Print

Closer Usage Debate

September 17, 2008

            It has become a common belief in the sabermetric community (population, 1,742) that the modern use of the Closer is framed by the Save Rule.   I read that I was the first person to make this argument; I don’t know whether I was or not.  If somebody else wants that credit they can have it, and if you want to hold me responsible, that’s OK.   Anyway, the argument as I understand it is this:  that once it became accepted that the top relief pitcher’s job was to “save” wins, then baseball people accepted or simply assumed that the top reliever should be used in save situations.   “It’s his job to save the wins, therefore we should use him to get the Saves.”   The de facto definition of a Closer’s job became “to pitch the ninth inning, if possible, in every Save situation.”   Sometimes he can’t pitch the ninth inning because he needs a day off; sometimes he needs to pitch an inning in some other situation to stay sharp.  But the core of his job is to earn Saves.

            The sabermetric community (population, 1,741. . ..some guy in Idaho just resigned) tends to regard this as an irrational choice.   The best reliever should be used, they argue, in such a manner that it maximizes his impact on the team win total.   Sometimes the highest-impact situation is a Save situation—and often it is not.   Managers are wasting resources, so say the sabers, by using their best relievers in “Save” situations rather than high-impact situations.  

            There are a lot of dimensions to this argument that I’m not going to get into here.  What I am focused on here is one historical aspect of this debate:  is it true that the modern bullpen evolved as it did because of the Save rule?   Mike Emeigh is a highly esteemed member of the sabermetric community. . .I think perhaps he was a three-term mayor.   In a recent post to SABR’s statistical analysis committee, Emeigh argued that it is not true.    I wish I could quote that post precisely, but unfortunately I’m not that organized.   “A careful study of the record shows,” wrote Emeigh. . ..I am paraphrasing. . . “that the usage of closers evolved as it did in an effort to protect their arms and keep them healthy, rather than in an effort to maximize Saves.”

            “I totally don’t believe it,” I fired back.   “What evidence are you talking about?”  Mike responded as follows, and I will quote his e-mail here in its entireity.  

How teams actually have used their bullpens since the save rule went into effect. This is a long study, and it's hard for me to distill it down to an E-mail, but I will try.

Roy Face went 18-1 in 1959, amidst much media attention. Jerome Holtzman made the argument that Face wasn't really ‘that’ effective, because he blew a lot of leads and ‘vultured’ a lot of wins during that stretch; he was the loudest voice promulgating the argument that the ace reliever's job was to PROTECT a lead - to SAVE the game, if you will. Holtzman developed the first "save" stat, which the Sporting News published.

Holtzman didn't develop the idea that there was such a thing as a "fireman" whose job it was to put out fires, though. By the 1959-1960 time frame, many teams in fact had ONE reliever whose primary job was to do just that - who would pitch virtually all of the late-inning high-leverage situations. There had been isolated instances of such relievers before, but the mid-to-late 50s saw a number of them come forward, mostly in the NL: Face, Lindy McDaniel, Don Elston, Don McMahon, Clem Labine. Some teams had two (Narleski/Mossi in
Detroit, Brosnan/Henry in Cincinnati
, Staley/Lown for the White Sox) but most teams quickly settled on a primary guy. When Holtzman developed the idea of the save, it was intended to reward these guys. I doubt Holtzman ever envisaged a situation where teams would use multiple guys to close out a game on a regular basis; at the time it was almost always just one, or at the most two, pitchers.

When the save became official in 1969, most teams still only had one pitcher for whom closing out high-leverage situations in the late innings became the primary job. That role was fairly consistent from about 1965 through 1990, although there were some managers (Tanner in
Pittsburgh
being the chief one) who started using other pitchers earlier in the game. Most ace relievers would come into about 2/3 of their high-leverage situations in the ninth, and about 1/3 of them prior to the ninth.

Free agency came into place in 1976. Right after that, teams (a) started spending significant money on top players, including ace relievers and (b) started cutting back on their use in situations OTHER than high-leverage situations. This was true of top-level starters, as well - they started seeing their usage in low-leverage late inning situations go down. That doesn't appear to be a coincidence; the more you have invested in a guy, the more you want to protect that investment. It could be that teams were spooked by what happened to Catfish Hunter after 1975 and to Bill Campbell after 1976 - I really don't know. But it's easy to trace the decline in both starter and ace reliever "innings" to the 1977-1979 period. However, while "innings" went down, high-leverage appearances didn't change significantly. The ace was still getting the vast majority of the late-inning high-leverage appearances.

The first true one-inning closer was not Eckersley. Eckersley became Oakland's full-time closer in 1988, but he was still making a significant number of high-leverage appearances prior to the ninth (and would continue to do so throughout most of the rest of his career, albeit fewer and fewer as he got older). The first "real" last-inning closers were
Baltimore's Gregg Olson and Toronto
's Tom Henke, in 1991. Both were made one-inning closers primarily because of injury concerns, and Henke, in particular, was spectacularly successful in the role; he came into games only eight total times all season before the ninth that year (it helped that he had Duane Ward in front of him, too). It should be noted that when Henke went to Texas after 1992, he went back into what was a more traditional Closer's role and (a) had just one more good season and (b) saw his career end fairly quickly afterwards.

By 1992, other teams were getting on the bandwagon:
Boston, with Jeff Reardon; Milwaukee
, with Doug Henry; the Yankees, with Steve Farr, and the Cardinals, with Lee Smith (who actually had started moving in that direction late in 1991). Eckersley, in 1992, was still getting into 20 games before the ninth, whereas Smith made just six such appearances. Most of the pitchers involved were (a) the best options available and (b) not terribly effective after more than an inning or so. Smith and Reardon were on the downside of their careers, and Henry and Farr were not, shall we say, supremely talented pitchers. In 1993, Mitch Williams (another guy with questionable effectiveness) took it one step further; he didn't pitch in a single high-leverage situation before the ninth inning all season, the first Closer to do so. It took a few more years after that, but by 2000 the last-inning Closer was more often the case than not.

It seems fairly clear that the teams that were early adopters of the last-inning closer did so because their options were limited. The first guys that pitched in that role were (a) injured veterans, (b) declining veterans or (c) pitchers with questionable effectiveness beyond a single inning - in addition to Doug Henry and Farr, some of the other guys who fell early into a last-inning Closer role were Mike Fetters and Jose Mesa.

There is not a lot of evidence that the decision to go to a last-inning closer - or indeed, the change from fireman to closer - was driven by the save rule; Eckersley, for one, pitched a lot in situations other than save situations. The save rule became official in 1969, by which time the majority of teams were already using, or trying to use, one pitcher in the late-inning, high-leverage role. The changes in that role between 1969 and the early 1990s appear to be driven mostly by free agency and a desire to limit the workload of the higher-salaried pitchers, and the move towards a last-inning Closer appears to be a combination of making use of guys with a history of injury and/or declining effectiveness, and guys who could be effective for an inning but not beyond that. And the changes in bullpen usage outside of the closer role are mostly a transfer of lower-leverage innings from good pitchers to lesser pitchers, keeping the better pitchers available more often for the higher-leverage innings.

And from what I can tell, LaRussa was not the primary driver behind the move toward larger and larger bullpens. An early adopter, yes, but never really the chief innovator. From what I can tell, it was mostly Tanner, in the late 70s in
Pittsburgh
, and Davey Johnson, with the Mets in the mid-80s, who really started the trend.
--
Mike Emeigh
 

There are a great many points there, and some of them are certainly accurate and some of them may not be entirely accurate.   I was trying to pick through this minefield of arguments and decide whether his argument is convincing.    Let’s go through it line by line. . ..

How teams actually have used their bullpens since the save rule went into effect. This is a long study, and it's hard for me to distill it down to an E-mail, but I will try.

Roy Face went 18-1 in 1959, amidst much media attention. Jerome Holtzman made the argument that Face wasn't really ‘that’ effective, because he blew a lot of leads and ‘vultured’ a lot of wins during that stretch; he was the loudest voice promulgating the argument that the ace reliever's job was to PROTECT a lead - to SAVE the game, if you will. Holtzman developed the first "save" stat, which the Sporting News published.

It has limited relevance to the core issues of the debate, but this does not strike me as an accurate summation of the Roy Face experience.   “Amidst much media attention. . ..” I would predict that if you studied and ranked the biggest baseball news stories of 1959, Roy Face wouldn’t make the Top 50.  

I have on my lap a bound volume of all Sports Illustrated issues between July 13 and December 21, 1959. . .let me see what I can find about Roy Face.   July 13, nothing. . ..key baseball article is about the White Sox, no mention of Face in the “Baseball’s Week” page.   July 20. . .half-sentence mention of Face in “Baseball’s Week”, as follows:  “Those spine-tingling Pittsburgh Pirates won three more games in extra innings (12 out of 13) and relentless Roy Face won two more in relief (14-0).”   Main baseball articles are about a lawyer trying to start a third league and a photo essay about plays at the plate.   July 27.. .nothing, baseball article is about Cleveland GM Frank Lane.  August 3, nothing . ..baseball articles are companion pieces about the Giants and the Dodgers.  August 10, nothing. .cover story is Nellie Fox and Luis Aparicio, also another article about the third league and a page about Larry MacPhail.   August 17, two sentences about Face in “Baseball’s Week”:  “Roy Face, on the other hand, gave up 15 hits and seven earned runs in 10 relief innings without losing.   His record went up to 15-0.”   Two-page article on the NL race (no mention of Face), two pages on rookie Willie McCovey and first installment of long profile of Larry MacPhail. 

August 24, no mention of Face. . .two-page article polling NL players on who will win the race, and second part of the article on MacPhail.    August 31, nothing. . .two pages on the Braves’ blowing a big series, two page glamour photo of Drysdale, third part of article on MacPhail.   September 7, nothing. . .”Baseball’s Week” has substantial entry on the Pirates, but no mention of Face.  Two-page article on the pennant races and long article about little league baseball in the Caribbean.   September 14, nothing. . .three short articles about the pennant races, no mention of Face.   September 21, this in Baseball’s Week:  “Roy Face finally lost one.   After 98 relief appearances and 22 wins over a two-year period, Roy gave up a losing run.  Said Face: `I should have lost before this.  I’ve been lucky and these guys have saved me time after time.  After all, Walter Johnson never got through a season without losing at least one.” Article about the Giants, Braves and Dodgers, article about the Dodgers, article about Reds manager Fred Hutchinson.   September 28, lots of baseball.  Two articles about the Giants (one about them tanking in the pennant race, long feature by Mark Harris about the city adopting the team, two-page World Series preview, two-page World Series scouting report, article about Bill Veeck and the World Series.   No mention of Face.    October 5. ..season wrapup.  Doesn’t mention Face. 

Sports Illustrated—stunningly slanted toward the National League-- mentions Face only three times in this period, never at any length, whereas they mention his teammates Dick Groat and Bob Friend at least eight to ten times each.  I don’t think Face’s season drew much media comment, frankly, but that’s a small point.   I have three other problems with this comment.   First, that Face “blew a lot of leads and ‘vultured’ a lot of wins during that stretch.”. . .I understand that Mike is not offering this comment for the truth of the matter, but as an explanation of Holtzman’s thinking.   But that aside, it’s pretty well out of whack.   Only two or at most three of Face’s wins were “vultured” from other pitchers (Retrosheet doesn’t have an account of one of the games.)   Mostly he picked up wins in games that he entered with the score tied or the Pirates behind.  As to his “vulturing” a lot of wins down the stretch, thirteen of Face’s wins were on July 12 or before.   He won only five games over the last eleven-plus weeks of the season.  

If you actually look at the games that Face was credited with winning, it’s kind of astonishing.   Just in the games that he won, Face pitched 3 full innings on May 3, 3.0 innings on May 13, 3.0 innings on May 31, 3.0 innings on June 8, 5.0 innings on June 28, 3.0 innings on June 28, 3.0 innings on August 9, and 4.0 innings on September 19.   Eight of his 18 wins, he pitched three to five innings. 

The statement that he vultured a lot of wins down the stretch, while it was not offered for the truth of the matter, is clearly not true, but there are still two more problems.  First, did anyone actually suggest, at the time, that Face had “vultured” wins?   I remember very clearly when that phrase, “vulturing wins”, crept into the baseball vocabulary.   It was 1966.  I don’t remember the term, “vulturing”, being used in this way before 1966, and I don’t believe that it was.

One can use the concept without using the word, but. . .did Holtzman—or did anyone—actually think in those terms in 1959?   Mike may have some contemporary quotes that he wants to show me, but I’d be surprised to learn that they did.   The first time I ever heard this kind of comment about Roy Face’s 1959 season was in the 1980s.   The criticism assumes that it is the reliever’s job to “save” wins—which didn’t become true until the late 1970s.   I frankly do not believe that this criticism of Face was ever offered at the time, because I simply don’t see how it could have been.   I don’t doubt that Holtzman may have remembered it that way after the fact, but I think it’s a serious anachronism.    It is reading the present onto the past.

Holtzman was trying to focus on and reward the job that Bob Elston was doing for the Cubs, “saving” a few games, but he was contrasting Elston not with Face, but with starting pitchers.  The starting pitcher gets a Win, the reliever gets nothing, let’s make up something for the reliever.   But even Elston had only 13 Saves in 1959.   “Saving” the game simply wasn’t anyone’s primary job in 1959. 


Holtzman didn't develop the idea that there was such a thing as a"fireman" whose job it was to put out fires, though. By the 1959-1960 time frame, many teams in fact had ONE reliever whose primary job was to do just that - who would pitch virtually all of the late-inning high-leverage situations.

Welllll. . .”virtually all of the late-inning high-leverage situations”, absolutely not.    In 1959-1960, about half of wins were complete games.  There were far more Complete Games than Saves.   A very high percentage of late-inning, high-leverage situations were handled by starting pitchers.  

Setting that aside, is it even true that “virtually all” of the late-inning, high-leverage situations that weren’t handled by starting pitchers were in the hands of one reliever on many teams?   It is not.   In 1960 the Milwaukee Braves had ten Saves from Don McMahon, nine Saves from Ron Piche, and nine from other pitchers.  The Baltimore Orioles had seven from Hoyt Wilhelm, five from Jerry Walker, and two each from five other pitchers.   The Chicago Cubs (3) had eleven from Elston, fourteen from other pitchers.  The White Sox (4) had ten Saves from Gerry Staley, five from Turk Lown, eleven from five other pitchers.  The Cincinnati Reds (5) had 17 from Bill Henry, twelve from Jim Brosnan, six from four other pitchers.   The Indians (6) had 14 from Johnny Klippstein, nine from Dick Stigman, and seven from six other pitchers.  The Tigers (7) had ten from Hank Aguirre, six from Dave Sisler, and nine from six other pitchers.   The Dodgers (8) had eight from Ed Roebuck, seven from Larry Sherry, and five from four other pitchers.   The Washington Senators (9) had 13 from Ray Moore, seven from Tex Clevenger, and fifteen from seven other pitchers.  The Yankees (10) had eleven from Bobby Shantz, nine from Ryne Duren, seven from Luis Arroyo, five from Bob Turley, and ten from six other pitchers.   The Kansas City A’s  (11) save distribution was 4-3-2-2-1-1-1.   The Giants (12) had eleven from Johnny Antonelli, five from Billy Loes, and ten from five other pitchers. 

Granting that there is a difference between “Saves” and “High Leverage Innings”, I would still argue that, in fact, a study of the distribution of high-leverage innings in 1959-1960 would show that at that time only three or four teams had even begun to concentrate those innings in a single reliever.   This statement, in my view, is simply not true. 

There had been isolated instances of such relievers before, but the mid-to-late 50s saw a number of them come forward, mostly in the NL: Face, Lindy McDaniel, Don Elston, Don McMahon, Clem Labine. Some teams had two (Narleski/Mossi in Detroit, Brosnan/Henry in Cincinnati, Staley/Lown for the White Sox) but most teams quickly settled on a primary guy.

            Again, I can’t see it that way.   McMahon was never a key focus pitcher like McDaniel, Elston and Face.    McMahon never saved more than 16 games in a season until the 1970s, and in 1960 he was basically a garbage man.  Labine’s last good year was ’58.  In reality, there were only two or three teams at this time that even knew who their top reliever was. 

When Holtzman developed the idea of the save, it was intended to reward these guys. I doubt Holtzman ever envisaged a situation where teams would use multiple guys to close out a game on a regular basis; at the time it was almost always just one, or at the most two, pitchers.

I agree that Holtzman certainly never envisaged a situation of multiple closers.   I confess that I may be missing the relevance of the point.   

When the save became official in 1969, most teams still only had one pitcher for whom closing out high-leverage situations in the late innings became the primary job.

Well, by 1969 the pattern of relief aces was developing.   But

a)      in 1969 starting pitchers were still pitching a lot of complete games, including a lot of late-inning high-leverage innings, and

b)      the top 24 relievers—one per team—accounted for 401 Saves, out of a major league total of 745, or 54%.   In 2007 the top 30 relievers accounted for 77% of major league Saves. 

That role was fairly consistent from about 1965 through 1990, although there were some managers (Tanner in Pittsburgh being the chief one) who started using other pitchers earlier in the game. Most ace relievers would come into about 2/3 of their high-leverage situations in the ninth, and about 1/3 of them prior to the ninth.

Well, in my view:

a)  Nothing at all was consistent about the use of bullpens between 1965 and 1990—nothing.   Bullpens in that period were revamped and revamped and revamped again.   To try to paint them all into one package with some vague comment about “most ace relievers” simply constitutes turning a blind eye to the sweeping changes that took place in those years.

b)  By using the vague term “high-leverage” situations, rather than the specific and well-documented term “Saves”, this makes a difficult point of reference for us. 

But in 1975, according to John Dewan, 16% of Saves were exactly three outs.   In 1980 this percentage was 18, in 1985 it was 23, and in 1989 it was 36.   Thus, the percentage of Saves that were three-out Saves had more than doubled in the time frame 1975-1989.  

This percentage did explode in the early 1990s, from 39 in 1990 to 70 in 1999.  But it was going up long before 1990, and it continued to go up after 1990.   There is a long, long, long pattern of increasing concentration of Closer’s opportunities into “it’s a Save opportunity; you pitch the ninth.”   The MOST active phase of this was the early 1990s—but the most intense phase of any long-term trend is normally in the middle, isn’t it?   This trend did not start or end in the early 1990s, and it did not primarily “occur” in the 1990s. 


Free agency came into place in 1976. Right after that, teams (a) started spending significant money on top players, including ace relievers and (b) started cutting back on their use in situations OTHER than high-leverage situations. This was true of top-level starters, as well - they started seeing their usage in low-leverage late inning situations go down. That doesn't appear to be a coincidence; the more you have invested in a guy, the more you want to protect that investment. It could be that teams were spooked by what happened to Catfish Hunter after 1975 and to Bill Campbell after 1976 - I really don't know.

I think this is generally correct.

But it's easy to trace the decline in both starter and ace reliever "innings" to the 1977-1979 period. However, while "innings" went down, high-leverage appearances didn't change significantly. The ace was still getting the vast majority of the late-inning high-leverage appearances.

The first true one-inning Closer was not Eckersley. Eckersley became Oakland's full-time Closer in 1988, but he was still making a significant number of high-leverage appearances prior to the ninth (and would continue to do so throughout most of the rest of his career, albeit fewer and fewer as he got older). The first "real" last-inning Closers were
Baltimore's Gregg Olson and Toronto's Tom Henke, in 1991. Both were made one-inning Closers primarily because of injury concerns, and Henke, in particular, was spectacularly successful in the role; he came into games only eight *total* times all season before the ninth that year (it helped that he had Duane Ward in front of him, too). It should be noted that when Henke went to Texas after 1992, he went back into what was a more traditional Closer's role and (a) had just one more good season and (b) saw his career end fairly quickly afterwards.

OK, well. ..let’s assume this is true.   So what? 

It seems to me that what Emeigh is saying. . .and maybe I’m missing the point; I’m a little dense sometimes. .. .but it seems to me that what he is saying is that

a)      the one-inning Save movement started about 1990,

b)      the first places it occurred were in Baltimore and Toronto, therefore,

c)      the motivations of the entire change can be inferred from the experience of Baltimore and Toronto.

If we assume that point (b) is true, then the only one of these statements that is true is (b).   The one-inning Save movement very clearly did NOT start about 1990.   The percentage of one-inning Saves had more than doubled from 1975 to 1989, and the percentage was already at a high level by 1989. 

And how in the world can anyone infer the general motives of other teams from the specific experience of Toronto and Baltimore?  I just don’t see how that follows.

By 1992, other teams were getting on the bandwagon: Boston, with Jeff Reardon; Milwaukee, with Doug Henry; the Yankees, with Steve Farr, and the Cardinals, with Lee Smith (who actually had started moving in that direction late in 1991). Eckersley, in 1992, was still getting into 20 games before the ninth, whereas Smith made just six such appearances. Most of the pitchers involved were (a) the best options available and (b) not terribly effective after more than an inning or so. Smith and Reardon were on the downside of their careers, and Henry and Farr were not, shall we say, supremely talented pitchers. In 1993, Mitch Williams (another guy with questionable effectiveness) took it one step further; he didn't pitch in a single high-leverage situation before the ninth inning all season, the first Closer to do so. It took a few more years after that, but by 2000 the last-inning Closer was more often the case than not.

It seems fairly clear that the teams that were early adopters of the last-inning Closer did so because their options were limited. The first guys that pitched in that role were (a) injured veterans, (b) declining veterans or (c) pitchers with questionable effectiveness beyond a single inning - in addition to Doug Henry and Farr, some of the other guys who fell early into a last-inning Closer role were Mike Fetters and Jose Mesa.

There is not a lot of evidence that the decision to go to a last-inning Closer - or indeed, the change from fireman to Closer - was driven by the save rule; Eckersley, for one, pitched a lot in situations other than save situations.

Wait a minute.  You’re arguing that because, in the 1980s, Dennis Eckersely pitched in a lot of non-save situations, this somehow proves that the Save rule was not a critical factor in the 1990s re-definition of the Closer’s role?   I’m not following you.

The save rule became official in 1969, by which time the majority of teams were already using, or trying to use, one pitcher in the late-inning, high-leverage role.

The changes in that role between 1969 and the early 1990s appear to be driven mostly by free agency and a desire to limit the workload of the higher-salaried pitchers, and the move towards a last-inning Closer appears to be a combination of making use of guys with a history of injury and/or declining effectiveness, and guys who could be effective for an inning but not beyond that.

Well, yes, there was a desire to limit the workload of high-salaried pitchers; that’s certainly accurate.   And it is true that many of the pitchers who were dominating as Closers were either injury-prone or not particularly effective when they pitched longer—like Eric Gagne and Mariano Rivera, both of whom were tried as starting pitchers and sucked at it.   That much of the argument is true.

And the changes in bullpen usage outside of the Closer role are mostly a transfer of lower-leverage innings from good pitchers to lesser pitchers, keeping the better pitchers available more often for the higher-leverage innings.

Not “mostly”, no.   They are a transfer of lower-leverage innings to less-qualified pitchers; that much is true.  There are also a lot of other things going on there.


And from what I can tell, LaRussa was not the primary driver behind the move toward larger and larger bullpens. An early adopter, yes, but never really the chief innovator. From what I can tell, it was mostly Tanner, in the late 70s in
Pittsburgh, and Davey Johnson, with the Mets in the mid-80s, who really started the trend.

It’s a side point, but I think LaRussa played a huge role.   LaRussa’s role in this transition is unique, in that he stayed ahead of the curve for a very long time.    I would put it this way:  that there has been a trend toward more and more and more pitching changes by ever-increasing numbers of relievers in motion since.  .well, since 1885.   Over that time many, many managers, including Leo Durocher, Walt Alston, Gene Mauch, Sparky Anderson and Whitey Herzog, were ahead of the curve early in their careers, but behind the curve by the end of their careers.  

But LaRussa was ahead of the curve in 1982, and ahead of the curve in 2006.   I don’t especially know that he was a key innovator in the use of Closers; that’s not what I’m saying.   I think he was critical in what I call the lunatic accumulation of middle relievers.   Since 1980 we’ve gone from four-man bullpens to seven-man bullpens, and LaRussa was always one of the guys pushing the average up.

Look, there are a couple of issues here that are (I think) just behind a cloud that I sort of agree with Emeigh about.   One is, sabermetricians don’t really understand this problem nearly as well as they think they do.    The problem has a lot more dimensions in real life than it does in a Strat-o-Matic League.  In Strat-o-Matic you can target your best reliever to the most critical innings with very few restraints.   In real life there are very serious and immovable restraints on this usage, whereas in real life it is much easier to target your Closer to Save situations.

The phenomenal effectiveness of pitchers like Papelbon, Rivera and Gagne is created in substantial part by the facts that

a)      they pitch a very limited number of innings, and

b)      they do so on what is, for the most part, a highly predictable schedule.  

They pitch about three times a week, and they know when they’re going to be in the game an hour before they come in, overstating the case a little bit.   It makes a huge difference.   You take them away from that schedule to target them at more high-leverage innings, they’re not the same. 

            A lot of people in the sabermetric community have ignored or underestimated these effects—and other dimensions to the problem as well, which we won’t get into.   I think this is implied in what Emeigh is saying, and I agree with that. 

            Second, health concerns and “fresh arm” concerns have been a driving force in changing the game from one in which relievers worked 130, 140, 150 innings (and more) to one in which they work 55 to 65 highly targeted innings.   That’s absolutely true.

            My problem is that Emeigh seems to be saying that because there are these other factors in the discussion, this somehow proves that the Closer’s role was not defined to capture the maximum number of Saves.   I don’t see the support for that.   I think it was defined to capture the maximum number of Saves for the Closer, and there is very considerable inefficient utilization of resources that results from this. 

            What Emeigh is pointedly ignoring, it seems to me, is that we have evolved into a condition in which there is a definition of the Save rule and a definition of the Closer’s responsibility, and there is essentially a perfect harmony between the two.    He thinks this is a coincidence.  I think it is not a coincidence.   

            Given where the game was in the 1960s, it is perfectly possible that, in the same way that starters were required to pitch five innings to get a Win, relievers could have been required to pitch two innings to get a Save.   This would have been a perfectly rational choice in 1969—not necessarily better than the choice that was made, but not necessarily worse.   In 1964, when Dick Radatz had what we now credit as a career-high 29 saves, for example, he did pitch two innings or more in 21 of the 29 games (and 1.2 in several of the others.)   Two-innings saves were common at that time. 

            Suppose that, in 1969, the Save rule had required two innings to get a Save.  If the usage of pitchers since that time has evolved independent of the Save rule, as Emeigh apparently believes, then this would imply that the same changes would have occurred post-1975 even had the Save rule required two innings.    That would mean that, in modern baseball, there would be virtually no saves, since there are (now) virtually no two-inning saves.  

            The question before you is, do you really believe that that would have happened?   Or do you believe that, in the pursuit of Saves and the glory they carry on their little silver booties, the job would have evolved so that ace relievers routinely pitched the two innings that they needed to pitch?

            Well. .. ..believe what you want to believe; it’s a free country.  I don’t believe for a moment that, once the reliever’s job was defined as earning Saves, ace relievers were going to walk away from those Saves.   It ain’t happenin’ that way.

 
 

COMMENTS (11 Comments, most recent shown first)

studes
Great article and fascinating debate. I personally have two pet peeves with the way relievers are used today, and I think one of them clearly shows the impact of the save rule.

First, I think managers should be much more open to using their aces in the eighth when the heart of the order is coming up in a close game. You hardly ever see that, but I think you should. Go ahead and use your number two reliever in the ninth -- the difference in batters makes it worth it. I find it hard to believe that ace relievers couldn't be ready an inning earlier than the ninth. Of course, I could be wrong about that.

Second, why aren't ace relievers used in the ninth inning of tie games? Those are very high leverage situations and they fit the "one inning, ninth inning" rule. Not using them in these situations is ridiculous, in my opinion. I think this is partially attributable to the save rule, which doesn't accrue in tie situations.
8:30 AM Oct 25th
 
Steven Goldleaf
Not sure if I’m agreeing or disagreeing with Bill here, but let’s take off on one remark of his: If “in 1969, the Save rule had required two innings to get a Save…, then this would imply that the same changes would have occurred post-1975 even had the Save rule required two innings” Taken a step further (or towards Mars) this implication gets even more interesting and possibly more true. If the save rule had rule had been written in any way differently, it would have affected how the best relievers are used, because saves= more salary, and salary is what the best players of any stripe are motivated by. So if the save rule had stated that, say, “a save shall be awarded to a maximum of one relief pitcher (the last relief pitcher if more than one fulfills the following description) who, in a victory that another pitcher earns a win in, after the sixth inning, records the most outs without a run being scored while he is in the game,” then we would see closers being used routinely in seventh inning jams, I think, or to start the eighth inning with the top of the order coming up, or some such variation. What would NOT happen is what happens now: the closer coming in to get the last three outs in the 9th with a three run lead. Those roles would go to a mop up guy, mostly likely, while the closer would have probably gotten two innings’ worth of outs earlier in the game. He would be prized for his ability to accrue the most outs, and not necessarily to get them as the final outs of the game. He would come out of the game, probably, as soon as the first run had been scored off him so as to leave him rested for a future game.

I believe this is one way that the stat, as defined, has affected strategy, and I find it has affected strategy for the worse. I dislike this stat because it discourages managers from thinking. The manager now has a ready-made excuse whenever his closer is ineffective: “Hey, that's his role,” neglecting to mention that the manager knew full well that his closer had pitched in three straight games, increasingly ineffectively, that another pitcher was ready to go, that this other pitcher was healthy while the closer had been vomiting prior to the game, etc. That was his role, ergo he pitched. QED. The manager is not responsible for picking personnel—the stat determines who is used and when.

It also makes all managers closer to being equal, the smart managers and the dumb ones alike. The manager who knows his players, knows his matchups, distinguishes a crucial spot from a merely scary one, has a huge advantage over one less gifted but, under the current save rule, that smart manager is going to use his smarts less often, because he's going to employ his closer in a fairly tightly regimented role, same as his dummy counterpart. Did I mention that I hate this rule?

3:15 PM Oct 18th
 
beta461
Although I generally agree with most of your points I’m uncomfortable with your line-by-line rebuttal. It seems to inherently favor the responder by taking an argument that is intentioned to be read in its entirety and dividing it into disjointed pieces making it much easier to refute.
11:13 PM Sep 22nd
 
Richie
The '3-run 9th inning lead with a somewhat tired closer' would provide a superb 'critical case'. No reason to bring the guy in then except for the stat. Yet it is regularly done, is it not?
11:12 AM Sep 19th
 
3for3
While I agree with the general premise of the article, one situation that the save rule definitely dictates strategy is the dreaded 3 run, 1 inning save. Somehow, I doubt closers would pitch these, especially when they are 'tired'.


2:41 AM Sep 18th
 
elricsi
Well you sure picked that dude apart.

It's obvious that relievers eyes light up with the save. They all want to close and they want as many saves as possible, saves first, team second.

One thing I think Bill can add to the discussion, is what has been his experience with arbitration and saves? Have relievers always been all about saves, or did their $$$ interest peak in a certain year?

P.S. Is it even possible to keep the trend going for more pitchers and less innings with a 25 man roster? I think they'll be going to a 27 man soon.
12:21 AM Sep 18th
 
DonM
For what it's worth . . .

In the Yankees' 154th decision of 1961, Roger Maris was trying to reach 60 home runs and avoid the dreaded asterfrick. The Yanks led 4-3 (I think), and Baltimore brought in Hoyt Wilhelm to pitch the ninth. He got Maris on a groundout, but the Yankees still won.

In Billy Crystal's movie entitled 61*, the event is recreated. "Foul!" cry the Yankees in unison. "Hey, he's only supposed to pitch when they're ahead!"

Wilhelm was definitely NOT a modern one-inning closer. The screenwriter is guilty of anachronism.

10:26 PM Sep 17th
 
wovenstrap
Bill describes the role of the bullpen as constantly evolving. Admitting that there is still a trend in favor of increased bullpen appearances per game in effect, can it be ventured that the 1995-2008 period is a period of *relative* quietude in this process? In 1995 John Wetteland was pitching the 9th for the Yankees when they were ahead by 1 or 2 entering the 9th, and 13 years later Mariano Rivera's job description could not possibly be more identical. It seems to me that the bullpen may be changing, but the Closer is not. And that has everything to do with the Save Rule.
10:18 PM Sep 17th
 
evanecurb
I believe the save rule has been a factor, but I can't think of how the extent to which it has been a factor can be accurately assessed. Fact is, many teams now have two or three good relievers, so it becomes less important which of them pitches the most high leverage situations. Keeping them on a predictable schedule as Bill describes it would seem to maximize everyone's effectiveness.
10:06 PM Sep 17th
 
MattDiFilippo
"As to his 'vulturing' a lot of wins 'down the stretch,' thirteen of Face’s wins were on July 12 or before. He won only five games over the last eleven-plus weeks of the season."

To nitpick, Emeigh actually had Holtzman saying that Face vultured wins 'during that' stretch, not down the stretch.

Otherwise, you've got me convinced. This is an interesting topic.
9:24 PM Sep 17th
 
77royals
If pitcher were still allowed to throw complete games, none of this would matter.
5:41 PM Sep 17th
 
 
©2017 Be Jolly, Inc. All Rights Reserved.|Web site design and development by Americaneagle.com|Terms & Conditions|Privacy Policy