Red Sox picking up veteran relievers like Chad Fox, Ramiro Mendoza, and Mike Timlin and labeled the Boston bullpen a “committee”.

So who was right? Well, the Red Sox did use ten different pitchers, including two in their rotation (Wakefield and Fossum), to save games. Is that unusual? Actually, it hasn't happened since the 1995 Detroit Tigers traded Mike Henneman (1.53 and 18 saves) to Houston for Phil Nevin just after the trade deadline and used nine different relievers not too effectively to finish out the year. Before that it hadn't happened since 1987 when Baltimore and Cleveland both let 10 relievers pick up saves. Here's the complete list (The first to do it was the Tigers in 1909 when every pitcher on the staff but one—he only appeared in two games—picked up a save, posthumously):

By the way, the only team to have more than 10 pitchers record a save since Bruce Sutter revolutionized the closer role was the 1979 Dodgers. Bobby Castillo led them with 7 saves (20.6% of the team's 34 saves) and a 1.11 ERA in just 24.1 innings. Don Sutton (1 save), Jerry Reuss (3), and Bob Welch (5 saves) were even in the mix. Fourteen different Dodgers finished games; eight finished 10 or more games. Now, that's a bullpen by committee.

Castillo's seven saves comprised just 20.6% of the Dodgers saves that year. That was the all-time low for a team saves leader…until this year. The Tigers' save co-leaders Franklyn German and Chris Mears recorded just 5 saves each. The Tigers had a woeful 27 saves as a team. Even so, that means that the saves leader for Detroit only registered 18.52% of the team's saves, the lowest ever. Here's the list of all team "closers" who recorded 30% or less of their team's saves:

OK, so back to the Red Sox. They certainly shared the saves around, but does that mean that the constituted a bullpen by committee? What is a bullpen by committee anyway?

Well, it seems to me that a bullpen by committee would not only share the saves among many pitchers, but they would share them pretty equally at least among the better relievers. What if we looked at the numbers for the pitchers who finished second and third in team saves.

Here is a table of the men who finished second in saves on their respective teams with the highest save totals all-time (co-team leaders are both listed):

Now here are the men who finished third (or fourth) on their teams with the most saves:

In 2003 the Red Sox’s saves leader was mid-season acquisition Byung-Hyun Kim with 16. Next was Brandon Lyon at nine, and then Chad Fox at three. Actually, if you look at the Red Sox game log for 2003, it’s pretty apparent that they were not employing anything like a bullpen by committee. They just had a succession of unsuccessful, putative closers or as James puts it, relief aces.

The 2003 season started with a bang for the Red Sox pen. Game one, March 31 at Tampa Bay, the Sox led 4-1 going into the bottom of the ninth. In what was potentially a save situation, Boston turned to Allan Embree. When he relinquished two runs on a Terry Shumpert home run, Chad Fox was summoned. Fox was acting as James’ relief ace, coming in with a one-run lead. It was also a save opportunity. Fox lost the game on a three-run, two-out, walk-off home run by Carl Crawford. And Grady Little started to stray from the relief ace construct that James laid out though Fox remained the closer for a short time.

In the next game Bobby Howry was given an 8-6 lead in the bottom of the eighth and he quickly lost it, giving up two runs in one-third inning. The Red Sox did win, 9-8 in the 16th, however, and Brandon Lyon pitched three solid innings to pick up the win. Howry did have a save opportunity (if he had pitched the final two innings and kept the lead). However, he was not acting as James’ relief ace since they are only employed with a one-run lead, in a tie ballgame, or when the team trails by one.

Game three Boston lead 7-5 in the eighth. Fox came in to record his first save though this was not technically an opportunity according to James in which to use the relief ace. Game four was a blowout and Fox rested.

The Sox then went to Baltimore and won a one-run game, 8-7. However, there was no save or relief ace opportunity because Boston led 8-1 going into the bottom of the seventh and 8-3 going into the bottom of the ninth. Ramiro Mendoza gave up four runs in the ninth. Again Fox was rested.

The wheels started to come off the relief ace concept in game 6. Boston and Baltimore were tied, 1-1, as the bottom half of the ninth began. Boston turned to Chad Fox as the relief ace in a non-save opportunity but an ideal opportunity according to James’ relief ace criteria. Fox spelled the always bubbly Pedro Martinez and quickly relinquished a one-out walk to B.J. Surhoff. Conine doubled, and with first open, Gobbons was intentionally passed to load the bases (a strategy probably not advocated by James). Fox went 3-0 to Miguel Batista, then worked a full count, but finally walked in the winning run.

Their next save opportunity (or relief ace opportunity) did not come for six games. On April 13, the Red Sox led the Orioles 2-0 at home. Starter Tim Wakefield came in with the two-run lead in the eighth and earned the two-inning save. However, it should be pointed out that this was not a relief ace opportunity. Fox had only been used for one scoreless inning during these six games, in a blowout game apparently to get a little work.

The next game, Boston led Tampa 5-1 in the top of the eighth. Ramiro Mendoza quickly allowed two runs and left with no outs, two men on, and a 5-3 lead. Mike Timlin let the Devil Rays toe the game on a Marlon Anderson one-out single, but stayed in the game and earned the win after the Sox scored a run in the ninth. This was technically a relief ace opportunity (after Tampa tied the game). Chad Fox was rested but was not used. It seemed that he remained the closer but that the concept of the relief ace was no more. It seems odd given that ESPN chose to criticize the Jamesian bullpen approach as “Closer by Calamity and Closer by Catastrophe” in the recap of a game in which James theories were not ever employed.

Even odder, Fox was used in the next game with Boston trailing 4-2 to lead off the eighth. Fox pitched a scoreless eighth and then earned his first win as the Red Sox scored 4 in the bottom of the eighth. Lyon came in for the save.

Fox then pitched a mop ninth three days later in a 7-3 win over Toronto. On April 20, the Red Sox and Jays were tied 5-5 in the top of the eighth at Fenway. Mike Timlin was used for the last two inning and the Red Sox won 6-5.

On April 22 in Arlington, Fox was then entrusted with a one-run lead (5-4) with one out in the eighth, after Timlin allowed three runs. He earned his second win with a 1.2 hitless innings. On April 25, Fox earned his third and final save in a Boston uniform, holding a three-run lead with two out in the eighth and two men on.

Meanwhile Lyon was being used to finish the blowout games and carried a 1.64 ERA through April 24. On April 27, Boston won a game 6-4 over the Angels in 14 innings. They had led 4-2 in the bottom of the eighth and turned to Lyon in the save non-relief ace opportunity. Lyon gave up a run and Fox was brought in for the ninth. He allowed the tying run in one-third inning, and that was the end of Fox as the Red Sox closer.

The Sox had long since abandoned the relief ace concept. On April 30, they had just tied the Royals 2-2 entering the eighth. Ramiro Mendoza who had pitched horribly to that point was left in the game and allowed two runs to score. The two runs scored after Lyon replaced Mendoza. The Sox won with three runs in the bottom of the ninth.

On May 1 Brandon Lyon was anointed the official closer with a save, his second, in a 6-5 win over the Royals. Lyon remained the closer pretty much until he handed the job over to Kim in July. Kim remained the closer for the rest of the year aside from a 4-inning save by Casey Fossum, a three-inning save by Bronson Arroyo, and a save by Mike Timlin in relief if Kim in the ninth on September 19.

So there you have it. The Red Sox had nothing close to a bullpen/closer by committee. They did follow James’ tenets for a short time but quickly abandoned them. From mid-April on they employed the same strategy as most any other team; they just got poor performance from the closer role.

When I think about bullpen by committee, I see Cleveland in 1993, Toronto in 1985, and LA in 1979. The Indians had a good group of relievers (Eric Plunk, Derek Lilliquist, Jeremy Hernandez, Jerry Dipoto, and Bill Wertz). All had a park-adjusted ERA between 20% and 92% better than the league average. None of them amassed more than 15 saves, but the first four had at least 8 each (and career highs for Lilliquist and Plunk). Also, each of the first four finished between 22 and 40 games.

The Blue Jays in 1985 had four relievers who recorded between 8 and 14 saves each and finished 19 to 51 games (Bill Caudill, Tom Henke, Jim Acker, and Gary Lavelle). Their top five relievers had park-adjusted ERAs between 31% and 109% better than league average (between 2.03 and 3.32).

The '79 Dodgers, I discussed above. They were led in saves by Bobby Castillo (7 with a 1.11 ERA), followed by Dave Patterson (6 with a 5.26 ERA), Bob Welch (5 with a 3.98 and 12 starts), Lerrin LaGrow (4, 3.41 ERA), Jerry Reuss (3 with a 3.54 ERA and 21 starts in 39 games), three others with two saves, and three with one save. Of the 8 pitchers on the staff that started at least 10 games, five appeared as relievers. They may have transcended the bullpen by committee mold and may have anachronistically approached the old John McGraw teams at the start of the twentieth century. McGraw solidified the use of relief ace, but would cannibalize his starters (Joe McGinnity, Christy Mathewson) to accommodate it.

Also, Sparky Anderson and his quick hook were highly influential in the history of the bullpen-by-committee approach. In his nine seasons in Cincy, he had five in which the pitcher who finished third in saves amassed at least 7. And only in five seasons did he have a reliever record 20 or more saves, even though the Reds had three 100-win seasons and just one with fewer than 88 wins during his reign.

Finally, the 2000 Baltimore Orioles should be a cautionary tale for anyone considering the bullpen by committee route. They started the year with a rotation of Mike Trombley, Buddy Groom, and Mike Timlin. They blew 22 of their first 49 save opportunities. The trio finished the year with 19 total saves and ERAs between 4.12 and 4.89. Finally, rookie Ryan Kohlmeier was given the job. He pitched well (2.39 ERA with 13 saves in 25 games), but fell apart in 2001 (7.30 ERA with 6 saves) and was out of baseball. The problem, as with the Red Sox's original configuration, was that the personnel was not strong enough or deep enough to fill out the entire bullpen and act as the closer as well.

Where To, Buddy (Groom)?

The excessive increase of anything often causes a reaction in the opposite direction.

—Plato "Shrimp"

All of this non-standard use of relievers got me to thinking about the state of relief pitching in 2003. There seemed enough clues to indicate that something related to relief pitcher use was afoot around baseball. With Detroit's all-time low save co-leader (by percentage of team saves), Boston's sharing the saves among ten pitchers, and the White Sox's splitting the closer job among three pitchers there seemed to be a change in the way relievers were used.

The closer role seems to reflect the economy during the current Bush administration. The elite are excelling (e.g., Gagne, Smoltz), the poor are floundering (the Tigers), and the middle class are getting squeezed (Ugueth Urbina, Armando Benitez, Billy Koch) and everyone is looking for a bargain (Rod Beck, Joe Borowski). Maybe I'm overstating the case. How do we know that there's anything more than the normal cyclical changing of the guard for a number of teams' closer role?

Well, here's a table for every year since the save stat became official of the average percentage of the team save leader's save total to the team's total saves, the yearly change in the percentage, and the average save total per "closer":

Note that greatest dropoff in leader-to-team save percentage was last year, the only year with a greater than 10% decrease. This came after a pretty steady increase following Dennis Eckersley's role redefining season in 1988 (the era that I call the post-modern closer era).

And it’s not as if the Tigers' closer issues skewed the data. There were a number of teams with low percentages:

Another indication that change is afoot is that the standard deviation from the average leader-to-team save percentage shot up to the highest in thirteen years and the second highest in the save era, especially odd since one would expect the standard deviation to drop as the majors expand (because of the additional teams being averaged):

Here’s one more illustration, expanding a table that I created in the Nineties section. It contains the percent of team save leaders who amassed a certain percentage of the team’s total saves. For example, the 100% column tells you the percentage of all “closers” who registered all of their team’s saves. Note how each bracket is increasing especially into the late Nineties and early 2000s until 2003:

So what's going on? Well, one thing is that teams are dramatically cutting payroll. That makes them question if paying perennially mediocre closers like Ugi Urbina and Bill Koch four million dollars is being fiscally responsible (or if it’s preferable to line the owners’ pockets instead). The impecunious A's seem content to mine for undervalued closers and then let them go when their price tag goes up. They have had four closers in the last five seasons (Billy Taylor, Jason Isringhausen, Koch, and Keith Foulke) and will have a new one in 2004 (Dusty Rhodes?).

I think that with the offenses back in obeyance, managers went back full bore to the tried and true closer role, which is if there is a save opportunity in the ninth, bring out the closer. However, as the media and the fans became more sabermetrically informed, they began questioning a strategy that left supposedly the best closer in the pen in the seventh and eighth when the game may be on the line. Oftentimes, once the ninth inning rolled around the save opportunity had already evaporated. So the cresting wave of one-inning closers broke and fell back this last season.

Another problem was the quality of some of the closers. They too often, like Greg Brady’s Johnny Bravo, merely fit the suit. Jose Mesa in Philly is a perfect example. He had a few years with high save totals and sub-3.00 ERAs, but given his wildness never seemed too secure on the mound. All of that came back to haunt him in his deplorable 2003 season. Again why pay someone four million dollars to come into a 4-1 game in the ninth and then walk the bases full while striking out the side?

So where to next? It seems that 2003 was not a one-season anomaly and rather a shift in reliever usage. Relief pitching strategy seems to change every ten or so years. It’s like they say that there is a war every twenty years or for each generation: baseball’s generations just cycle a bit more quickly. The current usage pattern started with Eckersley’s 1988 and became entrenched around 1990. Perhaps the offensive surge in the mid- to late-Nineties, two rounds of expansion, and/or the expansion of the middle relievers’ roles extended its shelf life.

With just a handful of elite closers, teams seem content to muddle through by jury-rigging the closer role. Indeed, many clubs seem to dissemble and re-assemble a bullpen almost every offseason. That’s what a surfeit of free agent pitchers will do for you.

I may be wrong and Gagne’s big 2003 season may be the clarion call back to the clearly defined, ninth-inning-only closers, but I doubt it.

One thing that would move the process along would be to redefine the outdated save rule by eliminating the automatic three-inning save and the automatic three-run-lead save. They could also make the hold stat official. Why not credit a reliever who holds a lead at an important junction. A hold may be more important to a game than a save, at least under its current configuration. If a closer’s saves and holds were citable in arbitration and free agency cases, then the closers would be more amenable to coming in with a one-run lead in the seventh.

One thing I think will probably not be tried again for some time is the bullpen/closer by committee. Even though the Red Sox never really employed it, they gave the bullpen by committee a bad name. A manager would be vilified in the press and by the fans if he chose to use one any time soon. The preferred method now seems to give a series of relievers the closer role on a trial period. If one succeeds, great, ride him until he fails and then get someone else.

Finally, here are the appearance and saves leaders for the decade so far:

Here’s an update for pitchers per role for the Aughts including the 2003 season (RA=Relief Appearances; P/G=Pitchers per game; #P=Number of pitchers in total):

Here’s the breakdown for starters:

Relief pitchers:

Swingmen: