All Saves Are Not Created Equal

This article was written by Gabriel Schechter

This article was published in the 2006 Baseball Research Journal


When the Fireman of the Year award was created in 1960, the term “fireman” had already been in use for more than 20 years, referring to a relief pitcher who entered the game to stop a rally. The connotation was that some emergency existed, requiring the rescue of one pitcher by another. It didn’t matter what inning it was; with a small lead and runners on base, the manager would bring in his best reliever to put out the fire.

After saves became an official statistic in 1969, a generation of relievers built their reputations as firemen who doused rallies as early as the sixth inning and pitched the rest of the way to record saves. More recently, a new breed of relief ace has emerged, one of many specialized bullpen roles. These save-gatherers are spared the hazardous duty of putting out fires in the seventh or even the eighth inning. Instead, their sole assignment is to saunter in at the start of the ninth inning, with a lead of three runs or less, and record the final three outs. They are called“closers” (“fireman” has become obsolete), a business term for the person who irons out the final details of a deal after others have done the legwork. In essence, the game is already won when the closer enters; it is only a matter of what the final score will be.

On the rare occasions when a closer is brought in with no save possible, announcers feel compelled to account for the aberration. Most likely he hasn’t pitched in a few days and needs the work. Some closers have admitted to pitching poorly with a four-run lead because they aren’t sufficiently motivated. It’s as if the manager gears his game strategy toward providing his closer with the chance to accumulate a lot of saves, compared to the earlier generations when the manager identified his best reliever and sought to get as many innings as possible from him, with victories and saves the by-product of quality work. For instance, when Mike Marshall set the record in 1974 with 208.1 innings pitched in relief, 93 of those innings came when he entered with his team losing or tied, and he became the winning pitcher in 15 of those 47 appearances.

 

Table 1. Career Save and Blown Save Totals

Player

G in Relief

SV Opps

SV

BS

Fingers

907

479

341

110

Gossage

965

463

310

112

Sutter

661

412

300

101

L. Smith

1016

616

478

103

Eckersley

710

484

390

71

Hoffman

821

556

482

56

Rivera

720

501

413

55

TOTALS

5800

3511

2714

608

 

Dennis Eckersley has said that “you can’t blame a pitcher for the way a manager uses him.” That is true, but we can assess the relative difficulty of their assigned tasks and their relative success in similar situations. Thanks to a wealth of data supplied by Dave Smith of Retrosheet, I’ve conducted numerous studies of seven relievers most prominently mentioned in debates about electing relievers to the Hall of Fame. Three are already enshrined: two firemen (Rollie Fingers and Bruce Sutter) and one closer (Eckersley). Two others were the top vote-getters among relievers on the 2007 ballot who were not elected: Rich Gossage and Lee Smith. The final two are the active pitchers regarded as the most likely to make the Hall of Fame someday: Mariano Rivera and Trevor Hoffman.

 

Table 2. Innings Pitched in Saves

Players

1/3

2/3

1

1 1/3

1 2/3

2

2 1/3

2 2/3

3+

Fingers

39

20

81

30

36

61

20

18

36

Gossage

24

23

70

33

35

73

16

12

24

Sutter

18

12

82

25

33

84

21

10

15

L. Smith

26

23

260

46

29

79

5

2

8

Eckersley

25

28

231

44

34

23

2

2

1

Hoffman

26

12

389

39

9

5

1

1

0

Rivera

16

11

299

49

27

10

1

0

0

TOTALS

174

129

1412

266

203

335

66

45

84

 

As a group, these seven standouts provide a vivid cross-section of the evolution from firemen to closers during the “saves era”. 

Did the elite relievers of the 1970s work that much harder than today’s elite, and do they deserve more respect for doing so? A reliever’s workload—his contribution to the team’s winning effort—is easily measured in innings pitched. One argument in favor of enshrining Bruce Sutter was that even though he ranks just 18th in career saves, he worked harder for his saves, often pitching two or three innings to do so. So in Table 2 I tallied exactly how many outs were recorded in every one of these pitchers’ saves.

The numbers tell us quite a bit. Sutter pitched at least two innings in 43.3% of his career saves, more than any of the others. Gossage and Fingers weren’t far behind, and Fingers pitched at least three innings in more than 10% of his saves. It is impossible to pick any member of this trio over the other two in terms of how hard they worked for saves.

Contrast their innings with those pitched by Eckersley, Rivera, and Hoffman. The great majority of their saves involved pitching one inning or less, with few appearances earlier than the ninth inning. Consider this: from May 27 through July 4, 1984 (39 days), Sutterhad more saves (nine) where he pitched at least two innings than Hoffman has in his whole career. Gossage did the same thing from August 15, 1980, through the end of that season, and Fingers accomplished it in a 53-day stretch in 1978. The earlier pitchers acted as their own setup men. These firemen put out the  fire and cleaned up after themselves.

Rivera’s work in postseason play proves that he is quite capable of shouldering a heavier burden. Manager Joe Torre has not hesitated to bring him in early. In 27 of his 34 post-season saves (79.4%), Rivera has entered in the eighth inning. A dozen times (35.3%), he has worked two full innings for a save. Throw in a career ERA of 0.81 in the postseason, and it’s no wonder that he is considered a shoo-in for the Hall of Fame.

The change from multiple-inning to one-inning closers is seen most dramatically in the career of Lee Smith. From 1981 to 1990, he carried a load similar to the earlier trio. From 1991 through the end of his career, he was used much the same way that Hoffman has been. The statistical breakdown reflects the shifting trend. From 1981 to 1990, 44.2% of Smith’s saves lasted one inning or less, a little more than Fingers & Co.; from 1991 on, that figure is a whopping 90.1%. His saves of 2+ innings went from 34% all the way down to 1.9%. Smith recorded his top four seasonal totals for saves after 1990, thanks to his managers lightening his workload as he grew older.

The biggest difference between the “old-style” firemen and the current crop of closers is the number of times they enter the game to start the ninth (or extra) inning, with no runners on base, the easiest situation for a reliever to face even with just a one-run lead.

Figures supplied by Tom Ruane of Retrosheet indicate that if the home team starts the ninth inning with a one-run lead, it will win roughly 85% of the time. Put the leadoff runner on first, and the percentage drops to 75%, the same likelihood as having runners on first and second with nobody out and a two-run lead, or the bases loaded and nobody out with a three-run lead.

Start the ninth inning with a two-run lead, and you’ll win about 93% of the time; with a three-run lead, it jumps to a 97% win rate. Current managers love to put in their big-time closer with that three-run lead in the ninth inning because victory is a near-sure thing, but it would be a near-sure thing no matter who pitched the final inning.

Trevor Hoffman has been used in this situation 124 times in his career, compared to 11 for Fingers, 14 for Gossage, and 16 for Sutter. When Fingers recorded three outs for a save, he started the ninth inning only 65.4% of the time; Gossage got to start the ninth in 72.9% of his three-out saves. That is, about a third of the time they got the call only when the previous pitchers put runners on base. In 1975, Gossage pitched in 62 games and only twice entered without runners on base. By contrast, Hoffman faced only one inherited runner the entire 2006 season (recording the easiest save imaginable, retiring one batter with a runner on first base and a three-run lead).

 

Table 3. Performance When Entering to Start the Ninth Inning

Player

One-Run SV/BS

Two-Run SV/BS

Three-Run SV/BS

Fingers

25/12

17/2

11/0

Gossage

21/6

17/2

13/1

Sutter

28/10

22/4

16/0

L. Smith

96/30

87/9

61/0

Eckersley

78/22

78/8

65/1

Hoffman

139/26

121/9

117/7

Rivera

94/20

102/3

94/2

TOTALS

481/126

444/37

377/11

 

Table 3 breaks down the performance according to the size of the lead. As a group, the seven relievers have gotten the save 97.2% of the time with a three-run lead and 92.3% of the time with a two-run lead, very close to Tom Ruane’s figures. The most striking thing is the high percentage of the time that the modern closers start the ninth inning, especially with more than a one-run lead.

Fingers and Gossage enjoyed this relatively carefree entrance in only one-sixth of their saves. It happened a little more often for Sutter, but still only 22% of the time. For Smith, it was 27.2% in the first part of his career, but 78.9% in the second part. Hoffman has had it very easy by this standard, with more than three-fourths of his career saves (78.2%).

Dan Quisenberry, the unjustly overlooked relief ace of the 1980s Kansas City Royals, advocated measuring saves by “degree of difficulty.” I have attempted to do just that in my studies, examining the various “save situations” in which a reliever enters the game. First, I looked at the most difficult jam, with not only the tying run(s) on base but the (potential) winning run as well.

 

Table 4. Performance When Entering with the Winning Run on Base

Player
WROB
SV
BS
Fingers 50 24 25
Gossage 45 22 21
Sutter 26 11 15
L. Smith 27 14 13
Eckersley 15 7 8
Hoffman 17 11 5
Rivera 10 3 7
TOTALS 190 92 94

 

The finding that jumps off Table 4 is that even the best relievers blow the save in this situation more often than not. Hoffman has done the best and Rivera the worst, though the more significant point is that their opportunities are so few compared to Fingers and the earlier relievers. Give Rivera as many appearances as Fingers with the winning run on base and, using his “success” rate, we’d be adding 28 blown saves to his career total. Conversely, putting Fingers in that spot as seldom as Rivera has faced it would lop off 20 blown saves from his total. This doesn’t even take into consideration what inning it is or how many outs there are.

The folks at Rolaids, who hand out the annual award for relief pitching, have tallied “tough” saves since 2000, defined as having the tying run on base when the reliever enters. In the past seven seasons, Rivera has more “tough” saves than any other reliever, 20, which happens to be only one more than John Hiller had just in 1973. The saves in Table 4 are “tough” by this definition. Here’s the data from entries with the tying run(s) on base but not the go-ahead run.

 

Table 5. Performance When Entering with Tying Run(s) On Base

Player
TROB
SV
BS
Fingers 118 77 35
Gossage 102 59 36
Sutter 80 47 31
L. Smith 59 37 20
Eckersley 42 28 11
Hoffman 36 25 8
Rivera 41 26 13
TOTALS 478 299 154

 

As in the previous table, Fingers has as many of these dangerous outings as Rivera, Hoffman, and Eckersley put together. As a group, these seven stalwarts recorded the save less than twice as often as they blew it, a measure of the difficulty of handling inherited runners. Taking these two tables together, the career “tough” saves add up to: Fingers 101, Gossage 81, Sutter 58, Smith 51, Eckersley 35, Hoffman 36, and Rivera 29. For Fingers and Gossage, more than half of their career blown saves came in these spots, and in nearly half of those blown saves, they entered the game in the sixth or seventh inning. For Rivera and Hoffman, most of their blown “tough” saves come in the eighth inning, virtually the only time they enter with inherited runners.

 

Table 6. Performance When Entering With Tying Run at Bat

Player
TRAB
SV
BS
Fingers 175 126 37
Gossage 174 116 42
Sutter 161 117 41
L. Smith 259 189 54
Eckersley 168 128 37
Hoffman 188 158 27
Rivera 164 129 26
TOTALS 1289 963 264

 

This data is, on the surface, more comparable, since all seven relievers faced this situation roughly the same percentage of the time, 36–39% of their save opportunities for the earlier guys and 32–35% for the later group, with Lee Smith at 42%. The career figures for Fingers and Eckersley are almost identical, as are those for Sutter and Gossage. Hoffman and Rivera have significantly higher ratios of saves to blown saves when facing the tying run at the plate, but a breakdown of the situations reveals why.

For Fingers, Gossage, and Sutter, more than two-thirds of their blown saves came when they entered no later than the eighth inning, meaning they not only had to get out of their first jam, they also pitched multiple innings and therefore had extra chances to blow the lead. Of Hoffman’s 27 blown saves in this category, only one came when he entered before the ninth inning; similarly, for Rivera it was only five out of 26. For them, this situation usually occurs when they enter to start the ninth inning with a one-run lead.

Add up the evidence and it’s clear that all saves are not created equal. Some save “opportunities” are gift-wrapped while others are booby-trapped. A whopping 59.8% of Hoffman’s career saves have come when he entered the game with no more peril than having the tying run in the on-deck circle. It’s even higher for Rivera at 61.7%. The percentage goes down as we look further back: Eckersley 58.2%, Smith 49.8%, Sutter 41.7%, Gossage 36.5%, and Fingers 33.4%. Compare the stats of the seven studs for these “easy saves” compared to the“tough saves” in which the tying run is on base.

 

Table 7. Tough Saves vs. Easy Saves

Player
Tough

SV/BS

Tough Ratio

Easy

SV/BS
Easy Ratio
Fingers 101/60 1.68 114/13 8.8
Gossage 81/57 1.42 113/13 8.7
Sutter 58/46 1.26 125/14 8.9
L. Smith 51/33 1.55 238/16 14.9
Eckersley 35/19 1.84 227/15 15.1
Hoffman 36/13 2.77 288/16 18.0
Rivera 29/20 1.45 255/9 28.3
TOTALS 391/248 1.58 1360/96 14.2

 

Fingers, with almost as many tough saves as easy saves, had a better success rate in those dangerous situations than Rivera, the most revered of current closers. Gossage’s success rate was virtually the same. Why is Hoffman’s ratio of tough saves to blown tough saves so much higher? Of the 102 career saves he has recorded in which he inherited runners, 65 came when he entered with two outs, and 26 of those were in the ninth inning. Only two of Hoffman’s 482 saves saw him enter before the eighth inning, compared to 75 for Fingers.

This perspective suggests the difficulty of devising a unifying formula to evaluate all save performances in  their  situational  context.  Such a formula must take into account the immediate danger when the pitcher enters, where the runners are, how many outs, the size of the lead, how far he is from the end of the game, and run support. For instance, the fire is blazing when you enter in the seventh inning, but your team gives you a six-run cushion for the last two innings. How much easier is your save than the one where you have to nurse a one-run lead after the seventh inning, and how much tougher than facing the winning run when you enter in the ninth inning?

 

I believe it’s possible to devise a formula which will satisfy Dan Quisenberry’s wish for a “degree of difficulty” for saves, and which can be calculated by any fan watching the game. Until that time, Table 8 contains a final look at how our seven elites measure up in the separate parameters when entering the game.

 

Table 8. Performance In Game-entering Parameters

 

Fingers

Gossage

Sutter

L.Smith

Eckersley

Hoffman

Rivera

TOTALS

1-run lead

2-run lead

3-run lead

4+-run lead

117-77

114-23

80-6

30-4

116-69

101-34

67-7

26-2

97-71

103-23

84-6

16-1

162-76

169-20

120-7

27-0

120-43

130-21

109-5

31-2

167-34

136-15

141-7

38-0

127-37

132-13

120-2

34-3

906-407

885-149

721-40

202-12

0 runners

1 runners

2 runners

3 runners

138-38

88-28

104-39

11-5

128-38

96-35

69-29

17-10

152-45

79-28

59-27

10-1

320-64

83-19

67-17

8-3

265-43

68-12

48-12

9-4

380-43

31-3

59-8

12-2

314-28

46-13

43-12

10-2

1697-299

491-138

449-144

77-27

0 outs

1 out

2 outs

171-59

79-27

91-24

160-58

73-40

77-14

180-61

56-24

64-16

347-74

53-14

78-15

257-42

62-16

71-13

394-47

22-4

66-6

309-34

38-10

66-11

1818-374

383-135

513-99

6th or 7th

8th

9th

75-50

125-27

141-33

52-36

142-56

112-23

46-22

142-56

112-23

15-12

153-48

310-43

5-4

101-30

284-37

2-2

53-10

427-44

1-3

87-23

325-29

196-129

790-231

1728-248

 

GABRIEL SCHECHTER has been a research associate at the National Baseball Hall of Fame’s library since 2002, and is the author of three baseball books.

© SABR. All Rights Reserved